# Janeen Baxter · Jack Lam · Jenny Povey · Rennie Lee · Stephen R. Zubrick  *Editors*

# Family Dynamics over the Life Course

Foundations, Turning Points and Outcomes

# **Life Course Research and Social Policies**

Volume 15

#### **Series Editors**

Flavia Fossati, Bâtiment IDHEAP, Quartier UNIL-Mouline, Lausanne, Switzerland Andreas Ihle, University of Geneva, Geneva, Switzerland Jean-Marie Le Goff, LIVES, Batiment Geopolis, University of Lausanne, Lausanne, Switzerland Núria Sánchez-Mira, Bâtiment Géopolis, Quartier UNIL-Mouline, Lausanne, Switzerland Matthias Studer, NCCR Lives, University of Geneva, Genève, Geneve, Switzerland

The book series puts the spotlight on life course research. The series publishes monographs and edited volumes presenting theoretical, methodological, and empirical advances in the study of the life course, thereby elaborating on possible implications for society and social policies applications. Topics appropriate for the series include, among others, the following: *Life Course Research and Social Policies*


Books commissioned for the series aim to encourage debates on life course research in various countries and regions across the world. Volumes in this series will be of interest to researchers, professionals and policy makers in social sciences and related felds.

The series is edited by a team of scholars affliated to the Swiss LIVES Centre: Flavia Fossati (UNIL), Andreas Ihle (UNIGE), Jean-Marie Le Goff (UNIL), Núria Sánchez-Mira (UNIL) and Matthias Studer (UNIGE).

If you are interested in flling a gap in coverage, providing a focus on a certain area, or contributing a new perspective or approach, we would be delighted to receive a book proposal from you. The book proposal should include a description of the proposed book, Table of Contents, unique or special features compared to competing titles, anticipated completion date, and CV with brief biography. The book proposal can be sent to the publisher, Evelien Bakker, at evelien.bakker@ springer.com.

Janeen Baxter • Jack Lam Jenny Povey • Rennie Lee • Stephen R. Zubrick Editors

# Family Dynamics over the Life Course

Foundations, Turning Points and Outcomes

#### *Editors*

Janeen Baxter Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research The University of Queensland Brisbane, QLD, Australia

Jenny Povey Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research The University of Queensland Brisbane, QLD, Australia

Stephen R. Zubrick Australian Research Council Centre of Excellence for Children and Families over the Life Course, Telethon Kids Institute University of Western Australia Perth, WA, Australia

Jack Lam Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research The University of Queensland Brisbane, QLD, Australia

Rennie Lee Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research The University of Queensland Brisbane, QLD, Australia

ISSN 2211-7776 ISSN 2211-7784 (electronic) Life Course Research and Social Policies ISBN 978-3-031-12223-1 ISBN 978-3-031-12224-8 (eBook) https://doi.org/10.1007/978-3-031-12224-8

© The Editor(s) (if applicable) and The Author(s) 2022 **Open Access** This book is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made. . This book is an open access publication.

The images or other third party material in this book are included in the book's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the book's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specifc statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

## **Acknowledgements**

This collection draws on research undertaken for the Australian Research Council Centre of Excellence for Children and Families over the Life Course (Life Course Centre). The Life Course Centre is a multidisciplinary, multi-partner collaborative research centre designed to investigate the transmission of social disadvantage over the life course and across generations and to provide evidence and solutions to reduce it. The Centre is a collaboration across four universities in Australia: The University of Queensland, the University of Sydney, the University of Western Australia and the University of Melbourne. Additionally, the Centre is partnered with experts in several international universities, as well as key government and non-government agencies in Australia working to design and deliver policies and programs to mitigate social disadvantage. The Life Course Centre was initially funded in 2014 for a period of 7 years, and we were fortunate to receive funding for a further 7 years in 2020. We gratefully acknowledge the funding and support from the Australian Research Council Centre of Excellence for Children and Families over the Life Course (CE140100027 and CE200100025), as well as our partners and colleagues who have contributed to the research and evidence underlying this collection.

Each of the chapters uses recent, unique and high-quality Australian data. In many cases, the data are drawn from longitudinal surveys funded by the Commonwealth Government of Australia. These include: the *Household, Income and Labour Dynamics in Australia Survey*, conducted by the Melbourne Institute: Applied Economic & Social Research and funded by the Australian Government Department of Social Services; *Growing Up in Australia*: *The Longitudinal Study of Australian Children*, conducted in partnership by the Department of Social Services, the Australian Institute of Family Studies and the Australian Bureau of Statistics; *Footprints in Time: The Longitudinal Study of Indigenous Children*, initiated and funded by the Australian Government Department of Social Services; *Building a New Life in Australia: The Longitudinal Study of Humanitarian Migrants* survey conducted by the Australian Government Department of Social Services; and the *Longitudinal Survey of Australian Youth*, funded by the Australian Government Department of Education, Skills and Employment. We are very grateful to the data managers, funders and participants in each of these studies and to the Australian government for their foresight and vision in building such a rich and diverse range of high-quality, publicly available longitudinal data assets for Australia. We hope that the analyses provided here will not only showcase the power of these data for answering critical scientifc questions but also provide useful evidence for policymakers and service providers working at the frontline to reduce social disadvantage in Australia.

The editors would like to particularly thank each of the chapter authors for generously agreeing to contribute their time and expertise to the volume. We are delighted with the quality of the chapters, the depth of the analyses and the coverage of life course stages reported here. We believe the chapters showcase some of the best Australian expertise in life course and longitudinal analyses, and we hope that readers will fnd the research engaging, informative and useful. The idea for this volume originally emerged at a Life Course Centre research retreat held in Perth, Western Australia, in 2019. As with all research ideas, successful execution requires a great deal of hard work and dedication, and we are grateful to our colleagues who stayed the course and to those who joined the project to ensure it could be delivered.

The editors were ably assisted by research and administrative support from Sarah-Ann Burger and Heidi Hoffman in the Institute for Social Science Research at The University of Queensland. We thank them both for their extremely valuable behind-the-scenes work on the volume and for helping to ensure its successful and timely completion. More broadly the Life Course Centre is supported by an exceptional team of professional staff that ensure the Centre operates seamlessly and productively across all of its many activities and commitments. All of the researchers in the Centre owe a huge gratitude of thanks to these staff for providing the strong foundations and governance on which we depend to successfully deliver our research.

We are very thankful to Springer, the series editors and our reviewers for their encouragement and support of our work. The opportunity to share our fndings with colleagues across the globe is invaluable and highly rewarding. Portions of some chapters have appeared previously in publication and are used here with permission. We thank the publishers for permission to reuse selected material.

Finally, we are deeply indebted to our families and friends for their ongoing support and encouragement. As each of the chapters demonstrate, families and personal relationships provide the essential, often invisible, support and resources that enable goals and opportunities to be realised. This book is no exception. Our loved ones have helped in numerous ways to ensure we were able to produce this volume, and we thank them for their love and support in this and all of our endeavours.

# **Contents**





# **Chapter 1 Introduction**

#### **Janeen Baxter, Jack Lam, Rennie Lee, Jenny Povey, and Stephen R. Zubrick**

Families are the starting point for life course journeys and all families want to give their children the best possible start in life. As children grow older and transition through key life course stages, such as schooling, entry to the labour market, relationship formation and parenthood, families continue to shape experiences and outcomes. There is ample evidence across many studies of the long reach of family background on opportunities and outcomes well into adulthood and across generations (Bradbury et al., 2015; Duncan & Murnane, 2011; Ermisch et al., 2012; Huang et al., 2021; Lersch & Baxter, 2021). But families vary in the extent to which they can provide the resources, skills and environments that enable children and young people to achieve their dreams. Understanding what these differences are, how they vary across social groups and how they are supported or impeded by key social institutions and across different social contexts is the focus of this book.

Our approach is guided by a life course perspective. Life course theory is a multidisciplinary perspective that foregrounds the importance of timing and sequencing of key life events, variations across time and context, connections between individuals through linked lives and agency for understanding variations in pathways and outcomes (Elder & Giele, 2009; Mayer, 2009). As discussed in Chap. 2, different disciplines emphasise different aspects of life course theory, but central to all

e-mail: j.baxter@uq.edu.au; j.lam@uq.edu.au; rennie.lee@uq.edu.au; j.povey@uq.edu.au

J. Baxter (\*) · J. Lam · R. Lee · J. Povey

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia

S. R. Zubrick

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Telethon Kids Institute, University of Western Australia, Perth, WA, Australia e-mail: stephen.zubrick@telethonkids.org.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_1

approaches is the notion of connections over time as individuals move through culturally and historically defned stages infuenced by past experiences, events and circumstances. A life course approach draws attention to both micro-level individual factors such as cognition, personality and biology as well as social and cultural forces at the macro level, like education systems, labour markets and social and cultural values. Importantly for us, a life course approach emphasises family relationships and resources as key to understanding how individual biographies develop and play out over time in both the short and long-term.

Life course journeys are always somewhat unique. There is debate about whether life courses are becoming more variable and less standardised over time or whether most people follow broadly similar, albeit different to previous generations, pathways (Macmillan, 2005). Some have suggested that individuals are increasingly able to pursue their own individual biographies free from many of the normative expectations and constraints imposed in previous times (Beck & Beck-Gernsheim, 1995). This may include delaying major life events, skipping some, adding new ones and reorganising the timing and sequencing of others (Macmillan, 2005). On the other hand, there is also evidence that most people still follow broadly normative pathways (Brückner & Mayer, 2005), although with some new stages added, such as emerging adulthood as examined in Chap. 8, cohabitation before marriage as discussed in Chap. 10 and a new stage following retirement but prior to old age, referred to by Moen (2016) as "encore adulthood."

Andrew Cherlin (2004, 2009) elegantly captured some of these ideas in relation to changes in the timing and ordering of marriage. As discussed in Chap. 10, Cherlin argued that many of the norms that once guided marriage have been weakened or lost as societies become increasingly secular and legal institutions recognize a diversity of family types—a process he refers to as the "deinstitutionalization" of marriage. But just as marriage has lost much of its practical signifcance through deinstitutionalization, it has taken on a new symbolic status according to Cherlin (2004). Marriage he argues, is now something to be celebrated and achieved as the capstone to one's life after other goals, such as having children, buying a house, establishing a career or steady employment, have been achieved. Marriage has thus shifted he argues, from being one early marker of adulthood to a prestige status that is a capstone of adult life.

As argued by Cherlin (2004) and others (see for example, McLanahan, 2004), at least part of the variability in life course pathways is a result of circumstance and opportunity, not choice. Not everyone has the same opportunities to achieve milestones at the same age and in the same way and some may never reach certain milestones. There is also evidence that individuals, especially young adults, are remaking and redefning traditional markers of adulthood, subverting conventional conceptions of adulthood taken-for-granted in prior generations (Hartmann & Swartz, 2006; Manning, 2020). Unfortunately, academic research often lags behind in developing theories and methods to understand the new life course pathways of young adults, beyond the normative and linear approach often undertaken (Manning, 2020; Roy & Jones, 2014) though there have been some new advances in recent years (Sánchez-Mira & Bernardi, 2021). Nevertheless, it has been clearly shown that individuals with fewer socio-economic resources will face larger hurdles in achieving certain life course milestones, such as developmental skills and competencies in early childhood, high school completion or long-term intimate partnerships, than those with greater socio-economic resources.

There have been enormous changes in Australian family life in recent decades, including married women's movement into paid work, increasing social acceptance of alternative family forms, legislative and policy support for gender equality and greater visibility and acceptance of non-conventional forms of sexuality and intimate partnerships. At one level, these changes suggest that families are more egalitarian than ever, and individuals have greater choice and agency without the constraints and restrictions that shaped the lives of previous generations. But we also know that economic inequality is substantial and possibly increasing (Huang et al., 2016; Davidson et al., 2018; United Nations, 2020), gender equality has stalled and is declining on some measures (World Economic Forum, 2019; WGEA, 2019) and social stratifcation along class lines means there are identifable groups who continue to have limited opportunities for fnancial prosperity and wellbeing, as measured by outcomes such as income mobility, fnancial security, social inclusion and loneliness (Scutella et al., 2009; McLachlan et al., 2013; Productivity Commission, 2018). These inequalities inevitably place constraints on life course choices.

Our aim in this book is to provide unique and up-to-date Australian evidence about contemporary life course pathways in Australia. As might be expected in an edited collection, the chapters vary in the extent to which they articulate how life course theory applies to their topics, but all are guided by broad life course principles in their focus on the role of families and personal relationships in the transmission of (dis)advantage over the life course. While focused on Australia, we hope that the issues covered are of relevance and interest to researchers across a range of countries. We use insights from life course theory and recent high-quality data to examine the transmission of social and economic inequalities within families over the life course. We examine variations across social groups, focus on key life course stages, and draw out the implications of the research for public policies designed to support families and ensure greater equality for all groups.

Although we do not focus explicitly on social class as an organising frame for our analyses, Australia, like many other advanced capitalist countries, is a country with deep class divisions and inequalities (Connell, 1977; Connell & Irving, 1979; Chamberlain, 1983, Baxter et al., 1991; Threadgold & Gerrard, 2022). Many of the outcomes we consider could be examined using the language and concepts of class analysis and our work aligns with the views of class theorists that disadvantage is social and relational and generated by the social positions that individuals occupy rather than individuals themselves. But consistent with the multidisciplinary and applied focus of our work, we use terms like social disadvantage, inequality of outcomes, and differential access to opportunities and resources rather than social class. We believe these terms are consistent with a life course approach, and they resonate more easily across disciplines and to policy makers and service providers charged with improving distributions of resources and access to opportunities.

Australia is in the fortunate situation of having several high-quality, governmentfunded, longitudinal studies that collect detailed information about life course journeys for the whole population, but also on specifc social groups, for example children, humanitarian migrants and Indigenous children. These studies have been running for many years and form the basis for much of the analyses presented in the chapters that follow.

The book is motivated by three broad research questions that have guided our selection of chapter topics:


#### **The Australian Context**

Australia is a wealthy, politically stable, liberal welfare country with high living standards, universal health care, relatively low levels of unemployment, a system of government-funded welfare safety nets and a relatively resilient economy. It is in many ways a "lucky country" with Australia one of the few countries that avoided an economic recession during the Global Financial Crisis in 2009 (Productivity Commission, 2018) and experiencing a sustained period of economic growth over the last 25–30 years, in part driven by a mining and investment boom. Labour force participation rates have grown from 63% in 1989 to 65% in 2016, and much more rapidly for women than men, by about 7 percentage points over the same period, and unemployment has dropped from almost 8% in the early 1990s to under 5% in 2021 (Productivity Commission, 2018; ABS 2021). But the Australian economic story is not straightforward and the statistics above mask other trends—rising housing costs, low wage growth, a concentration of women in part-time employment, and rising gender wage gaps—that leave some groups behind and point to persistent economic and social disadvantage (Davidson et al., 2020; Productivity Commission, 2018). This means that not all Australians have the same opportunities to move smoothly along a standard normative life course pathway with some groups consistently missing out. Key social groups that fall into this category include women, migrants (particularly refugees and humanitarian migrants), Aboriginal and Torres Strait Islander groups, and LGBTQI+ people.

Relative income poverty, defned as half of the median Australian disposable income, is the level deemed necessary to maintain an acceptable standard of living (Davidson et al., 2020). In 2020, 3.24 million Australians (13.6% of the population) were estimated to be living below this poverty line, including 774,000 children (17.7% of all children) (Davidson et al., 2020). Material deprivation represents the balance between resources and the capacity to meet essential needs and affected just under 12% of Australians in 2015–2016 (Productivity Commission, 2018). Multiple deprivation (lacking at least 2 of a list of defned 'basic needs') is more common among Indigenous Australians (40% of Indigenous Australians), single parents, those with a disability that severely restricts their ability to work, and those receiving welfare as their main source of income (Productivity Commission, 2020).

Gendered differences in lifetime economic earnings for Australian women remain substantial. In May 2020, the average female wage was 86% of the average male wage (full time adult average weekly ordinary time earnings) and had not changed since the previous year (ABS, 2020b). Women are more likely to work part-time, with women forming 68.7% of all part-time employees in 2019 (DESE, 2019). Among parents with a child under six, 94.4% of men participate in the labour force, compared to 65.5% of women, and among these employed parents, less than 10% of men are working part-time, compared to two thirds of women (ABS, 2020b). In 2017, retiring women had 47% less super than men (Dawson et al., 2020).

In addition, Australian occupations are segregated by gender. Only 46.5% of Australians work in gender-mixed organisations; that is organisations which are neither male- or female- dominated (WGEA, 2019). This gender segregation is signifcant as female-dominated industries continue to pay lower salaries to employees on average. In 2015, the median starting salary for graduates with a Bachelor's degree was \$2000 higher for men than for women (ABS, 2016). However, this gap has been decreasing since 2012. Educational pathways also vary by gender, with 40% of women aged 25–29 and only 30% of men having attained a Bachelor's degree or above in 2016, but men aged 15–24 three times more likely to be enrolled in an apprenticeship or traineeship than women (ABS, 2016).

Single parenthood predominantly falls on women. Of the 1 million single parent families in Australia (families classifed as two related people, at least one of whom is over the age of 15, living in the same household), 64.8% (663,000 families) have dependents. Of the single parent families with dependents, 81.6% are single mothers (ABS, 2020c). Single motherhood is associated with lower workforce participation compared to single fathers, with 57.6% of single mothers in employment compared with 76.2% of single fathers (ABS, 2020c). Single parents are less likely to be employed when their children are young and become more likely to be employed as their children age, with this trend stronger among single mothers.

Educationally, migrants are not disadvantaged on average in comparison with the general population. Of adult migrants (those who arrived in Australia after the age of 15), 75% have a non-school qualifcation, compared to 63% of the general population (ABS, 2020d). Around 45% of adult migrants arrive in Australia with one non-school qualifcation, while 44% complete one (or another one) in Australia. However, there are important differences across migrant groups with some non-English speaking background groups faring considerably worse than their Englishspeaking counterparts in terms of employment and earnings and in particular, refugees and humanitarian migrants showing much worse outcomes on a range of indicators than migrants entering Australia though other means (Fozdar & Hartley, 2014; Hugo, 2014).

There are stark contrasts in outcomes for Indigenous compared to non-Indigenous populations in Australia. Despite government targets since 2008 for closing the gap between Indigenous and non-Indigenous outcomes across a range of areas including child mortality, early childhood education, school attendance and completion, reading and numeracy, and employment and life expectancy, only two indicators (early childhood education and school completion) were defned as on track in 2019 (Commonwealth of Australia, 2019). Life expectancy at birth in 2015–2017, for Aboriginal and Torres Strait Islander males was estimated to be 71.6 years, 8.6 years less than life expectancy at birth for non-Indigenous males and 75.6 years for Aboriginal and Torres Strait Islander females, 7.8 years less than life expectancy at birth for non-Indigenous females (ABS, 2018b). Aboriginal and Torres Strait Islander adults and children have a suicide rate double that of the non-Indigenous population (ABS, 2020a). In the period 2015–2019, suicide was the leading cause of death among Aboriginal and Torres Strait Islander young people, accounting for 32.4% of deaths among Aboriginal and Torres Strait Islander children. In 2017, the Indigenous child mortality rate (0–4 years) was 2.4 times that of non-Indigenous children (Department of Prime Minister and Cabinet, 2019).

These differences are further differentiated by geographic location. Fifty percent of Aboriginal and Torres Strait Islanders living in remote areas are unemployed compared to 39% in non-remote areas (ABS, 2018a). Only 25% of Indigenous people aged 20–64 years living in remote areas reported completing Year 12 at the 2016 census compared to 39% of those living in non-remote areas (ABS, 2018a). More broadly, for the Australian population as a whole, educational outcomes refect remote and regional disadvantage. Living remotely is associated with one third of the likelihood of going to university compared to living in a major city (Cassells et al., 2017).

There is also evidence in Australia, like other western countries, of intergenerational inheritance of disadvantage. One example is educational attainment. Nearly two-thirds of people in Australia with a parent who attained a university qualifcation also attain a tertiary qualifcation compared to 21% of people whose parents highest qualifcation is Year 10 or below (Cassells et al., 2017). Another example is government support. Young people from families reliant on welfare are much more likely to receive welfare support themselves. Bubonya and Cobb-Clark (2019) fnd that the main mechanism is incomplete high school, due to disrupted educational conditions and a lack of fnancial support.

The Australian welfare system has been categorised as broadly consistent with a liberal welfare democratic regime, according to Esping-Andersen's typology (1990). Countries in this category are characterised by means-tested welfare payments, low levels of social security payments, and a focus on employment as the main pathway to self-suffciency and independence. But Australia also has a relatively generous aged care pension system and universal health care. The Australian government recently initiated a "Try, Test and Learn" fund based broadly on a social

investment approach. The intention of the fund was to support trials of new codesigned social interventions aimed at moving those most at-risk of long-term welfare dependence into employment. Several priority groups were identifed of particular concern including young parents, migrants and refugees, older unemployed and young students and carers. Interventions were primarily tailored to individuals in one of three work transition phases: work-ready; developing work readiness; and those with limited capacity to work due to experiencing nonvocational barriers. Evaluation of the trials was undertaken using government administrative data, surveys and some in-depth interviews with participants and service providers. Critics of the fund point to the potential of this approach to surveil the poor by identifying and monitoring outcomes of certain "problem" groups, a tendency to overlook the complexity of individual lives by focusing on a single outcome (employment), and the assumption that movement off welfare is always a desirable outcome (Staines et al., 2020). While these are all valid concerns, it is also the case that the fund was broadly consistent with a life course approach that recognises the importance of intervening early to support new opportunities. The fndings pointed to the importance of targeted programs that take account of specifc circumstances and especially the level of work-readiness of individuals. Those who are work-ready may beneft from programs designed to support direct pathways into employment including matching with potential employers, support to identify employment opportunities and specifc skills training. On the other hand, others may require support to address non-vocational barriers such as health, education and housing to increase their capacity to work or study in the future (Baxter et al., 2021).

#### **Structure of the Book**

All editors of the book, and most of the chapter authors, are connected to the Australian Research Council Centre of Excellence for Children and Families over the Life Course (Life Course Centre). The Life Course Centre is a multidisciplinary collaborative research centre investigating the transmission of social disadvantage over the life course and across generations. The editors and authors come from a range of disciplines and backgrounds including sociology, psychology, economics, epidemiology, neuropsychology and social policy, brought together through our involvement in the centre and its ongoing research and policy agenda.

The volume commenced with the idea of bringing together the wealth of research on family dynamics and social disadvantage generated by researchers in the Life Course Centre into a single collection that sheds light on the ways in which families both generate and mitigate social disadvantage. The original idea was to distil key fndings into a collection that showcased research highlights from the Centre. As we discussed the structure and format of the book and approached chapter authors about key areas, the volume quickly morphed into a collection of predominantly new research on key life course stages and transitions and social disadvantage.

In all cases we asked authors to highlight how a life course approach informs their research approach and results, the implications for understanding outcomes for socially diverse groups, and the policy implications of their work. Although we expect this book will primarily be read by colleagues in academia, we hope that our research is also useful for those working outside the academy to develop and deliver policy solutions aimed at reducing social disadvantage. Each chapter presents an overview of the latest evidence and issues, uses high quality, recent, data to examine key processes and outcomes, and discusses implications of the fndings for public policies and programs designed to mitigate social and economic disadvantage.

The book organises chapters according to key stages of the life course—the early years, adolescence and young adulthood, adulthood and later life. We commence with an overview of life course theory and discussion of recent advances and new directions in life course theory in Chap. 2. Here we outline the main principles of a life course approach, examine differences across disciplines and discuss new advances and directions in life course theory.

The next three chapters focus on the early years of the life course. In Chap. 3, Zubrick, Taylor, Christensen and Hancock use rich national longitudinal data from the *Longitudinal Study of Australian Children* to examine vocabulary growth, literacy and subsequent school attendance in children aged 4 years and upwards. Their results show surprising variability among children in their early vocabulary and literacy growth which in turn lead to substantial gaps in growth and development at later years. Moreover, the circumstances driving these developmental variations vary widely across groups, highlighting the importance of targeted interventions matched to places and circumstances. The authors conclude that while universal interventions for all children are appropriate in the early years, more targeted programs that address specifc barriers will be more effective at older ages. Further, fairness, equity and empowerment of individuals, families and communities must be at the heart of services in order to ensure optimal engagement with, and outcomes from, services. These conclusions resonate strongly with many of the chapters in the book, and particularly those examining educational outcomes.

In Chap. 4, Fatima, Cleary, King, Solomon, McDaid, Hasan, Mamun and Baxter examine the importance of cultural identity for the developmental outcomes of Aboriginal and Torres Strait Islander children. This research uses data from a unique study, the *Longitudinal Study of Indigenous Children*, designed to assess the issues facing Indigenous children, and their families and communities and to provide research and policy evidence to improve opportunities and outcomes. The authors explore social and emotional wellbeing in Indigenous Australian children and investigate whether cultural identity protects against social-emotional problems. The results clearly show that Indigenous children with strong cultural identity and knowledge are less likely to experience social and emotional problems than their counterparts. While Indigenous children face many of the same circumstances and issues challenging developmental growth of non-Indigenous children, they are also subject to additional challenges brought about by white colonisation, loss of sovereignty and dispossession of lands. This chapter contributes to growing recognition of the role of cultural identity in promoting strong health and social outcomes of

Indigenous Australians and further evidence to support the change from a defcit narrative to a strengths-based discourse for improved health and wellbeing of Indigenous Australian children.

Chapter 5 turns our attention to children of humanitarian migrants (refugees). These children come to Australia under vastly different circumstances than migrants entering the country under other types of visas. Forced migration and displacement due to violence, persecution, or natural disasters, combined with often perilous journeys and time spent in refugee camps and processing centres, mean that these children face enormous challenges in integrating into their new country. Lee and Cheung draw on data from *Building a New Life in Australia*, a longitudinal study of humanitarian migrants in Australia, to examine the origins and premigration experiences of refugee children, the structure of refugee families and households and variations in adaptation and language by national origin and gender of the children. Their results show that about half of the sample of children have post-traumatic stress disorder, with refugee girls faring worse than boys. Girls also have poorer self-rated health but do better than boys on standard measures of strengths and diffculties. There are also important differences in outcomes according to national origin with children from Central Asia, Iraq and Afghanistan faring poorly on health, psycho-social measures and English literacy compared to children from other regions suggesting that prolonged confict in countries of origin have a large impact on outcomes for refugee children. Lee and Cheung conclude that while policies and programmes around housing, healthcare, education, employment and language programs need to be comprehensive to enable refugee children to transition to full enjoyment of all opportunities afforded in Australian society, targeted policies are also required to address the challenges faced by children from specifc countries.

The next three chapters examine outcomes for adolescents, young adults transitioning from education to employment and emerging adolescents. In Chap. 6, Povey, Plage, Huang, Gramotnev, Cook, Austerberry and Western examine parental engagement in adolescent education using data from the *Longitudinal Study of Australian Children* as well as material from in-depth interviews from a recent study investigating the impacts of COVID-19 on education outcomes of highly disadvantaged students. They show that parent engagement with schooling has a positive effect on students' outcomes such as self-concept, mental health and aspirations in early and middle adolescence, even when accounting for family and school context factors. Further, parent engagement in late adolescence, with students from highly disadvantaged backgrounds, continues to be important for positive student outcomes.

Tomaszewski, Perales, Xiang and Kubler focus on young people's journeys into and out of university in Chap. 7. They investigate differences in pathways through the higher education system for individuals from low and high socio-economic status backgrounds using three fagship Australian datasets, the *Households, Income and Labour Dynamics in Australia* survey, the *Longitudinal Survey of Australian Youth* and the *Australian Census Longitudinal Dataset*. They show large gaps in access to university as measured by enrolment by socio-economic status with barriers manifesting among low socio-economic status students at the age of 15, compared to their high socio-economic status peers. At the participation stage during ages 18–22 years, there is evidence of barriers at the individual, area and school levels. Finally, their results show the effects of socio-economic background extend beyond graduation and although there is evidence of a 'catch up' effect for some groups, it can take them several years. Tomaszewski and colleagues argue for early multi-pronged and targeted interventions, well before the age of 15, in order to align educational opportunities and trajectories for low and high socio-economic status students.

Ribar and Wong examine emerging adulthood, the period where young people are legal adults, but without the full responsibilities and autonomy of independent adults. For Chap. 8, Ribar and Wong use data from the *Households, Income, and Labour Dynamics in Australia* Survey and defne emerging adults as those who are either co-residing with their parents or receiving direct fnancial transfers. They fnd that the proportion of 18- to 25-year-old Australians who are emerging adults has increased over time with two-thirds engaged in education. Although most emerging adults are in employment, the majority are on casual contracts (short-term positions with no leave entitlements) with low average earnings and low fnancial resources. The majority attain independence by about age 22, but almost 10% return to the family home in the year after leaving, indicating the fragility of early independence for some groups. Ribar and Wong's analyses provide a rich picture of this new life course stage with their analyses highlighting signifcant variability in life course pathways for different social groups.

Chapters 9, 10 and 11 turn to life course experiences in adulthood including the challenges of combining work and family, entering long-term intimate partnerships and becoming parents. Broadway and Kalb show that the percentage of men and women in full-time employment has decreased in Australia since the 1970s, and particularly for men. This trend, combined with increases in the availability of parental leave and greater opportunities for other kinds of fexible work, including working from home, should lead to better work-family balance for both men and women. But the evidence presented in Chap. 9 shows that the burden of care work still falls disproportionately on women. Not only does this lead to high stress and short-term inequitable outcomes for women, it also presents challenges later in the life course. The consequences include lack of retirement savings and a signifcant poverty risk for women in older ages. Furthermore, Broadway and Kalb show the impacts of care on employment are often more evident for low-skilled women, which in turn jeopardises children's health, education, career aspirations and future earnings prospects, potentially deepening a cycle of entrenched disadvantage and intergenerational transmission of poverty. Broadway and Kalb provide a convincing case that policies to improve work-family balance will have a multitude of positive benefts, not just for families but also society more broadly.

In Chap. 10, Lee and Baxter examine differences in who gets married, who cohabits and who stays single and the consequences for fnancial and health outcomes using data from the Households, Income and Labour Dynamics in Australia survey. Like emerging adulthood, cohabitation has developed as a typical new life

course stage, with over 80% of Australian couples cohabiting for a period prior to marriage, and about 18% cohabiting at any one time (Qu, 2020). The percentage of people staying single is also increasing (Qu, 2020). Lee and Baxter fnd that married individuals are more advantaged than their cohabiting and single counterparts across several measures of socioeconomic wellbeing. Married individuals are more likely to be employed, own a home, and have access to emergency funds. Lee and Baxter also investigate outcomes across different partnership statuses to investigate whether marriage selects more advantaged people or has protective effects. There is evidence for both selection and protection effects in these trends, with marriage both selecting more advantaged individuals and providing more protective effects than cohabiting or remaining single. Overall, the analyses show that those who transition from single to married have the strongest mental health and fnancial wellbeing while those who do not partner have the worst outcomes. These differences persist over time and are consistent for men and women, though men have higher wellbeing, employment, and fnancial security than women.

Chapter 11 focuses on parenting, arguably one of the most signifcant life course transitions affecting outcomes for both adults and children. Burke and Dittman explain the importance of individual, family, community and policy contexts in shaping parenting practices and highlight the impact of social disadvantage on parenting styles. They detail the complex interplay between the contextual and individual factors that shape parenting experiences and behaviours and whether aspects of parenting and disadvantage may be passed from one generation to the next. Parental support programs can be very effective in improving parenting practices, but often these programs are delivered to specifc groups, ignoring evidence that such programs can be very effective for all parents and are an important means of interrupting the intergenerational transmission of social disadvantage. Government funding of population level programs is critical and Burke and Dittman advocate for expansion of such programs to parenting at other stages of the life course, such as during emerging adulthood and grandparenthood.

Family dynamics across generations amongst lesbian, gay and bisexual (LGB) people is the focus of Chap. 12. Campbell and Perales investigate relationships between LGB people and their parents and whether the quality of these relationships affect parenting aspirations later in the life course. They show that despite Australian society becoming less heterosexist, a minority of LGB people surveyed in the *Households, Income and Labour Dynamics in Australia* study still experience negative relationships with their parents. The data show that these disparities emerge early, with a turning point evident at age 11, and they persist across the life course. They also fnd that that the less satisfed people are with their relationships with their parents, the less likely they are to desire to have children of their own. Specifcally, gay men who express parenting desires were substantially less likely to expect to fulfll those aspirations if satisfaction with their relationships with their own parents was low. Campbell and Perales propose that internalised homophobia, negative self-concept, stigma and low social support may be mechanisms underlying this fnding. They conclude that equality of outcomes for LGB people must include support and education for parents, as well as continuing to challenge heteronormative social structures.

In Chap. 13 Lam, Dickson and Baxter shift our attention to the later stages of the life course when adults are potentially at greater risk of social isolation and loneliness. At a time in the life course when individuals may have both a strong need for social connections and support and more time to spend with friends and family, older adults may fnd themselves more isolated and disconnected due to retirement, fewer resources and declining health and mobility. The degree to which people feel connected and supported has been shown to be related to physical and mental health, as well as morbidity and life expectancy (Hawkley & Cacioppo, 2010; Luo et al., 2012). Lam, Dickson and Baxter use qualitative interview data with 50 older adults living in the community to highlight variations in experiences of loneliness. The results show that those who maintain strong social ties over the adult life course report less loneliness, but events such as the loss of a spouse, or relocation are disruptors that can render individuals more susceptible to later-life loneliness. There is also evidence here of the long reach of childhood events and experiences in shaping feelings of loneliness and social connection, pointing to the complex ways in which loneliness is shaped by both contemporary and temporally distant experiences.

There is increasing emphasis in Australia on research engagement and translation of fndings to address current social problems and issues. Translating fndings from academia to public policy is not always straightforward. Researchers typically design projects around academic interests and in relation to available data and funding scheme priorities, while policymakers normally require more focussed research on a specifc issue and with much shorter time-horizons than is usual for academic research. In Chap. 14, Reddel, Hand and Lata take up these issues, outlining some of the history and challenges in translating research on family dynamics and social disadvantage to Australian social policy. Their chapter canvases the way research has infuenced key Australian social policies such as the design of the social security system, childcare policies, work-family balance policies, gender equality strategies and support for vulnerable families. Reddel, Hand and Lata conclude by proposing a comprehensive set of criteria to support research and policy impact which would help to reform ad hoc approaches to research and policy partnering and impact and improve outcomes for families over the life course.

Finally, in Chap. 15 we refect on the key fndings from the book, discuss emerging directions and trends in life course research and outline what we believe are some of the key challenges still to be addressed. We argue that life course theory provides a powerful framework for showing that social disadvantage accumulates across the life course and that social institutions shape outcomes and variations across social groups. But we suggest further work is needed to understand the specifc mechanisms of change leading to the accumulation of disadvantage and how these mechanisms vary across place, space and time. Individuals occupy multiple social spaces and identities at the same time, as suggested by theories of intersectionality, and contexts and circumstances may change rapidly as shown most recently by the COVID-19 pandemic. Theories that explain mechanisms of change must therefore be contextual and supported by evidence from high quality data and must move away from average outcomes to assess experiences of disparate social groups across place and time.

Families are possibly more important than ever. As Australia, like many other countries, navigates its way through the COVID crisis and the many health, social and economic challenges thrown up by the pandemic, it is likely that families will play an increasingly important role in buffering these negative impacts. It is families that we often turn to in times of crisis, and families that are the backstop for the support and resources we need to survive and move forward when other sources of support fail or prove insuffcient. But families are not all equally able to provide these supports. To say that families are important is, of course, only the starting point. The more important task, as social scientists and policy makers, is to unpack how this occurs, under what circumstances families contribute to positive or negative outcomes, how these accumulate or decline over the life course and how they vary across cultures and groups and what supports are needed to bolster family resources. These are the broad goals of this book.

#### **References**


Western, M. (2021). *Try, test and learn evaluation—Final report*. Prepared for Department of Social Service. Institute for Social Science Research—Publications Brisbane, The University of Queensland, Institute for Social Science Research. https://www.dss.gov.au/sites/default/ fles/documents/11\_2021/ttl-evaluation-fnal-report\_0.pdf

Beck, U., & Beck-Gernsheim, E. (1995). *The normal chaos of love*. Polity.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 2 Families, Life Courses and the Intergenerational Transmission of Social Disadvantage in Australia**

#### **Rennie Lee, Jack Lam, Janeen Baxter, Jenny Povey, and Stephen R. Zubrick**

At the most fundamental level, a life course approach means understanding that outcomes for individuals and groups are best explained when we take account of the experiences of previous events and the transitions and pathways followed through life to arrive at a particular point. This is intuitively straightforward. If we are to explain why one child attains high grades and achieves well at school for example, but another does not, it is useful to know something about the life history of each child, their family background and the life events that have potentially shaped their learning outcomes. But a life course approach is much more than simply knowing about pathways and life events. It also means understanding how context and circumstances have shaped the types of pathways available to individuals, the role of important others in opening or blocking pathways, how institutional and organisational settings shape experiences and the importance of timing and sequencing of events and transitions experienced along that pathway (Elder & Giele, 2009; Mayer, 2009). A life course approach therefore seeks to understand the actions, behaviour and experiences of individuals by combining insights about the individual as well as the broader social forces surrounding them.

Compared to some other approaches that either only focus on individual level drivers of outcomes, as is common in some psychological theories, or institutional level factors, as is common in some economic theories, a life course approach is a

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia

e-mail: rennie.lee@uq.edu.au; j.lam@uq.edu.au; j.baxter@uq.edu.au; j.povey@uq.edu.au

S. R. Zubrick

R. Lee (\*) · J. Lam · J. Baxter · J. Povey

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Telethon Kids Institute, University of Western Australia, Perth, WA, Australia e-mail: stephen.zubrick@telethonkids.org.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_2

comprehensive and powerful framework for addressing both individual and societal outcomes and the connections between them (Alwin, 2012). Of course, to effectively undertake a life course approach for understanding social outcomes requires collecting, or otherwise accessing, large amounts of information about individuals, families, places, policies, institutions and historical contexts. This approach will often be time consuming and expensive and potentially beyond the scope of researchers. Nevertheless, life course approaches have become very popular amongst researchers and policy-makers, gained currency across disciplines, and developed alongside increased availability of longitudinal panel and administrative data (Mayer, 2009). Such data sources allow us to investigate how lives unfold over time and identify key milestones, transitions, and outcomes across a wide range of areas. A range of disciplines, including epidemiology, criminology, sociology, economics and psychology, have increasingly adopted life course approaches, or some variation, as part of their conceptual toolbox to explain various outcomes (Alwin, 2012).

This chapter provides an overview of the key elements of a life course approach, explains some of the key differences across disciplines and the strengths and weaknesses of a life course approach for understanding the role of families in the transmission of (dis)advantage. We discuss how a life course approach may help to explain intergenerational disadvantage in Australia including in the early years, during adolescence and adulthood and present recent evidence on intergenerational inequality in Australia. In doing so, we outline some of the conceptual and substantive issues guiding the analyses presented in later chapters on the early years, education, labour markets, marriage, parenthood and ageing, as well as explain how a life course approach might assist to understand outcomes for specifc social groups such as migrants and refugees, Indigenous children and LGBTQ+ groups.

#### **What Is a Life Course Approach?**

#### *What Is the Life Course and Why Should We Study It?*

The life course perspective focuses on understanding life-long human development as embedded in historical social context (Elder et al., 2003). Focusing on "changing lives in changing contexts" (Elder & Shanahan, 2006, p. 667), the life course describes the trajectories of a human life from birth to death, structured and shaped by age-graded social roles, and historical and interpersonal contexts (Katz et al., 2012). In life course research, event trajectories are compared across individuals or groups on the basis of timing, duration, and rates of change (Giele & Elder, 1998). The perspective is widely referenced in multiple disciplines because it unites individual and institutional factors, guiding lines of inquiry that are appropriate to studying increasingly diverse populations, in times of substantial social change (Alwin, 2012; Elder et al., 2003).

#### *The Origins of the Life Course Perspective*

The life course perspective is presented by Elder and Giele (2009) as a research paradigm originating from theoretical developments in the 1960s and growing to great infuence in social science over the ensuing half-century. Following a turn in the 1950s towards 'contextualisation' of research fndings through social history, this theoretical work sought to investigate connections between the signifcant social changes occurring at the time and life patterns under study, considering historical events and cohort differences. At the same time, as populations aged in the United States and Europe, more academic attention was directed towards later life stages, with theories moving toward understanding ageing as a life-long developmental process (Alwin, 2012; Elder et al., 2003). Work on ageing by Bernice Neugarten in the 1950s and 1960s, and Matilda Riley and colleagues in the early 1970s, has been credited with connecting life course research with temporality, by discussing age-defned social roles and the normative dimension of transitions and timing, and the importance of birth cohort to life chances, respectively (Elder & Giele, 2009; Elder et al., 2003). This focus on time and temporal organisation provides the common foundation for cross-disciplinary collaboration and the accumulation of a body of knowledge, which has contributed to the emergence of the paradigm in social research (Giele & Elder, 1998). This research depends on longitudinal methodologies capable of tracking individuals over time. From the early 1970s onwards, the life course perspective became the model for developing national surveys and quantitative social research (Elder & Giele, 2009).

#### *Differing Defnitions in the Life Course Literature*

Any review of the literature shows a broad glossary of 'life' concepts complicated by overlapping and conficting defnitions for key terminology. Among the terminology which require defning are life stages, life cycles, and the life span. The term life stage is underpinned by a conceptualisation that across the life course, individuals age sequentially through a series of roles or states (Alwin, 2012). These life stages, understood to be biologically- and socially defned phases, such as infancy, childhood, adolescence and adulthood, help explain the timing, duration and transitions between age-graded social roles within the life course. Life stages do not have defned boundaries and may vary in quantity and duration according to life circumstances and historical and geographical contexts.

The life cycle, using terminology drawn from the biological sciences, describes the complete sequence of life stages between birth and death (Alwin, 2012). The concept is widely regarded to have limited analytical utility in the social sciences, because it arguably implies that this sequence is universal and fxed, failing to refect the socially-defned nature of life stage transitions and the increased prevalence of divergent life course trajectories, such as in reproductive decisions (Alwin, 2012). However, Hogan (2001) suggests that because the life cycle describes the framework for age-graded social roles, it sets the standard which then gives meaning to life course studies of transitions and timings. The life cycle concept is also suggested to be useful for understanding intergenerational processes of socialisation, and processes of societal change as cohorts successively replace each other over time (Alwin, 2012; Elder & Giele, 2009).

Life span is used in general terms to refer to the extent of a life, which for life course research sets the scope for inquiry, in which the life course is nested as the life trajectory of transitions and timings (Elder et al., 2003). However, in psychology, the life span theory of human development defnes aging as a life-long process of within-person change over time, adapting to individual and environmental contexts (Alwin, 2012; Oris et al., 2009). This defnition bears strong similarities with the considerations of life course research. Interdisciplinary work in the early 2000s by scholars from life course sociology and life span psychology initially aimed to link the two concepts, but has since moved into identifying where they diverge, as the life span developmental perspective is primarily interested in changes to functional capacities and behavioural adaptation, looking at the resources available to individuals, and self-regulatory strategies, whereas life course sociology is interested in institutional infuences on diverging life course trajectories (Alwin, 2012; Oris et al., 2009; Mayer, 2009). In Elder's theorisation about the life course, he integrates life span by emphasizing that human development and aging are lifelong, continuous processes.

#### *Paradigmatic Principles*

Elder outlines four elements to the life course paradigm, which can be understood in a 'hierarchy of generality'—location, linked lives, human agency, and timing. The frst, historical and geographical location, highlights that social and historical events affect lives through providing specifc opportunities and constraints. This is typically provided by studying birth cohorts, which use a historical defnition of age by birth year (Elder & Giele, 2009). The second element, linked lives, underscores how lives are lived in relation to other people and infuenced by them, and that these relationships, particularly kinship ties, can be enduring across historical events (Katz et al., 2012). Signifcant new relationships, such as formed through long-term partnering or marriage, can cause shifts in the composition of social ties. The third element is human agency, which accounts for the variations in life courses within given constraints, as humans 'planfully' construct their own lives. Elder and Giele (2009) describe the discrepancies between individual lives and the age-graded life course as a 'loose-coupling'. Finally, the fourth element is the timing and sequence of life events that infuence the life course. The antecedents and consequences of a life transition will be different depending on the event's timing (Heinz & Marshall, 2003). The impacts of the sequence of events in a life course will also depend upon social and normative expectations and the decisions that individuals make in response.

#### **Life Course Approaches and Intergenerational Inequality in Australia**

A life course approach is particularly useful for understanding the intergenerational transmission of inequality. Within social science there is a long and strong tradition of research examining intergenerational transmission of opportunities and outcomes showing that disadvantage accrues over generations and that family background plays a key part in shaping outcomes for individuals (Blau & Duncan, 1967; Bowles et al., 2005; Duncan & Brooks-Gunn, 1997). While debates continue about how much intergenerational inequality exists in different countries, how to measure it and whether more unequal societies have higher or lower levels of social mobility (Corak, 2006, 2013), there is general agreement that family background is important for understanding social disadvantage, including across multiple generations (Bowles et al., 2005).

In Australia, there is strong evidence of intergenerational transmission of inequality. For instance, Huang et al. (2016, p. 373) using panel data from the Households, Income and Longitudinal Dynamics in Australia study (HILDA) found intergenerational earnings elasticity, a measure of the extent to which parental earnings determine their children's earnings, of approximately 24–28%. This broadly accords with earlier studies, although depending on the methodological approach and the data used, measures of earnings elasticity in Australia vary between about 20% and 35% (Leigh, 2013; Mendolia & Siminski, 2016). In recent work, Deutscher and Mazumder (2019) estimate intergenerational mobility using income tax data from 1991 to 2015. They report an elasticity measure of around 18%, signifcantly lower than estimates based on survey data. Importantly they also show meaningful variations across geographical locations both across the country and within cities.

Intergenerational inequality is not only transmitted via earnings. Other work in Australia has focused on intergenerational transmission of welfare receipt with a systematic review in 2014 identifying about 30 empirical studies on this topic in Australia, substantially fewer than undertaken in many other western countries (Perales et al. 2014). The most recent work in this area shows that children of parents who receive welfare payments are almost twice as likely to be on welfare payments when they are adults compared to their counterparts (Cobb-Clark, 2019; Cobb-Clark et al., 2017). Moreover, there is evidence that children of welfare recipients need more intensive support and for longer time periods. The intergenerational correlation was particularly strong for those on disability payments, payments for those with caring responsibilities, and parenting payments for single parents. Overall, this suggests that parental disadvantage may be more harmful to children's later life outcomes if it is more strongly driven by circumstances rather than personal choice. In later work Bubonya and Cobb-Clark (2021) unpack some of the mechanisms linking parental and offspring welfare receipt correlations. They argue that the primary mechanism is the failure to complete high school. Adolescents in welfare-reliant families experience more disruptions in their schooling and less fnancial support from their families leading to lower educational attainment. Risk-taking behaviour is also identifed as a key mechanism underpinning intergenerational welfare reliance.

Other research has focused on the transmission of educational (dis)advantage from grandparents and parents to children (Hancock et al., 2018), the transmission of health outcomes across generations (Huang, 2020; James et al., 2020), the transmission of attitudes (Perales et al., 2021), joblessness (Curry et al., 2019), and the transmission of wealth (Lersch & Baxter, 2020). In sum, there is clear evidence of intergenerational associations which show that the playing feld is not level for all children and that if we are to fully understand and address social and economic disadvantage in Australia, we need to understand not only the life course journeys of individuals, but also their family background.

#### **Family Background and Intergenerational Inequality**

Families can reproduce advantage and disadvantage by directly transferring genetic, social, economic, and cultural resources between parents and children, as well as by indirectly infuencing life choices and pathways through shaping opportunities, experiences and orientations. While advantages and disadvantages can be passed from parents to children, the intergenerational inheritance of disadvantage also has particular consequences for communities, and governments. Over time, it contributes to enduring differences between population sub-groups, and the entrenchment of deep and persistent disadvantage. In this section we consider how family dynamics are associated with the transmission of inequality focusing specifcally on family type, parenting time, parent employment characteristics (Lam et al., 2018), transmission of norms, values, orientations and resources (Salimiha et al., 2018). We present some recent fndings from Australian research on intergenerational transmission of inequality in Australia. The research underscores how families matter in terms of children's outcomes over the life course.

#### *Family Type*

In Australia, non-traditional families are on the rise with an increase in the proportion of children in shared residence arrangements (Nielsen, 2014). Therefore, understanding the implications of varying family types on children's outcomes is of growing importance as children in non-traditional families, such as one-parent, blended, and step-families, show a higher prevalence of mental disorders, lower levels of cognitive ability, and lower academic outcomes than children in original families (Carlson & Cocoran, 2001; Lucas et al., 2013; Perales et al., 2016, 2017). The causal direction of this relationship is debated. Specifcally, there is debate about whether family type has an independent effect on children's socioemotional and behavioural outcomes (Carlson & Corcoran, 2001; Fomby & Cherlin, 2007; Pearce et al., 2013) or whether the socioemotional and behavioural challenges often associated with mental health disorders lead to parental stress and family breakdown (Wymbs et al., 2008).

Regardless of the causal pathways, family type remains an important consideration in understanding intergenerational inequality as adults continue to experience these effects of family type growing up. To illustrate, Bernardi et al. (2019a) and Lersch and Baxter (2020) found that individuals whose parents separate during childhood had less economic wealth as adults. Adult children whose parents separated during childhood had reduced education and earning capacities, unstable family structures, and a lower fnancial planning horizon in adulthood. Additionally, Bernardi et al. (2019a) found that adult children who did not live with their birth parents from ages 0 to 18 experienced a "wealth penalty" throughout their adult lives. A wealth gap started around their early 30s and grew over time. In sum, this body of work shows that childhood family structure continues to shape the outcomes of adult children over their life course. Overall, there is strong evidence that parental separation or divorce is negatively associated with children's skills, education, wages, wellbeing, and own family behaviour as adults (Lam, 2020).

#### *Parenting Time*

Child development is intimately tied to the quantity of resources provided by parents, such as parental time with children, which provides opportunities for children to establish safe and secure relationships with parents. (Lam et al., 2018). In general, parenting time can refer to the time spent parenting by mothers, fathers, or both parents (Amato & Rivera, 1999; Cano et al., 2019; Lam et al., 2018). There are mixed fndings on the effects of parental time on children's outcomes (Cano et al., 2019; Milkie et al., 2015) because mothers and fathers provide different kinds of contributions to children's development (Amato & Rivera, 1999), though it remains unclear whether fathers' and mothers' parenting are conceptually different (Fagan et al., 2014). Nonetheless, we draw on Fagan et al.'s (2014) argument and conceptualization of a more general model of parenting rather than on emphasizing separate conceptualizations of mother's and father's parenting behaviour. Thus, our review will focus broadly on parenting time, without much differentiation across mothers and fathers.

How parents spend time with their children is positively associated with children's socioemotional and behavioural, competencies and non-cognitive skills. Parents are particularly crucial because young children develop a secure attachment with them. Likewise, parents represent a key source for children's acquisition of these traits and skills. For instance, fathers' time with children, often in the form of leisure and play, provides children with important social skills (Paquette, 2004). Additionally, parenting time has important consequences for children's academic outcomes (Philipps, 2011). How parents spend time with their children is not the same and time spent doing educational activities are more important for children's

cognitive outcomes than other activities (O'Flaherty & Baxter, 2019; Cano et al., 2019).

#### *Parent Employment*

Parent's employment is another important factor shaping children's long-term outcomes. However, the effects of parental employment on children's outcomes varies. On one hand, working parents have greater household income (Coley & Lombardi, 2013), better maternal mental health (Roxburgh, 2012), and expose children to formal child care (Gialamas et al., 2014), all of which can improve children's socioemotional outcomes. On the other hand, too much time in employment can have adverse effects such as causing strain on the family in terms of work-family balance and confict (Hsin & Felfe, 2014; Kelly et al., 2014) especially if mothers are working long hours. There is evidence that maternal employment has resulted in increased participation of fathers in child care (Gottfried et al., 2002). Nonetheless, parental employment can also negatively affect children's behaviour by limiting parents' time with children. These factors may negatively affect children's socioemotional outcomes.

In general, Australian studies have shown evidence that maternal employment is associated with children's improved socioemotional outcomes, though the studies vary in their fndings about the magnitude of the effect (Huerta et al., 2011; Lombardi & Coley, 2017; Salimiha et al., 2018). For instance, Hadzic et al. (2013) found that mothers engaged in long work hours showed worse child behaviour, such as higher hyperactivity and inattention, though they found no effect of paternal employment on children's behaviour.

Likewise, there is evidence of an association between parents who work longer hours or have lower job security and poorer child behavioural patterns (Lam et al., 2018). Again however, the evidence is mixed. Lombardi and Coley (2017) focused on maternal employment and found a neutral effect on children's skills, while Huerta found a negative effect on children's cognitive development (Huerta et al., 2011). Lam et al. (2018) found that maternal time but not paternal time was a more important factor in children's behaviour. Father's longer work hours was associated with better child behaviour but children of mothers who worked long hours has poorer behaviours (Lam et al., 2018). One possible reason is that mother's work hours is associated with fewer opportunities for mother-child interactions and mother-child bonding. While mothers are more likely to protect their childcare time than fathers (Bianchi, 2007), the number of hours mothers work will still affect how much time they can spend with their children.

#### *Transmission of Norms, Values, and Resources*

The intergenerational transmission of disadvantage and advantage can occur through the transmission of norms, values, and resources. Parents may transfer socioeconomic resources, such as wealth to their children. This may occur via parents' educational attainment, income, mental health, and occupation for example. Parents' socioeconomic resources may be transmitted to children by shaping children's wellbeing outcomes. There is evidence that children whose parents have lower work hours and earnings during early childhood, are more likely to experience early-adult disease, such as hypertension, arthritis, and limitations on daily activities. Raising the average income by \$5000 over 4-year period would reduce the risk of adult arthritis and hypertension. Overall, this suggests that early low SES environments are associated with immune changes in human children (Ziol-Guest et al., 2012).

In addition to resources, family background can shape children's outcomes via norms and attitudes. For instance, parents who experience periods of unemployment or joblessness can inadvertently yet negatively affect children's attitudes toward work and education (Curry et al., 2019). Children may internalize parents' inconsistent employment patterns, which may subsequently shape their own perceptions about continuous labour market participation or future aspirations. Another instance may be locus of control or the belief that there is a causal relationship between one's own behaviours and the consequences for their lives (Baron & Cobb Clark, 2010). Cobb Clark et al. (2019) found a correlation between parents' selfcontrol and their children, indicating that children who exhibit higher levels of social control also have parents who report higher levels of social control. Additionally, parents transmit gender ideologies to their children, which could contribute to the reproduction of gender inequalities (Perales et al., 2021). In sum, this suggests the importance of looking beyond economic and fnancial resources in the intergenerational transmission of inequality to norms and values that may also determine important life outcomes.

#### **Social Groups and Intergenerational Inequality**

In this section, we examine intergenerational inequality in relation to specifc disadvantaged communities and marginalized social groups. Social disadvantage tends to be concentrated in low-income communities and other disadvantaged social groups, including Indigenous Australians, non-English Speaking Background (NESB) migrants, humanitarian migrants, and single parent households. Below, we review intergenerational inequality across these four groups.

#### *Indigenous Australians*

In general, relative to non-Indigenous populations, Indigenous populations in Australia show large disparities in health outcomes, educational attainment, and other life chance measures (Salmon et al., 2019; Walter et al., 2017). Nonetheless, most Indigenous children show good or excellent health (Anderson et al., 2017). Given the history of colonialism, European settlement, dispossession of lands and discriminatory treatment of Indigenous people in Australia, intergenerational transmission may have less to do with parents and more to do with the trauma and inequities imposed by these broader circumstances. Furthermore, non-Indigenous and Indigenous children differ in several ways, which may affect their outcomes and the role of parental characteristics on these outcomes. For instance, Indigenous children tend to live in larger households (Biddle & Yap, 2010) with more kin and family members. Additionally, around 75% of Indigenous children live in a household that is either a single-parent household, contains adults with lower than Year 12 education, and no employment (Biddle & Yap, 2010). Around 21% of Indigenous children live in a household that shares all three characteristics compared with 5% of non-Indigenous children (Biddle, 2011). Indigenous children also tend to live in households belonging to the lowest income quintiles (Katz & Redmond, 2010) and socioeconomic disadvantage affects early childhood education attendance (Biddle, 2011).

Given these household characteristics, understanding intergenerational changes may require us to look beyond parents. To illustrate, De Bortoli and Thomson (2010) and Trudgett et al. (2017) found that parental SES may not work the same way as it does for non-Indigenous children's schooling outcomes. Therefore, to understand mechanisms that positively shape Indigenous children's schooling outcomes, we may need to look beyond those that are positively associated with non-Indigenous children (Trudgett et al., 2017). These issues are further explored in Chap. 4 where cultural identity and connection is identifed as one of the key protective factors in shaping outcomes for Indigenous children.

Overall, the Indigenous population in Australia is relatively young so understanding their outcomes can further our knowledge about this population and how to successfully transition to adulthood (Biddle, 2011). We acknowledge that while parents play a role, the outcomes of Indigenous children are shaped by various outcomes, including geography, connection to cultural background, and racial discrimination (Lovett, 2017; Thurber et al., 2015). A life course perspective in particular is useful for understanding the intergenerational transmission of inequality among Indigenous children. More broadly, a life course approach that takes a holistic view across time, lives, generations and relationships has been argued to be a useful framework for improving health and social outcomes of other Indigenous communities (Theodore et al., 2019).

#### *Migrant Background Populations*

The effects of migrant background on social outcomes in Australia are typically understood via high-income or English-Speaking Background (ESB) countries or middle- and low-income countries or non-English Speaking Background (NESB) countries. While immigrants from ESB countries typically have outcomes that are on-par or better than the host population, immigrants from NESB countries tend to show greater disadvantages. For instance, NESB migrants tend to show lower levels of labour market outcomes, such as lower incomes (Katz & Redmond, 2010) and lower labour market participation relative to native-born Australians and immigrants from ESB countries (Wilkins, 2008).

There are mixed fndings about whether these differences among immigrant adults are observed among their children. In general, children from NESB immigrant families in Australia still face challenges in their integration with some of this stemming from racial discrimination in school or growing up in a new context away from their connections in their origin country (Katz & Redmond, 2010). To illustrate, children from NESB immigrant households belong to households with lower incomes (Katz & Redmond, 2010) than their ESB counterparts and the overall Australian population. Nonetheless, it is also possible that this is concentrated in certain national origin groups, such as Vietnam and Lebanon, that tend to experience greater poverty and material deprivation. Likewise, children of NESB migrants tend to show lower levels of wellbeing (physical and mental health) and higher rates of obesity than the general population. These are related to lower SES and neighborhood location though these factors do not completely account for these disparities (Katz & Redmond, 2010; Zulfqar et al., 2018).

In contrast, Cobb-Clark and Nguyen (2012) found that children of NESB show educational advantages over their ESB and native-born Australian peers. Additionally, Washbrook et al. (2012) did not fnd any signifcant differences in the development of young children of foreign-born and Australian-born families. In part, this may be explained by the fact that although parents may experience challenges in the labour market, many are still highly educated or arrive from higherincome origin countries (Washbrook et al., 2012). Overall, this may suggest a relatively optimistic outlook for immigrants' children, even those with parents from NESB backgrounds who face more challenges.

#### *Refugees*

Unlike other groups of immigrants, refugees are a particularly vulnerable group. Refugees show low levels of labour market participation and income, limited English profciency, poor mental health, and poor housing conditions (Colic-Peisker & Tilbury, 2007; Fozdar & Hartley, 2014; Hugo, 2014). Like other migrants these patterns tend to vary in relation to country of origin and gender (Perales et al., 2021). Refugee women are particularly disadvantaged and show low rates of paid employment and higher rates of psychological distress (Delaporte & Piracha, 2018; Jarallah & Baxter, 2019). Refugee children face multiple challenges related to family instability, lower health outcomes, and educational attainment. They have been identifed as a high-risk group for poor mental health (Lau et al., 2018; McMichael et al., 2011) and not completing secondary school (Correa-Velez et al., 2017). For instance, Correa-Velez et al. (2017) found that nearly a decade after settlement, nearly 38% of refugee youth in their study had left school prematurely. Likewise, many still continue to experience long-term effects of social exclusion related to discrimination (Correa-Velez et al., 2015).

Nonetheless, much work has noted that refugee youth are still making strides relative to their parents' human capital and socioeconomic characteristics, which may indicate intergenerational progress (Daniel et al., 2020; Zhou & Bankston, 1998). Similarly, despite their challenges, they appear to make improvement with time in Australia. For instance, Ziaian et al. (2013) found large improvements in the socioemotional wellbeing of refugee children and adolescents within a few years of resettlement. Overall, while refugee children still face challenges in their integration, understanding their starting points is an important reference point. More work and longitudinal data are needed to understand their outcomes as they transition to adulthood to assess the extent to which refugee youth are making progress relative to their parents.

#### *Single Parent Households*

Individuals in single parent households experience high rates of mental illness, shorter lifespan, poorer physical health, and food insuffciency (see review in Jovanovski & Cook, 2020). There is also evidence that the fnancial situations of single-parent households are becoming more dire with increases in relative poverty rates. In 2016, 15% of single-parent households were in poverty compared with 25% in 2018 (Broadway & Vera-Toscano, 2020).

Single parent households are showing decreases in their use of formal child care. In 2016, 52% of single-parent households used formal care compared with only 35% in 2018 (Broadway & Vera-Toscano, 2020). Although the exact mechanism is unclear, single parent households may stop using child care due to their lower income and high cost of child care. For many single parents in low-income employment, the take home income after child care does not justify employment. However, this may constrain their labour market opportunities if they do not have reliable child care.

#### **New Developments in Life Course Theory**

Life course theory is useful for explaining the intergenerational patterns reviewed above. The origins of life course theory date back several decades, but commentators have noted a growth in references to the term life course in academic papers throughout the 1980s and 1990 and a surge in the 2000s and 2010s (Bernardi et al., 2019b). Furthermore, recent growth in academic journals and research centres with the term *life course* in their title has also been remarked upon (Mayer, 2009; Bernardi et al., 2019b). In a 2006 *Science* paper, Butz and Torrey identify longitudinal surveys as the "Hubble telescope" of the social sciences arguing that the "fundamental challenge in the social sciences is moving from complicated correlations to useful prediction" (2006, p. 1898), Longitudinal surveys enabling consideration of life events and transitions provides a major step forward in moving social science from description to prediction.

While few would dispute the wealth of important new information provided by longitudinal surveys, other accounts of the potential of life course and longitudinal approaches are more measured (Macmillan & Hannan, 2020; Mayer, 2009). Ulrich Mayer, one of the founding fathers of life course theory, argues that the feld has been highly successful in several areas—instilling a life course approach and new data and methods across disciplines, examining cross-national variations in institutional contexts on life courses, assessing the impact of sudden societal change on life courses and investigating the relationship between health and life courses (2009). But he also identifes areas where important progress in life course research is yet to be made, including identifying causal linkages across human lives, the interaction of psychological processes of development and socially embedded life courses and the development of an integrated overarching life course theory as opposed to a set of concepts and heuristics (Mayer, 2009).

In a related vein, Macmillan and Hannan (2020) argue that life course research has failed to make the most of opportunities to exploit natural experiments to develop causal explanations. They argue that by identifying important moments of historical change, variations in social structures, or policy changes it is possible for life course research to credibly develop causal arguments about social experience and human development. The assumption of such work is that groups of people are indistinguishable prior to a naturally occurring event or social change, that they are randomly affected by the event and sorted into identifable treatment and control groups enabling identifcation of mechanisms of cause and effect. Macmillan and Hannan urge life course researchers to take up opportunities for these kinds of natural feld experiments and to pursue causal theories and evidence.

Arguably one of the most innovative developments in life course theory in recent years is the work of Bernardi et al. (2019b) who have taken up the challenge of developing an overarching integrated life course theory. They defne the life course as a multifaceted process of individual behaviour and develop a theory aimed at explaining the nonlinear dynamics of individual behaviour over time and across multiple dimensions. These dimensions are defned as inner-individual (e.g., genetics, biology, physiology), individual (e.g., education, social status, gender, citizenship), and supra-individual (e.g., socio-cultural environment ranging from networks and relationships to country institutional features). Importantly life course theory must be dynamic they argue, recognising the agency of individuals who make bounded or constrained choices based on their beliefs, experience and expectations of their actions.

They use the heuristic device of a *life course cube* to outline the complex interdependencies across these dimensions that explain contemporary life courses. The core axes of the cube are time, domains and levels. Time refers to the history of a life course, current experiences and future outcomes. Domains cover areas such as work, family, education, leisure that are interdependent and overlapping. Levels denotes the interdependence that connects individual agency with the life courses of other people, the external societal structures and the inner-individual dispositions and orientations. They argue that the life course cube offers a parsimonious heuristic device for integrating various approaches to life course research across disciplines. They recognise that not all life course research will address all elements of the cube, but that it serves as a reminder of the full complexity of interdependencies, levels, domains and dimensions that underly life course processes and outcomes.

In more recent work Sanchez-Mira and Bernardi (2021) have further argued for a more developed theoretical conceptualisation of time in life course research that goes beyond notions of time as absolute (linear, chronological and uniform) and incorporates understanding of relative time. Time has been a central concept in life course approaches but is often understood as linear and unidirectional following a chronological clock and calendar at a uniform pace. In contrast, Sanchez-Mira and Bernardi (2021) argue for a concept of relative time that is *multidirectional* with the remembered past and anticipated future infuencing agency and decision-making in various ways, *telescopic* with individuals acting on the basis of differing time horizons where they zoom in and out of past and future events in ways that shape their actions, and *elastic* where individuals do not experience time as continuous, uniform or linear, but rather at different tempos and paces and distorted in various ways through subjective perceptions. Time is thus relative to individuals rather than absolute and an integral part of understanding agency as opposed to an objective measure of cohort, period or ageing.

Overall, this brief review shows the power of life course theory to explain social disadvantage over the life course and across generations. But it also highlights the need for further new developments in life course scholarship to move beyond a set of related concepts to an integrated overarching theoretical framework. There is lively and constructive debate occurring in the literature about these new ideas with scholars pushing the life course framework forward in useful ways (Mayer, 2009, Bernardi et al., 2019b). Importantly these new directions are aimed at informing and aiding more powerful empirical studies with implications for study designs, measurement, methodology and interpretation. And there are continuing moves to collect, access and analyse longitudinal data that enables a life course approach to understanding a wide range of social outcomes and to effectively inform policy directions.

#### **References**


Giele, J. Z., & Elder, G. H. (1998). *Methods of life course research*. SAGE.


Leigh, A. (2013). *Battlers and billionaires: The story of inequality in Australia*. Black Inc.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 3 Early Years and Disadvantage: Matching Developmental Circumstances in Populations to Prevention and Intervention Opportunities**

#### **Stephen R. Zubrick, Catherine Taylor, Daniel Christensen, and Kirsten Hancock**

At the outset of the twenty-frst century, Australian social and economic circumstances have prompted a relentlessly increasing demand for, and supply of, a range of Early Childhood Education and Care (ECEC) services as parents—particularly women—have moved into the labour force to support their family needs and protect their own futures. This demand has produced a decade of volatility in the number, design, and provision of ECEC services that seek to improve the development of infants and children. The distribution, cost and regulation of the standards and quality of ECEC services have been the subject of intense debate and scrutiny. Along with these developments, and because of the demand for such services at everearlier points in child development, there has been a rising focus upon the instrumental role that early years opportunities play in establishing and advancing developmental capacities important for onward learning achievement—particularly those that concern cognition and social skills.

Accompanying all of this has been a burgeoning scientifc evidence base linking early infant and childcare programs, and the nature of the opportunities and expectations that they provide, to the subsequent capabilities of children to achieve more optimal outcomes in education and onward into adult life. Consistent with a life course approach, early infant and childhood programs, typically in the epoch from birth to the commencement of formal schooling, have been identifed as one of the key mechanisms to break the intergenerational cycle of deep and persistent disadvantage, and to reduce inequalities in educational outcomes across socio-economic groups (Heckman, 2006). For example, foundational research such as the Perry/

S. R. Zubrick (\*) · C. Taylor · D. Christensen · K. Hancock

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Telethon Kids Institute, University of Western Australia, Perth, WA, Australia e-mail: stephen.zubrick@telethonkids.org.au; catherine.taylor@telethonkids.org.au; daniel.christensen@telethonkids.org.au; kirsten.hancock@telethonkids.org.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_3

High Scope and Abecedarian studies has demonstrated the long-term benefts of early childhood education programs for children from disadvantaged backgrounds (Campbell et al., 2012; Schweinhart, 2013). Such fndings have prompted great enthusiasm among the Early Years sector.

And yet, amid the impressive array of scientifc evidence about the positive impact and role of early experiences on subsequent neurobiological (Shonkoff, 2016), cognitive (Heckman, 2006) and non-cognitive maturation (Gutman & Schoon, 2014; Niles et al., 2006) in typically and atypically developing children, the science on the *delivery* of programs that actually change the level and rate of development and growth in infants and young children is far more sobering. In their meta-analysis of 84 carefully controlled experiments (i.e., "trials") of early infant and child interventions that met minimum scientifc standards spanning the period from 1955 to 2010, Duncan and Magnuson (2013) confrmed that such programs were indeed effcacious across a range of outcomes. However, the authors were explicit in noting that the average weighted effect size across all programs was small (0.21, measured in standard deviation units). Duncan and Magnuson also noted that programs designed by researchers imparted larger effects (0.39) than did programs designed by non-researchers (0.18) and that programs before 1980 produced signifcantly larger effect sizes (0.33) than did those that began after 1980 (0.16). This decline of program effect sizes over time likely refects the fact that children in the control groups were increasingly being exposed to some centrebased care and were also benefting from improved home environments (notably increases in maternal education) relative to those children in the control groups of earlier epochs.

In Australia, the ECEC sector provides a range of child care and preschool services. Families partake of ECEC via a mixture of informal care (grandparents, other relatives, friends, neighbours or nannies) and formal care (provided predominantly through long day care and outside-school-hours care) and preschool programs which are structured play-based learning programs delivered by early childhood teachers. The proportions of these types of care exhibit age-dependent variations: Typically, prior to the age of 1 year, parents use more informal care while from ages beyond 1 year increasing amounts of formal care are used (Baxter 2015; Baxter & Hand, 2013). Children may be exposed, at any age, to both informal and formal care. In 2014, among children aged 0–4 years, 22% used formal care only, 18% used informal care only, 12% used both types of care with the remaining 45% of children not attending child care (Australian Bureau of Statistics, 2014). Long day care was the main source of formal ECEC to children under school age, with its use peaking at age 2–3 years although it is notable that about 30% of 1-year-olds and 4-year-olds were also in long day care (Baxter, 2015).

Not all families use ECEC and where they do, there are striking differences in the mix and use of ECEC by families that have differing structures and capabilities. For example, single parent families are early and heavy users of formal care: Where single parents are employed, 65% of their 0–2 year-old children are in formal ECEC (Baxter, 2015). This contrasts to 49% of children in couple families where both parents are employed. These proportions drop to 23% where a sole parent is not employed and to about 14% where one of the parents is not employed in a couple family. From the age of 3 years onwards the use of ECEC by parents markedly interacts with offers of, and expectations about, enrolment in education in the period before children are expected to commence full-time in Year 1. In dual employed families, 32% of children were in preschool, while the percentage was lower (23%) in families of employed single parents. In combination, the variation in effect sizes between different services and the different rates of enrolment of children in ECEC by families with differing structures and capabilities, is seen to either deepen existing, or impart onward developmental inequalities in outcomes among specifc groups of children. These inequalities extend also to the quality of ECEC received by some children in relation to others.

ECEC classrooms within the lowest socioeconomic status (SES) areas have been found to have lower levels of instructional support provided to children than those in the highest SES areas (Tayler, 2016). Such differences accumulated across prekinder and kinder programs (approx. commencing ages 4–5 years) and led to children being approximately 3.3–4.9 months behind their peers in more advantaged neighbourhoods (on measures of verbal ability). These differences in the quality of, and child growth responses to, early childhood programs, suggest that the pragmatic considerations of how well such opportunities are arranged and resourced to deliver developmental advantage to children is of substantive importance. The enthusiasm for these opportunities is evident by their growing promotion and use. The promise of improved life chances for children who participate in them has enormous popular appeal and is policy relevant in societal settings that seek more equal outcomes for all children. These programs are designed with the intent to change the level and rate of developmental growth in participating children. However, the empirical data on the actual starting levels and onward rates of developmental growth of children are strikingly absent in contemporary research fndings. What is available suggests caution in enthusiasm for the size of growth effects imparted to children by such programs, the uniformity of these effects across different groups of children and the need for greater precision, reach, and proportionality in the delivery of these programs.

With these observations as our motive, we add to the literature on the delivery of early years opportunities by describing typical growth patterns in very young children. We go on to estimate the extent to which developmental circumstances accelerate or retard changes in early cognitive growth and later literacy and numeracy. Finally, we assess the implications of these fndings on the subsequent population reach and actual participation of children in these programs particularly regarding the very groups often in the target range for program beneft.

#### **Measuring the Developmental Growth of Children Is Not Easy**

The defning feature of infancy and childhood is growth, and longitudinal studies are a mainstay for evidence about growth in early life. Yet, many contemporary longitudinal studies are extremely challenged to meet the requirements of consistent, repeated measurement of growth characteristics that permit appropriate growth modelling (Zubrick, 2016). Moreover, growth phenomena are not easy to capture during periods when children are rapidly changing.

Some measures, such as weight, height, girth, head circumference, muscle mass and strength lend themselves more easily to repeated measures with the same metrics over long periods of time. Other growth characteristics are considerably harder to repeatedly measure—some cognitive abilities including reading and numeracy, and time-based activity measures appear in longitudinal studies and may do so with enough consistent repetition to allow growth modelling. In contrast though, growth of subjective decision making, risk aversion, emotional regulation and introspection are examples of important developmental characteristics for which growth measures are either absent or studies using them over time, lacking.

In this chapter we use early childhood growth in receptive vocabulary to observe important aspects of growth and its variability and to characterise patterns and circumstances of this growth relevant to the design and implementation of developmental opportunities in the early years. Vocabulary is a robust indicator of development across the lifespan (Powell & Diamond, 2012; Vasilyeva & Waterfall, 2011). Comprising the words we understand, receptive vocabulary can be measured from about 8 months of life onwards and thereafter throughout life. Vocabulary growth is rapid, expanding from infancy to about 200 words at age two to 20,000 words at age 8 (Anglin et al., 1993; Fenson et al., 2007). The Peabody Picture Vocabulary Test (PPVT) is one such measure that is well-standardised and can be used repeatedly across the life course (Dunn et al., 1997). Unless otherwise noted, it is used as the principal measure throughout this chapter.

In the fndings we present here, we model the growth of vocabulary from variables principally collected in the *Longitudinal Study of Australian Children* (LSAC) (Taylor et al., 2011). The generalisations we derive from these fndings describe modelled data. These models are best thought of as heuristic, or exploratory, devices to illustrate general principles. Where there are signifcant limitations to our generalisations arising from these models, we describe these.

#### **Risks That Predict Low Language Performance Are Weak**

There is an extensive literature on the selection and modelling of risks and their association with child development outcomes with cumulative risk models being frequently constructed (Evans et al., 2013). A range of risks is examined using statistical techniques to permit them to be added together to calculate a cumulative total. Despite their advantage, and simplicity of interpretation, cumulative risk models are limited in their capacity to capture aspects of duration and timing, nor do they provide insight into underlying, and perhaps more universal, mechanisms that link adversities to outcomes of interest (McLaughlin & Sheridan, 2016). In addition to this, developmental risks not only have the propensity to cluster or aggregate over time as well as spatially, but individuals within the same gender, ethnic category or social class are likely to share common constellations of risks (Kagan, 2018). These constellations, as we will show, impart differing rates of growth.

Language development is truly one of the endowments unique to human beings and so clearly manifest through the interplay of the child's endowments with their environmental circumstances. The clear growth of language would suggest a virtually unimpeded, steady, and "rocket" like trajectory upwards. And yet, systematic studies of language growth return a picture of surprising variability within and between children in the growth of their vocabulary specifcally and language more generally. The initial onset of language development is principally dominated by factors internal to the child (Zubrick et al., 2007) and most children who start late, subsequently catch up (Reilly et al., 2010). This has signifcant implications for those interested in early intervention strategies that seek to promote child development as well as early identifcation of children for specifc treatment.

Table 3.1 shows risks associated with lower language ability at the age of 4 years. A total of 28 candidate predictors was selected from the LSAC data based on empirical evidence. Sixteen of these variables were individually associated with vocabulary growth differences of at least 6 months over a six-year period (Christensen et al., 2014; Taylor et al., 2013). In the multivariable modelling, ten of these variables were associated with lower vocabulary growth.


**Table 3.1** Individual risks associated with lower receptive vocabulary at age 4 (effect size d > 0.30)a

Source: Longitudinal Study of Australian Children

Notes: a Variables are bolded where an effect size of d > 0.30 corresponds to about 6 months of growth. b Bolded entries are predictors retained in the multivariate model

This growth modelling of individual risks also showed that a large set of wellselected and measured predictors explained very little of the variation in children's language growth from 4 to 8 years of age. For most risk variables, adjusted effect sizes were negligible to small. The total amount of variance in vocabulary growth that the risk variables account for was an additional 7% *after* adjusting for advancing age, which accounted for another 52% of the variance in vocabulary development. This 7% is a surprisingly small percentage of increase in variance accounted for by the risk variables over and above that accounted for by the child merely getting older.

#### **Predicting the Children Who Will Need Help Will Miss Many Children**

The fndings here are important because they signal a critical feature about child development: It is characterised by large variability within children and between them. The fndings show that age is the biggest explanatory source of variation related to vocabulary growth. Closer scrutiny of the children's growth reveals additional insights. First, when age is appropriately accounted for, vocabulary growth varies by where children start their growth (Fig. 3.1, upper). Those in the bottom 15th percentile at age 4 display more rapid growth relative to children whose starting positions are higher. Second, when these individual children are observed over this same period they display striking positional movement (Fig. 3.1, lower). Some children progress to higher levels of vocabulary performance but some also display declines in their performance.

What does this mean when we try to predict who needs help? Our predictive models reveal that we are very good at predicting those children who will *not* have low receptive vocabulary status at age 8. Of the children we predicted at age 4 to *not* have low receptive vocabulary at age 8, we will have only misclassifed 5.9% of these who actually go on to have low receptive vocabulary at age 8. This might be deemed a tolerable level of predictive error.

However, we are typically interested in predicting those children who are likely to be developmentally challenged—after all, it is these children we would like to offer programs to improve their early developmental skills. Unfortunately, for the children at age 4 who were predicted to go on to have low receptive vocabulary at age 8, 74.2% of these children actually go on to be classifed as *not* having low receptive vocabulary at age 8. In other words, it is possible that considerable resources will be deployed towards children who, in all likelihood, would not need intervention. Decisions would need to be made as to whether this represents a tolerable use of resources for monitoring and/or intervention to assist the 25.8% of children classifed at age 4 who would otherwise have a low receptive vocabulary status at age 8. In summary, our fndings exhibit low predictive utility (Christensen et al., 2014).

**Fig. 3.1** Vocabulary growth in children as a function of their initial vocabulary at age 4: Average growth (upper) and positional change (lower). (Source: Longitudinal Study of Australian Children)

Early child development is certainly characterised by growth—and some of it, quite rapid. But our fndings and those of others demonstrate that development is also highly variable and characterised by children moving up and down in their skill levels. These features align with the observation that most early development interventions are characterised by small effect sizes. After all, interventions are designed based on assuming or predicting which children will beneft (or not) from early interventions by using predictors identical or similar to the variables in our models. These interventions are "pushing into" growth characteristics of the children that are in and of themselves highly variable. This variability would contribute to weaker effect sizes.

It would be easy to adopt a gloomy view about the benefts and use of early developmental programs based on observations like these. Instead, our view here is to posit that high variability is a basic feature of early child development and that it invites a broader understanding of how policies, funding and practices might be arranged to infuence better outcomes for all children. We turn to this in the following sections.

#### **Moving from Individual Risks to Describing Developmental Circumstances**

The variability in children's vocabulary growth from ages 4 to 8 makes an important transition to reading literacy at age 10 years. This permits an extension of the period over which to observe stability and change in growth characteristics (Fig. 3.2) from age 4 to age 10 (Zubrick et al., 2015).

**Fig. 3.2** Positional movement in language at ages 4, 6, 8 to literacy at age 10. (From Zubrick et al., 2015, doi:10.1371/journal.pone.0135612.g002. Source: Longitudinal Study of Australian Children)

The measures in this model have been standardised to capture children performing at or below the 15 percentile in vocabulary (ages 4–8) and reading (age 10). The children above the 15th percentile are classifed as having middle to high performances. In this series of observations on the same children, there are 24 or 16 possible combinations of movement in and out of "low" to "middle-high" levels of performance. With this, it becomes possible to focus on *patterns* of stability and change as early development progresses. Children can remain stable (i.e., in the same performance category) or they can change category as they develop.

When studied this way, the most prevalent developmental *pattern* was for children to start in the middle-high category and remain so—69% of the children studied were performing in the middle-high group throughout the 6-year period of observation. We refer to this pattern as "Developmentally Enabled".

One of the least common developmental patterns was the stable low pattern: Only 1% of children were persistently low at all age points. There are, of course, intermediate pathways in which there is change in developmental status which results in improving patterns (8%), declining patterns (10%), and fuctuating patterns (12%) of development. These fndings highlight that children move in and out of developmental vulnerability and that the patterns observed here continue to point towards high levels of variability in developmental status through these early years.

Can the sixteen individual exposures to risk for developmental vulnerability that we display in Table 3.1 be used instead to describe clusters or classes of risk circumstances that would better characterise developmental growth, its variability, and how interventions might be considered? While risks to development undoubtedly accumulate the pattern of their association, this does not suggest simple additivity. Different risks accumulate for different children.

The average number of risk exposures at the age of 4 for any one child was about 2.5 risks. 14% of children were exposed to none of the designated risk factors, almost two thirds of children had two or more risk exposures, with 42% of children experiencing three or more risks. These sixteen possible risk factors allowed for 216 (65,536) possible combinations of risk factors. But of these, 1585 combinations were observed in the data.

Figure 3.3 displays a "heat map" of the dichotomised individual risks from Table 3.1 that are associated with low vocabulary growth. This fgure shows how some risks cluster together more strongly than others. We used a statistical technique to identify substantively meaningful clusters or "classes" of risks within which participants have a similar response pattern (Christensen et al., 2017). The results of the latent class analysis distinguished six different groups of risks—or, *developmental circumstances*—that children experience. These circumstances are not mutually exclusive. Our statistical approach instead estimated the probability of children belonging more to one group over another and we assigned children to the group for which they had the largest estimated probability.

The frst developmental circumstance (i.e., the reference group) we described as Developmentally Enabled. This group made up 46% of the sample. On average, each child in this group was exposed to only 1.0 risk at age four. The distinguishing feature of children who were Developmentally Enabled was consistently lower than

**Fig. 3.3** The clustering of risks for low language development at age 4. (See Christensen et al., 2017. Source: Longitudinal Study of Australian Children)

population average proportion for each of the risk factors, with a likelihood of zero risk for teenage motherhood, being in families with four or more siblings, and the study child not being read to at all.

The next developmental circumstance comprised Working Poor (20%) families. On average, each child in this group was exposed to 2.8 risks. Relative to the overall population proportion, this group had a similar proportion of mothers who were unemployed (44%) when the child was 4 years old. But, children in this group were more likely to exhibit low school readiness, have mothers with low education, have four or more siblings, and live in disadvantaged areas. Non-English speaking status families were not in this group.

The third circumstance, which we termed Overwhelmed (10%) was typifed by multiple risk factors across all domains. On average, each child in this group was exposed to 6.1 risks. Relative to both the Developmentally Enabled and the population average, this group had an increased likelihood of all risk factors, other than maternal non-English speaking background.

The fourth circumstance was characterised by a combination of factors that were unique to the individual child and which we termed Developmental Delay (9%). On average, each child in this group was exposed to 3.8 risks. Higher proportions of children in this group had low temperamental persistence (47%) and reactive temperament (40%).

Making up 8% of the sample, the ffth circumstance was typifed by Low Human Capital and can be contrasted to those in the Developmentally Delayed circumstance because of the higher proportion of teenage mothers (9%) and maternal low education (53%) in this group. Maternal unemployment (65%) is very high in this circumstance relative to the population proportion. On average, each child in this group was exposed to 3.8 risks. This group has the highest proportion (97%) of families in the lowest income quintile and has the highest healthcare card use (84%). Importantly, the proportion of children in the Low Human Capital group with low school readiness (14%) is comparable to the overall population average.

Finally, the sixth circumstance we describe as Resource Poor non-English Speaking (7%). On average, each child in this group was exposed to 4.7 risks. This group included 44% of the non-English speaking mothers. It had an increased proportion of mothers with maternal psychological distress (42%), low parenting consistency (35%), four or more siblings (11%), low income (33%), healthcare card (35%), neighbourhood disadvantage (35%), and not reading to the study child (14%). This group did not show any increased likelihood for study child Indigenous status, child reactive temperament, or teenage motherhood.

With these groups now described, we wanted to know if their rates of developmental growth, as measured by vocabulary and literacy differed in initial starting levels and onward rates of growth.

#### **Children in Different Developmental Circumstances Have Different Rates of Growth**

When we studied the rate of vocabulary growth in children from age 4 to age 8, there were striking differences in the rate of growth for children who were in each of these different developmental circumstances (Christensen et al., 2017). The fndings demonstrate that developmental circumstances were associated with both differing initial starting levels of vocabulary ability and then, differing rates of onward growth. Importantly where there were similarities in starting levels of vocabulary and onward growth, there were differences in developmental circumstances that would need different policy prerogatives and intervention approaches in order to deliver better targeted interventions.

Developmentally Enabled children were characterised by "on time" vocabulary growth. That is, at each point of observation at ages 4, 6, and 8, these children's vocabulary performance was, on average, appropriate for their age—they were "on time". This group of children was used to compare the vocabulary growth of children in other developmental circumstances. What this showed was that children in the Working Poor circumstance started 5.8 months behind at age 4 and were almost 5.9 months late, or behind, in their vocabulary growth by the time they were 8 years of age—that is, they fell slightly further behind. Children in the Developmental Delay circumstances started 9.6 months behind and were 7.2 months late by the time they were 8 years of age, and children in the Low Human Capital circumstances were 6.1 months behind at age 4, and 4.7 months late in their vocabulary growth by the time they were 8 years—in other words, children in these groups were slowly catching up.

Children in the remaining groups showed even larger differences. Those children in the Resource Poor Non-English Speaking circumstance started 26.3 months behind at age 4, but had closed this gap to 10.4 months by the time they were 8 years. Children in the Overwhelmed circumstance started 18.9 months behind and were over a year behind—13.1 months—at age 8. As can be seen, each of these groups had different rates of growth with the Non-English Speaking group having the highest (most rapid) rate of growth. It could well be that these children would close this gap were we to observe them over a longer period of time.

How persistent are these patterns and circumstances over time? As noted above, longitudinal data on developmental growth are diffcult to come by. One domain, however, that offers some scope for such study is school—specifcally the growth of reading and numeracy. Australia conducts the National Assessment Program-Literacy and Numeracy (NAPLAN) gathering information on these student competencies in school years (i.e., Grades) 3, 5, 7, and 9 covering the developmental range from approximately age 8 to 15 years. Using NAPLAN data, and a similar composition of ten risk exposures, continued to produce four classes of risk circumstances that represented substantively different developmental circumstances that infuenced the onward growth of NAPLAN literacy (Taylor et al., 2019).

The largest group (62%) was again characterised as Developmentally Enabled and was characterised by a very low exposure to any of the modelled risks. Another 25% were in a circumstance characterised by, principally, Sociodemographic Risks with low maternal education, low family income, indigeneity, and area disadvantage predominating. Smaller proportions of children (11%) were in a Child Development risk profle with low school readiness, lower non-verbal intelligence and vocabulary, and low task attentiveness and behavioural problems. Finally, a very small proportion (2.4%) of children experienced Double Disadvantage with high exposures to risks spanning both Sociodemographic and Child factors.

These groups revealed differences in the growth of their reading competencies over time relative to the Developmentally Enabled group. Students whose circumstances were Developmentally Enabled in Year 3 had average reading achievements in Year 9 benchmarked at 10.4 years—they were ahead of their year level. However, students with a sociodemographic risk profle were 1.2 years behind the reference group in Year 3 and by Year 9, they had fallen 2.1 years behind their developmentally enabled peers. Students with a Child Development Risk profle were 2.0 years behind the reference group in Year 3 and by Year 9, had fallen 3.3 years behind their developmentally enabled peers. Finally, students with a Double Disadvantage risk profle were 2.7 years behind the reference group in Year 3 and by Year 9 were 5.3 years behind their developmentally enabled peers. These are substantial gaps in growth and development.

Comparing developmental effects for vocabulary and for reading competency in two studies from the same birth cohort illustrates that developmental circumstances are associated with marked differences in rates of growth in these capacities, and that the impact of developmental circumstances on school outstrips the effect of developmental circumstances on vocabulary.

#### **Developmental Circumstances Reduce Participation in Interventions**

Developmental opportunities and expectations are—of course—designed and offered with the intention of assisting children in their early development and, particularly, to assist those children who are vulnerable to or falling developmentally behind. Many of these opportunities and expectations take the form of participation in play groups, library programs, and early childhood education and care programs, and entry to, and attendance at school. Participation in these is explicitly designed to encourage optimal development, prevent developmental delay or disadvantage, or close emerging skill gaps in some children relative to others.

What are the relationships between the developmental circumstances we describe here and the likelihood of participating in these opportunities? School is one such opportunity and is notable because it is a socially sanctioned developmental expectation that is typically legislated and mandated for all children. Attending school is expected and school attendance offers one type of measure of exposure to a developmental opportunity that is explicitly organised to change child development.

Using LSAC data, Hancock et al. (2018) identifed four classes of risk exposure (e.g., developmental circumstances). Most children (56%) were exposed to minimal risk, 20% were exposed to parenting, child development, and mental health risks only, 15% were exposed to a greater extent to fnancial risks only, and 9% had a higher probability of exposure to all risks. Hancock et al. (2018) found a great deal of heterogeneity in the association between persistent non-attendance at school and developmental circumstances. One-third of persistent non-attenders were actually in the low-risk circumstance. But, persistently non-attending children were eight times more likely to face circumstances characteristic of the high-risk group than regular attenders.

In another study, Taylor et al. (2021) examined a large population cohort of very young children from ages birth to 5 years to examine the participation of these children and families in universal early childhood services located in Tasmania, Australia. These universal services included community-based child health services, a parent-child early learning programme and entry into and participation in ECEC in Kindergarten.

Taylor et al. (2021) found patterns of participation by families and children and characterised these as Regular, Low and High service use. These service use patterns emerge at the outset of a child's birth and remain stable over time and across different service types. That is, families with low participation in services tended to have low contact with all services throughout the fve-year period. The fnding that low service use was consistent across service types suggests that risk factors

infuencing levels of service use are not necessarily sector- or service-specifc. The propensity or capability to participate appears to be a relatively fxed characteristic of children's family and social circumstances. What these data show is that some vulnerable children remain in persistent circumstances that challenge current models of service delivery to reach them.

#### **Intervention and Prevention Opportunities Matched to Places and Circumstances**

The birth of a child into a family is most often heralded with celebration and excitement and is typically a welcome event. So much of the common narrative about child rearing and child development is narrated through the dual lens of time and money—two of the resources parents typically mention in the balance between work and family. In truth, time and income are but two of the several resource domains essential in the development of children. In addition, human capital, social capital and psychological capital are needed above and beyond just time and money (Zubrick et al., 2005, 2014). Child development is about growth and change. And in this chapter our focus has been on a narrow, albeit critical, component of that growth: Namely, vocabulary and literacy in the population of Australian children. These are two of the markers of the wider skill that language development entails.

What we have seen so far is that early growth (as measured by vocabulary acquisition and later literacy) is typically highly variable. Children of similar ages may commence their growth from higher starting points and with higher capacities than other children. Some children grow more rapidly than others and some children who may be developmentally "on track" at one point in time, may fall behind their peers at other points. The variability of this growth makes the prediction and selection of some children over others for developmental services, on the basis of their growth and performance at one point in time, at best, ineffcient and at worst, inequitable.

What we have also seen is that while individual risks are individually and collectively poor predictors of changes in child development, their clustering, patterning, and persistence nonetheless imparts changes in the starting levels of their growth and the onward rate of this growth over time. Quite importantly, while some children outwardly appear to have *similar* rates of slower development, the underlying circumstances associated with this growth are *different*. These differences in developmental "circumstances" as we call them here, invite different approaches to how governments, agencies, and services plan and deliver developmental expectations and opportunities for young children and their families (Taylor et al., 2020).

Table 3.2 details some of the policy and service prerogatives for child development interventions and prevention strategies for which there is robust evidence of effcacy. These strategies are best thought of in terms of their impact on large populations or sub-populations of children. Clearly there are important universal


**Table 3.2** Policy and service prerogatives for child and family development services and interventions

Notes: a Parenting programs, local mother's groups, play groups, community development aimed at quality opportunities for children/families, recreation facilities, safety and area enhancements, library programs (book sharing), uptake of preschool and K provision, social marketing. b Decennial child development surveys, broad-based child development report-card based on administrative data, census estimates, AEDC, NAPLAN

strategies that apply across all developmental circumstances. Critically, children and families experiencing other circumstances would beneft from a wider, more selected or targeted range of services that address some of the unique features of their lived lives.

At the outset, it is important to appreciate that the absence of a tick in many of the boxes of Table 3.2, does not imply that the policy or service is not needed or relevant to other developmental circumstances. Instead, the indicators in Table 3.2 highlight where policies and service prerogatives are particularly relevant to the developmental circumstances we have studied here. Table 3.2 refects where there are specifc opportunities to improve developmental expectations and opportunities and how they impact on particular groups in need.

Two features of this schema should be highlighted here. First, consideration of the mix of these policies and services can be tailored to match developmental characteristics of community populations. Populations of children and families vary in their developmental circumstances and this variation is frequently spatially determined: By suburbs, regions, and areas. "Place-based" developmental planning can better match developmental circumstances to policy and service delivery. Population demography, service use characteristics, and data integrated from a variety of sources can be used to better select the mix of intervention opportunities for specifc areas. Where funds are made available, how might these characteristics guide improvements best tailored to lift developmental capacities in these areas?

Second, it is important to also note the level at which the proposed policy or service prerogative is targeted. For example, a unique feature of families with Low Human Capital is the preponderance of children whose mothers are very young. These children, relative to children in the Working Poor or Overwhelmed circumstance, show a gradual growth in their early vocabulary. While these children undoubtedly beneft from universal services, their parents (mothers particularly) would also beneft from policies that support family planning and maintain maternal engagement in education and onward training and employment. In this instance, the target level of the intervention is a parent.

#### **Conclusion and Discussion**

This chapter has focused on variations in population child outcomes as they relate to changes and patterns in the growth of child language, literacy, and subsequent school attendance. These are benchmark performances that portend future capabilities as theorised by life course approaches to human capital development. We believe the beneft of modelling the characteristics of growth in early childhood is evident and that the fndings suggest important ways of thinking about how we address early child development in the context of disadvantage.

Child development as modelled here shows us that it is highly variable over time and it is much more unpredictable than is typically presented. There are striking variations in the growth of young children that can be demonstrated in their starting capacities and in the onward rate of growth of these capacities. This variability predisposes children to move in and out of vulnerability and consequently their need for appropriate services and inventions.

The predictive utility, that is, the reliability and effciency of predicting which children will do poorly, is very poor using a wide range of empirically selected risks. Predicting with certainty which children are behind or falling behind in their development will inevitably misclassify many at one point in time. Indeed, some who are not in need early will be found to subsequently need help. This means services and developmental opportunities need to have "open doors" across a wide developmental age range.

What emerges from this view of early child development is not so much about what "works" to improve child development—there is an ample evidence base detailing the range of effective strategies (see Table 3.2)— but rather, a challenge of "how" we arrange the provision of these opportunities through services and policies to meet the developmental circumstances of children. In addressing how we arrange provision of services and policies, there have been consistent calls for early developmental interventions and opportunities that are proportionately or progressively universal (Lynch et al., 2010; Marmot et al., 2010); That is, individuals across the population are entitled to benefts proportionate to their needs. But what does this actually mean in practice? How is "proportionality" or progressive universalism achieved?

Our fndings suggest ways of thinking about this regarding early years interventions. First, recognise that at the earliest points of development, early years services and opportunities are best provided universally to all families and children with essentially no access or selection thresholds placed on means or needs. This is the entry point in life where services can assess, reach, invite, and establish developmental opportunities and begin to make provision for those in greater need. The major challenge here is ensuring that universal services are available and that they are obligated to pro-actively reach the population fairly and equitably.

Second, adjust intervention intensity and complexity against the distribution of developmental circumstances in area populations. Intervention intensity is more than just increasing the "dose," or providing more of the same of an existing program or service offered to a family or child with greater needs. For example, families who are Overwhelmed are restricted in their very capability to participate. The issue for these families is not access to service—but the effort and manner by which services *reach* for them. Addressing this may entail eliminating social barriers by adjusting or implementing policies that guarantee fairness and equity and that provide additional means for reach, access and participation.

Third, in considering the challenges of establishing proportionality, Carey et al. (2015) highlight the essential need to "ensure that decisions and actions are taken as closely as possible to citizens through a multi-layered system" (i.e., subsidiarity). In Australia there is current enthusiasm for "place-based co-designed" services that are responsive to local circumstance and needs, and that are overseen and governed at the local level thereby empowering individuals, families, and communities. Place-based co-design is certainly in line with the Carey et al. schema for proportionality. But as the authors point out, subsidiarity is not an invitation for the higher levels of governance (State and Federal) to abandon their unique responsibilities to govern, legislate and fund in ways to guarantee fairness and equity in provision, access and reach.

In concluding this chapter we return to where we started. There is ample evidence that establishes that expectations and opportunities for children in their early years have a signifcant impact on their onward capabilities and choices as theorised by life course perspectives and as shown by many empirical studies. While we believe that prevention interventions in the early years have small effects this is not to say that we believe these effects to be trivial or insignifcant. To the contrary, it is our view that the variability of growth within and between children refects the very nature of how development typically occurs.

What we also show is that disadvantage is heterogenous in its effects on child development. This heterogeneity reveals itself in the different developmental circumstances we document in this chapter. Many children are developmentally enabled and progress through their early years relatively unimpeded and without delay and with parents who are seeking optimal outcomes for them. But for many other children development is less orderly, more variable, and accompanied by individual, family and community circumstances that are associated with slower growth and gaps in their development that widen substantially into the middle years. For these children, developmental science documents what works to effectively prevent or reduce these gaps. Our work in this chapter invites a broader research and discourse in how we arrange these expectations and opportunities and create circumstances for better guided actions on the part of governments, agencies and individuals.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 4 Cultural Identity and Social and Emotional Wellbeing in Aboriginal and Torres Strait Islander Children**

**Yaqoot Fatima, Anne Cleary, Stephanie King, Shaun Solomon, Lisa McDaid, Md Mehedi Hasan, Abdullah Al Mamun, and Janeen Baxter**

*Our identity as human beings remain tied to our land, to our cultural practices, our systems of authority and social control, our intellectual traditions, our concepts of spirituality, and to our systems of resources ownership and exchange. Destroy this relationship, and you damage – sometimes irrevocably – individual human beings and their health.*

(Pat Anderson 1995)

The First Nations people of Australia comprise two similar but distinct traditional cultural groups—Aboriginal peoples and Torres Strait Islander peoples, with unique and rich cultural beliefs, practices and knowledge (Australian Institute of Aboriginal and Torres Strait Islander Studies, n.d.). Aboriginal and Torres Strait Islander peoples include all Indigenous people of the Australian mainland and Indigenous peoples of the island of Tasmania and the Torres Strait, a strait located in the northernmost extremity of the Australian mainland that connects with Papua New Guinea. In 2016, an estimated 798,365 Aboriginal and Torres Strait Islander peoples lived in Australia, representing 3.3% of the Australian population (Australian Bureau of

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia e-mail: fatima.yaqoot@uq.edu.au; mamun@sph.uq.edu.au; j.baxter@uq.edu.au

A. Cleary · L. McDaid · M. M. Hasan

Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia e-mail: anne.cleary@uq.edu.au; l.mcdaid@uq.edu.au; m.m.hasan@uq.net.au

S. King · S. Solomon Murtupuni Centre for Rural and Remote Health, James Cook University, Mount Isa, QLD, Australia e-mail: stephanie.king1@jcu.edu.au; shaun.solomon1@jcu.edu.au

Y. Fatima (\*) · A. Al Mamun · J. Baxter

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_4

Statistics, 2019). While acknowledging the diversity of Aboriginal and Torres Strait Islander cultures and identities across Australia, for this chapter, hereafter, we use Indigenous Australians as a collective term for describing both Aboriginal and Torres Strait Islander peoples.

Indigenous Australians have the longest, continuing and adapting culture in the world. For example, evidence of ritual burials dates cultural practices of Australian Aboriginals from 41,000 years ago, with other evidence dating Australia's Aboriginal occupation from over 60,000 years ago (Australian Human Rights Commission, n.d.). With over 250 languages and many hundreds of dialects, Indigenous Australian culture is diverse, vibrant and can be understood and expressed by different Indigenous Australians in different ways (Hampton & Toombs, 2013). Within this diversity reside core concepts such as family, kinship, relatedness and connectedness, which are the basis of Indigenous Australian worldviews and the philosophy underpinning Indigenous Australian social organisation, cultural identity and cultural practices (Grieves, 2009). These concepts are highly consistent with a life course approach, as outlined in Chap. 2, that emphasises linked lives, the importance of family background, intergenerational connections, contextual and environmental infuences, and the impacts of cumulative advantage and disadvantage over time.

As a result of European settlement, Indigenous Australians have suffered devastating loss of sovereignty and dispossession of lands, waterways and customary law, reduced access to their ancestral lands and intergenerational trauma (Sherwood, 2013). Consequently, ongoing disadvantage in education, employment, housing and health outcomes has contributed to appalling inequity in health and wellbeing outcomes between Indigenous and non-Indigenous Australians (Australian Institute of Health and Welfare, 2020). Nonetheless, despite the adverse impacts of colonisation, Indigenous Australians have demonstrated formidable cultural resilience in responding to historic and contemporary impacts of colonisation (Berry et al., 2010). There is growing recognition of the role of cultural identity in promoting positive health, social, educational and economic outcomes of Indigenous Australians (Roth, 2011). The emerging evidence strongly connects the health of an Indigenous person to the health of their family, kin, community, and their connection to Country, culture, spirituality and ancestry (Dudgeon et al., 2020). The Australian National Indigenous Reform Agreement also highlights the critical role of connection to culture for Indigenous Australians' emotional, physical and spiritual wellbeing (Steering Committee for the Review of Government Service Provision, 2019).

Understanding the potential protective role of cultural identity is particularly important when considering the health and wellbeing of Indigenous youth who have a disproportionately higher burden of poor health than their counterparts (Australian Institute of Health and Welfare, 2018; Dickson et al., 2019). The literature from various Indigenous communities worldwide highlights the positive impact of cultural identity on the health and wellbeing of Indigenous youth. For example, a strong sense of cultural connection is associated with reduced anxiety and depression among Indigenous Sami youth from Arctic Norway (Bals et al., 2011), reduced suicidal ideation among American Indian youths (Yoder et al., 2006) and reduced suicide risk among Canadian Aboriginal youth (Chandler et al., 2003). Gee et al. suggest that strong culture builds resilience, facilitates life balance and offers protection against adverse life experiences for Indigenous children (Gee et al., 2014), which is particularly important in the changes in the transition from childhood to adolescence and increased vulnerability to poor mental health (Christensen et al., 2017).

Compared with the adult population, the role of cultural identity and mental health outcomes is relatively less explored in Indigenous Australian children (Salmon et al., 2018, Lopez-Carmen et al., 2019). This chapter aims to fll part of this knowledge gap. Utilising the Primary Carer responses to questions about cultural identity and social and emotional problems within the Longitudinal Study of Indigenous Children (LSIC), we explore the social-emotional wellbeing of Indigenous children in LSIC and assess to what extent cultural identity reduces their risk of social-emotional problems.

#### **Cultural Identity for Aboriginal and Torres Strait Islander Peoples**

Indigenous Australians have a unique physical and spiritual connection to a country/place with unique knowledge and belief systems. The literature on Indigenous Australian culture is vast and diverse, with multiple efforts to defne this multi-faceted construct (Dockery, 2009). While the commonly used defnitions of cultural identity are based on an individual's self-awareness or self-knowledge (Usborne & Taylor, 2010), the studies based on LSIC data defned cultural identity in terms of "children knowing and understanding who they were and where they were from" (Martin, 2017). However, there is no clear consensus within the literature on how to best measure cultural identity among Indigenous Australians. A recent literature review of descriptors of Indigenous Australian's cultural identity identifed six broad, frequently cited cultural domains (Salmon et al., 2018):


Within each of these domains exist sub-domains; for example, cultural expression and continuity contain identity, traditional practices, arts and music, community practices, and sport (Salmon et al., 2018). Other authors have explored the conceptualisation and measurement of Australian Indigenous cultural identity within the

context of the factors associated with connection to culture. For example, Dockery et al. used data from the 2002 National Aboriginal and Torres Strait Islander Social Survey to propose two broad dimensions of connection to culture: cultural identity (spoken languages, recognition of clan, tribal group or language group and recognition of homelands) and cultural participation (attendance at, or participation in, cultural and related social activities) (Dockery, 2009). Dockery used this conceptualisation and operationalisation of connection to culture to explore how the connection to Indigenous culture might shape Indigenous Australians' engagement with education and training (Dockery, 2011). Despite the nuances and diversity in defning cultural identity, there is a universal understanding that culture is core to identity and sense of self and what it means to be healthy and well for Indigenous Australians (Colquhoun, 2012).

#### **The Longitudinal Study of Indigenous Children**

The rich and unique data from LSIC (also known as Footprints in Time), a large cohort study of Indigenous children in Australia, offer an untapped opportunity to examine the link between cultural identity and social-emotional wellbeing in Indigenous children. The LSIC is a longitudinal study conducted by the Commonwealth Department of Social Services, Australia. Unlike the data examined in the previous chapter, which is a national longitudinal sample of all Australian children, LSIC focuses specifcally on Indigenous children and was designed to provide a source of information on what helps Indigenous children to gown up strong and healthy. The study commenced in 2008, involving 1759 Aboriginal and Torres Strait Islander children aged 6–18 months at baseline (B cohort) and a cohort of children aged 3.5–5 years at baseline (K cohort) (Thurber et al., 2015). Participants were recruited through purposive sampling from 11 diverse sites across Australia, covering a wide range of socioeconomic status, rural and remote locations and cultural groups. During each annual follow up, a face-to-face survey was conducted with a parent or primary carer of the child (B and K cohort) and the child (K-cohort). So far, data for 11 waves of the LSIC cohort has been released.

In addition to sociodemographic, lifestyle and health-related variables, the LSIC dataset offers rich information on the cultural knowledge, cultural identity, extended family and community, and strengths of Indigenous culture. Further details on the study and methodology are provided elsewhere (Department of Families Housing Community Services and Indigenous Affairs, 2009; Dodson et al., 2012).

Since only wave 8 (2016) of the LSIC cohort captured data on cultural identity variables for children, this chapter is based on wave 8 of the cohort. In addition to cultural identity, wave 8 also collected data on social-emotional wellbeing and other key sociodemographic variables for children. Considering a meaningful understanding of cultural identity might be diffcult for young children, we restricted the analysis to the older cohort (K cohort) of wave 8 (mean age 9.3 years (±1.52)).

#### *Key Variables*

In the LSIC cohort, children's social and emotional wellbeing, our dependent variable, is determined by primary carer responses to the Strengths and Diffculties Questionnaire (SDQ) (Goodman, 1997). The overall score (range 0–40) is computed by summing scores across the frst four subscales of the primary carer reported emotional symptoms, conduct problems, hyperactivity and peer problems, with a higher score indicating more social-emotional problems. Summed scores are grouped into three categories: normal, borderline and of concern. Children in the "of concern" category indicate a substantial risk of mental health problems. Based on the recommendations of the LSIC technical group, we used the SDQ cut-off ≥14 for dichotomisation ("normal/borderline" and "of concern") to identify the children at risk of poor mental health (Thurber et al., 2019).

Cultural identity, one of our main independent variables, was based on the mean score of the following four items reported by the primary carer: (1) "Study Child (SC) feels good about being Indigenous in class"; (2) "SC wants to share things about being Indigenous in class", "SC feels safe about being Indigenous in class" and "SC wants people in class to know that he/she is Indigenous. Each of these items was coded as "never", "sometimes", and "always." Cultural knowledge, another key independent variable, was based on mean scores of the items exploring the child's knowledge of (1) clan/tribe, (2) their people and (3) family stories/history. Each of these items was coded as "Yes" or "No".

#### *Covariates*

The selection of covariates was guided by previous evidence on the link between cultural identity and mental health of Indigenous people (Dockery, 2011, Colquhoun, 2012). Child-related covariates included gender and age. Weekly family income after deductions were reported by the primary carer and categorised as "less than \$399, \$400–\$599," "\$600–\$999," and "\$1,000 or more."

Major life events were derived from the list of major events experienced by the family in the last 12 months. These events are not necessarily negative and can be any event in life with a signifcant impact on a person's wellbeing (Wilkins, 2012). In the LSIC cohort, some life events are related to the normal human life cycle, for example, births, deaths, and marriages, while others are related to external stressors such as the loss of a job, fnancial hardship, or social isolation. The 15 possible events covered in the LSIC data are: pregnancy, sickness, death, job loss, arrested or jailed or police problem, divorce, humbugged (harassed for money), mugged, robbed, assaulted, worries about money, alcohol or drug problems, child upset by family arguments, child scared by other people, child cared for by someone else for at least 1 week. Based on the LSIC technical group's recommendations, we used the cut-off ≥3 to dichotomise major life events (Thurber et al., 2019).

Geographical remoteness was measured using the Level of Relative Isolation (LORI) scale (Department of Health and Aged Care and the National Key Centre for Social Applications of Geographical Information Systems (GISCA) 2001). The LORI scale indicates the relative distance of families from population centres (coded as "none", "low isolation", "moderate isolation", "high/extreme isolation"). Area-level disadvantage was measured using the Index of Relative Indigenous Socioeconomic Outcomes (IRISEO) scores. The IRISEO deciles are based on nine socioeconomic status measures (SES) such as employment, education, income and housing, and rank SES for an area in which an individual resides relative to other Indigenous Australians (Biddle, 2009). In the LSIC cohort, areas were categorised as having the "lowest (IRISEO 8–10)," "middle (IRISEO 4–7)," or "highest (IRISEO 1–3)" level of disadvantage.

#### **Analyses and Results**

Initial analyses examined basic descriptive statistics (for example, frequency, mean, standard deviation) to summarise information related to sociodemographic, cultural identity and social-emotional wellbeing. Regression analyses were subsequently used to identify factors associated with increased risk of social and emotional problems. The results of logistic regression models are reported as odds ratios (ORs) and 95% confdence intervals. A p-value of <0.05 was adopted as a signifcance threshold for statistical signifcance. However, it is worth mentioning that using p < 0.05 for statistical signifcance is merely a convention and should not be used refexively to determine the size or importance of the observed effect (Baker, 2016). The interpretation of quantitative analyses should be based on a combined consideration of the conceptual framework, confdence intervals, p-value and sample size (Concato & Hartigan, 2016). All statistical analyses were undertaken using Stata IC 15.0 (Stata Statistical Software, College Station, Tx, USA).

As highlighted in Table 4.1, social and emotional wellbeing was explored in 498 Indigenous children (mean age 11.0 years, SD ± 0.51). The majority of survey respondents were the mother of the child (83.2%). A signifcant proportion of the study participants identifed as Aboriginal (88.9%), and the rest identifed as Torres Strait Islander (6.1%) and both Aboriginal and Torres Strait Islander (4.9%). Approximately half of the sample was male (50.2%), 10% were living in extreme geographical isolation, and 18.8% of children were living in the most disadvantaged areas. Approximately one-quarter of the participants (24.3%) reported a family income of less than \$600/week. Nearly half (47.1%) of the study participants had experienced ≥3 major life events in the 12 months before their interviews. About one-third (30.3%) of study children had "of-concern" SDQ scores. The mean scores for cultural knowledge and cultural identity were 0.60 (±0.36) and 1.70 (±0.39), respectively, suggesting that, on average LSIC children had fairly high cultural knowledge and a strong sense of cultural identity.


**Table 4.1** Distribution of sociodemographic, family and geographical area related variables in LSIC children (based on wave 8 data)

Note: Increased risk of mental health issues,\*\*signifcance level p < 0.05, \*\*\*Based on IRISEO deciles. Source: Longitudinal Study of Australian Children Wave 8, 2016

#### **Cultural Identity and Social and Emotional Wellbeing in Indigenous Children**

Results from regression analyses shown in Table 4.2 suggest that age, geographical isolation, and area-level disadvantage are not associated with social and emotional wellbeing in Indigenous children. However, there was a signifcant gender difference as females had lower odds of high social and emotional problems than their male counterparts. Children who experienced major life events (≥3) also had higher odds of social and emotional problems. In comparison, cultural knowledge and cultural identity seemed to play a protective role and reduced the odds of social and emotional problems. In multivariable analysis, even after controlling for socioeconomic disadvantages, the protective effect of cultural identity was still evident. High


**Table 4.2** Association between cultural identity and poor social and emotional problems ('of concern' SDQ scores) in LSIC children (based on wave 8 data)

Note: Based on IRISEO deciles, signifcant associations (p < 0.05) are highlighted in bold. Source: Longitudinal Study of Australian Children Wave 8, 2016

scores on cultural knowledge (OR: 0.49; 95% CI: 0.28–0.88) and cultural identity (OR: 0.42; 95% CI: 0.22–0.79) were associated with signifcantly reduced odds of social and emotional problems in Indigenous children.

#### **Discussion**

These results suggest that a signifcant majority of the Indigenous Australian children that participated in LSIC are experiencing a high burden of social and emotional problems and are at increased risk of poor mental health. However, children with strong cultural identity and knowledge are less likely to experience social and emotional problems than their counterparts. The potentially protective effect of cultural identity further highlights the need for strengths-based approaches to reduce mental health issues in Indigenous children. Shifting from a defcit narrative to

capitalising on Indigenous culture as a strength can lead to better engagement, uptake, and delivery of mental health programs and achieve better outcomes for Indigenous children.

Cultural identity is a key factor affecting the health and wellbeing of Indigenous children, who, due to rapid changes in globalisation, colonial disruption and undermining of Indigenous cultures, face greater challenges in understanding their identity from past, present to future self. Many young Indigenous people's social interactions and experiences are affected by past and current social realities, including negative stereotyping, racism, and outlawing Aboriginal languages (Stoneham et al., 2014; Macedo et al., 2019). These negative experiences strongly affect selfworth and are linked with self-deprecation in young people (Wexler, 2009). However, through cultural strength and resilience, Indigenous people have contributed to better outcomes for their people. Having a strong cultural identity and knowledge helps young Indigenous Australians to make positive social connections with people in their family and broader community and feel a sense of belonging (Renshaw, 2019). In turn, this promotes resilience, enhances self-esteem, and protects from poor mental health, offering opportunities for living life to full potential (Dudgeon & Walker, 2010).

Similar to our results, evidence from other Australian studies also highlight the protective effect of cultural identity and cultural knowledge in improved health outcomes for Indigenous Children. For example, the Western Australian Aboriginal Child Health Survey reports that children whose carers were more fuent in an Aboriginal language had lower risks of emotional or behavioural diffculties (Zubrick et al., 2005). The work by Colquhoun et al. highlights that knowing their cultural heritage, background, language and connection to Country and community is integral to Indigenous children's sense of belonging and pride and helps them attain emotional strength to face challenges in life (Colquhoun, 2012). The 2008 National Aboriginal and Torres Strait Islander Social Survey fndings highlight positive associations between cultural participation and cultural identity and perceived wellbeing, happiness and mental health (Dockery, 2011).

Studies from different Indigenous communities suggest that leveraging the strengths of Indigenous culture is a largely untapped opportunity for addressing Indigenous disadvantage (Dockery, 2020). The majority of the health programs still fail to recognise the strengths of Indigenous knowledge/culture and do not align with Indigenous people's needs and expectations, and therefore are inherently ineffective in Indigenous communities. The evidence has made it clear that health and wellbeing programs and services cannot be effective unless program/service planning, design and delivery are centred on cultural identity and cultural pride (Colquhoun, 2012, Australian Institute of Health and Welfare, 2013; Kingsley et al., 2013). Therefore, understanding how Indigenous youth connect with their culture and its application in mental health program design and delivery is crucial to addressing the growing trends of poor mental health in Indigenous youth.

Indigenous Australian culture is dynamic and continues to evolve and develop in response to historical and contemporary circumstances (Commonwealth of Australia, 2013). As highlighted by a life course approach children's lives are shaped by the social environment in which they are born and raised. The 'linked lives'perspective is particularly relevant to the life course of Indigenous children, given the role of the family in children's lives, social structures and vast kinship that are an important part of Indigenous children's lives (Biddle, 2010). There is established evidence on intergenerational transmission of trauma experienced by elders and family members to Indigenous children leading to poor health and wellbeing outcomes (Australian Institute of Health and Welfare, 2019).

However, connection to family and kinship are also important sources of cultural knowledge and play a big role in strengthening cultural identity (Colquhoun, 2012). Children's perceptions and understanding of cultural identity are strongly infuenced by parents' cultural knowledge sharing, their sense of Indigeneity, and the value of cultural heritage to them (Martin, 2017). Children's understanding of cultural identity is also shaped by whether they live on Country (ancestral land), have opportunities to participate in cultural practices, spend time and maintain meaningful relationships with people in their family and wider community (Jackson-Barrett & Lee-Hammond, 2018). These opportunities are limited for children in out-ofhome care who move further from their cultural identity and community (Richardson & Osborn, 2007). Therefore, for Indigenous children who need to be removed from their homes to protect them from harm, protecting and strengthening their cultural identities must be a key priority for child welfare services.

Indigenous children have signifcantly greater social and emotional problems, mental health issues, psychological distress and suicide rates than their counterparts (Priest et al., 2012, Department of Health, 2013). Identifying interventions and approaches that support better uptake and delivery of services is vital for improving Indigenous youth's mental health outcomes. It is promising to see that policymakers are now recognising that interventions centred on cultural identity and connections are critical to improving Indigenous Australians' health outcomes. Though nonrandom sampling and small sample size limit the generalisability of our fndings, our results further support the long-awaited shift in the defcit narrative focus to a strengths-based discourse. Strong evidence on the role of cultural identity reinforces the need for interventions centred on Indigenous knowledge and leadership to offer effective solutions for improving the health and wellbeing of Indigenous people.

#### **Conclusion**

Reducing the gap in health and wellbeing between Indigenous and non-Indigenous Australians is a critical national priority. Unacceptably high rates of poor mental health and suicide in Indigenous people indicate that current health and wellbeing services that excessively focus on defcit correction fail to improve outcomes. Socioeconomic disadvantages are linked with varying levels of psychological distress in Indigenous Australian children. However, attachment with Indigenous culture, clan and community, and cultural identity are individual assets that contribute to the health and wellbeing of children, buffering the negative effect of disadvantage in Indigenous children. Therefore, for better engagement and impact of health and wellbeing programs, policymakers and services providers must take a different approach and offer health interventions and services capitalising on the strengths of Indigenous culture and cultural identity.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 5 Refugee Children in Australia: Wellbeing and Integration**

**Rennie Lee and Sin Yi Cheung**

The United Nations estimates that at the end of 2019, there were 29.6 million refugees in the world and another 4.2 million individuals who were waiting for the outcome of their asylum claims (UNHCR, 2021). Displacement continues to grow as a result of climate change, confict, and hunger. Currently, most of the world's refugees come from Syria, Venezuela, Afghanistan, South Sudan, and Myanmar. Australia settles around 13,000–14,000 refugees annually with children comprising around 30–40% (Henley & Robinson, 2011). Since the start of Australia's humanitarian program in 1945, Australia has resettled over 800,000 refugees (Phillips, 2015).

Unlike immigrants who arrive via family or employment, refugees are a particularly vulnerable group who may face greater integration challenges. Many refugee adults and children fee violence and persecution. Forced displacement often means brutal uprooting and dismantling of communities and families, with some refugee children feeing as unaccompanied minors. Additionally, many refugees arrive with few social networks and economic resources implying signifcant challenges to their onward journeys over the life course.

Most refugees arrive in Australia as families (McMichael et al., 2011). The Australian Government considers immediate family members to be spouses or de facto partners and children under age 18, though households will likely incorporate those outside of a nuclear family. In fact, the United Nations High Commissioner for Refugees (UNHCR) posits that refugee integration is more successful among

R. Lee (\*)

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia

e-mail: rennie.lee@uq.edu.au

S. Y. Cheung Cardiff University, Cardiff, UK e-mail: cheungsy@cardiff.ac.uk family units than individual refugees (McMichael et al., 2011). However, the integration of each family member will differ and for refugee children in particular, their integration may well differ from that of their parents, who arrive in the host country as adults.

Many refugee children arrive in Australia during a formative time in their lives and face different challenges than their parents. From a life course perspective, the integration and wellbeing of refugee children is shaped by the timing and context of migration, including their age at migration, country of origin, and refugee status. While migration in itself represents an important transition that infuences one's life trajectory and future experiences, this will differ for refugee children who migrate at various ages and sometimes under stressful conditions related to confict and trauma (Kim et al., 2018).

Refugee children will grow up in a new society and culture, needing to learn and communicate in a new language and adjusting to schooling in a foreign education system (Henley & Robinson, 2011; Joyce et al., 2009). While adults rely on established diasporic communities and support networks, refugee children are more frmly grounded in host country education systems (Nunn et al., 2014). In turn, refugee children face several barriers that are strongly related to their wider wellbeing outcomes (Correa-Velez et al., 2010; Trickett & Birman, 2005). To illustrate, perceived discrimination and bullying greatly impede the ability to develop a sense of belonging and also negatively impact their educational outcomes (Correa-Velez et al., 2017), which can drastically alter life courses as education is a key indicator of integration outcomes (Strang & Ager, 2010). Further, peer diffculties in school are associated with lower adjustment and worse health outcomes (Lau et al., 2018). Importantly, resettling in a host country with new belief systems and values will likely challenge their socio-cultural adjustment, which has implications for their wellbeing. To successfully integrate, refugee children will need to overcome these barriers.

One aim of Australia's humanitarian programme is to facilitate refugees' full social, economic, and civic participation as well as psychosocial health and wellbeing (Correa-Velez et al., 2010). While the effects of the humanitarian programme have primarily focused on refugee adults, the programme is likely to have more prolonged and sustained effects on refugee children, given their young age at arrival. Refugee children arrive as permanent residents and who they grow up to be has profound consequences on Australia's social and economic future. In this spirit, this chapter focuses on several aspects of refugee children's integration, namely wellbeing, language profciency, and academic achievement (Due et al., 2016). Educational settings are important spaces for socialization and human capital accumulation, which allow refugee children to enhance their social and emotional health (Joyce et al., 2009).

This chapter provides a demographic portrait of refugee children in Australia across national origins using nationally-representative data. We address the following questions: (1) What are the origins and premigration experiences of refugee children? (2) What is the family and household structure of refugee children? And (3) How do refugee children's psychosocial adaptation and language profciency differ by national origin and gender?

#### **Integration of Refugee Children in Australia**

While there is growing knowledge about the integration of contemporary refugees in Australia, we know far less about the outcomes of refugee children and youth. If refugee children grow up to have poor outcomes as adults, this will have a knock-on effect on the socioeconomic wellbeing of the country. Importantly, failure to fully integrate refugee children into the Australian society will only serve to deepen existing inequalities between the newcomers and the host population.

#### *The Role of National Origin for Refugee Children's Integration*

Refugee children arriving in Australia may face greater risks because of their preand post-migration experiences of loss, separation, trauma, and disruption (De Anstiss & Ziaian, 2010). These experiences may vary by national origin or region of displacement (Portes & Rumbaut, 2014). For instance, the reasons that immigrants leave or fee their origin country refect the conditions in their country of origin (Levels et al., 2008). In turn, the structural conditions of the country of origin can affect the outcomes of refugee children in the host country. For instance, humanitarian migrants may face more premigration traumatic experiences such as violence, torture of family members, separation from parents, personal injury, or living in a refugee camp (Hadfeld et al., 2017). Such premigration experiences are linked to refugee's country of origin and thus, this suggests that disaggregating refugee's experiences by national origin may illuminate our understanding of the ways in which these experiences shape their integration in the host country. Thus, national origin serves as a useful proxy to capture the connection between premigration experiences and the integration of refugee children.

In general, studies tend to treat refugees as a homogenous whole or focus on a few specifc groups (Correa-Velez et al., 2017). For instance, Lau et al. (2018) study the adjustment of refugee children and adolescents but do not examine them by national origin. Rousseau et al. (1998) focuses on broad regions when assessing refugee children's emotional profles. Likewise, qualitative research on refugee children often rely on small-scale, interview-based studies involving one or a few national origins (Joyce et al., 2009; Nardone & Correa-Velez, 2016). While these studies are useful for generating theories, they do not allow us to generalize any patterns among a large and diverse population of refugee children. In part, this is driven by data limitations. To our knowledge, detailed data on refugee children from multiple national origins are rare. Australian Census microdata provide detailed information on visa category and national origin but it does not include any

information about wellbeing and premigration characteristics. Other large-scale and longitudinal surveys in Australia such as HILDA do not have a large sample of humanitarian immigrants, especially refugee children. In turn, we know little about how refugee children's premigration characteristics and national origin shape their integration experience in Australia.

National origin is important as refugees from different countries of origin experience different types and levels of exposure to trauma and premigration circumstances (Fazel et al., 2012; Bean et al., 2007; Rousseau et al. 1998). When examining psychological distress, Bean et al. (2007) found that patterns of distress differ by national origin, with unaccompanied asylum-seeking children from Eritrea, Ethiopia and Guinea displaying higher distress scores compared to those from other African countries and China. The national origin group that refugee children belong to matters for their integration as the coethnic community can be a source of support and provide resources that enable integration (Fazel et al., 2012; Liebkind, 1996).

#### **Humanitarian Migrants in Australia**

From 2018 to 2019, the largest number of humanitarian visas were granted to individuals born in Iraq (41.5%), the Democratic Republic of Congo (12.4%), Myanmar (11.7%), Syria (10.7%), and Afghanistan (7.7%) (Department of Home Affairs, 2019). Most contemporary refugee groups are viewed as racially different to the mainstream society who have distinct religious affliations, with Afghans and Iraqis likely to be more visually distinctive in Australian society than Asian or European groups (Waxman, 1999). These differences will likely infuence how refugees are received in Australian society and will differ to the experiences encountered by refugees from previous waves that primarily consisted of Bosnians, Iraqis, and black Africans (Colic-Peisker, 2008). A brief country profle of incoming refugees serves to contextualise the circumstances under which they arrived in Australia.

#### *Iraqi Refugees and Immigrants*

Roughly 77% of Iraqi migrants in Australia enter through the offshore Humanitarian program (Taylor & Stanovic, 2005) and approximately half of them arrived prior to 2007. Since 2015, there has been a signifcant focus on the settlement of Iraqi refugees following the Syrian confict; between July 2015 and December 2017 a total of 13,567 Iraqi refugees migrated to Australia with nearly 8000 in NSW and over 4000 in Victoria (Collins et al., 2018). Despite Islam being the offcial religion in Iraq, 45% of Iraqis in Australia are Catholics or Assyrian Apostolic Christians (Evason, 2015), probably a result of sectarian conficts driving out religious minorities from the country. While the Australian Government actively encourages settlement in regional areas, humanitarian migrants tend to settle in areas with members

of their own communities. For Iraqi refugees, this is exemplifed by the large concentration of settlements in the Fairfeld and Liverpool areas of Western Sydney and the Hume City area of Melbourne (Collins et al., 2018). Colic-Peisker and Tillbury (2007) highlight the increased diffculties faced by recently arrived Iraqi refugees in the Australian labour market, pointing to their 'visibly different' ethnic identity. By settling in established co-national communities, these diffculties can be mitigated through ethno-specifc support networks (Waxman, 1999).

#### *Iranian Refugees and Immigrants*

Following the Iranian Revolution in 1979 and the subsequent Iraq-Iran war in the 1980s, Australia saw an infux of humanitarian migrants with the establishment of a program aimed at protecting religious minorities from the region (DHA, 2018b). This has allowed for successive generations starting in the 1990s to pursue visas under the Family Streams of the migration program (Adibi, 2008). In general, Iranian immigrants show greater representation in skilled migration intakes compared to others in the region, largely due to the increasingly precarious economic climate in Iran (DHA, 2018a). Consequently, Iranians in Australia are engaged in white collar labour at higher rates relative to other Middle-Eastern migrant groups and have higher pre-migration education and qualifcations (Adibi, 2008). The majority of Iranians settled in Australia arrived after 2007. Three quarters speak Persian at home while only a third of Iranian refugees in Australia are Muslim, with nearly 30% reporting no religion (Evason, 2016a, b). Perhaps due to the secular tradition prior to the Islamic Republic, Iranians are also known to be more progressive in their cultural interpretation of Islamic practices such as allowing women the choice of wearing a hijab or not. While it is unclear how the overall composition of Iranian refugees compares to their counterparts arriving as non-humanitarian migrants, the Iranian immigrant community in Australia provides an indication of the context that Iranian refugees encounter upon arrival.

#### *Afghani Refugees and Immigrants*

Decades of confict in Afghanistan have caused signifcant forced displacement. Afghan refugees constitute the second largest refugee population in the world (UNHCR). Most Afghanis experience irregular migration (Koser & Marsden, 2013; Nardone & Correa-Velez, 2016). In turn, many may arrive in Australia without authorization and may have transited through other neighbouring countries, such as Pakistan or Iran, before arriving in Australia (Nardone & Correa-Velez, 2016). The UN refugee agency estimates that over half of all asylum seekers in Indonesia seeking resettlement are from Afghanistan (Karlsen, 2016). This represents a common transit point for many Afghani refugees, who were often left in limbo when seeking

settlement in Australia. Afghani refugees in Australia particularly beneft from established communities of compatriots, and the recently arrived Afghanis have reduced likelihood of secondary migration compared to other migrant groups (Waxman, 1999). Over 90% of Afghanis in Australia are Sunni Muslims and 60% are men. They also tend to be younger with a median age of 31 compared to Iranian and Iraqi immigrants. While the majority speak Dari at home, a signifcant proportion also speak Hazaraghi and Persian (excluding Dari) as well as other languages, making Afghanis one of the most linguistically diverse ethnic communities (Evason, 2016a, b).

#### *Middle East/North African (Egypt, Syria) Groups*

As a result of an ongoing civil war beginning in 2011, over fve million Syrians have fed to nearby countries in the region, including 3.6 million in Turkey (UNHCR, 2020b) and over 130,000 in Egypt (UNHCR, 2020a). Among Egyptian refugees, many are Coptic Christians, a religious minority that continuously faces persecution in their origin country. In response to this crisis, the Australian Government announced an additional 12,000 places in its existing humanitarian resettlement program for those feeing confict in both Syria and Iraq (DSS, 2015). Between July 2015 and December 2017, over 13,000 Syrian humanitarian migrants settled in Australia (Collins et al., 2018). As of the 2016 Australian census, 41.3% of Syriaborn people in Australia had arrived between 2012 and 2016 (DHA, 2018c). Syrian settlement in Australia is more evenly dispersed across Australian States with less established communities compared to other migrant groups feeing confict in the region. Although the majority of Syrian refugees in Australia have secondary school education, a good proportion of them are highly qualifed. Syrian women also have a relatively high labour market participation due to men fghting and killed in sectarian conficts (Collins et al., 2018).

#### *Central Asian (Pakistan, Myanmar, Nepal, Sri Lanka) Group*

Much of the humanitarian intake of migrants from the Central Asia region are resettled from countries such as Malaysia, Indonesia, and Thailand (Karlsen, 2016). Of the 154,000 refugees and asylum seekers in Malaysia, over 140,000 were from Myanmar (UNHCR, 2015). Between 2005 and 2020, Australia has resettled 13,380 Myanmar-born refugees who had sought asylum in Thailand, and a total of 22,462 (UNHCR, 2020c).

Due to increasing ethnic unrest in Bhutan in the 1990s, over 100,000 Bhutanese sought refuge in nearby Nepal, and at the beginning in 2008 Australia began resettling over 5000 Bhutan-born refugees (DHA, 2018a). Most of these refugees are Lhotshampa or ethnic Nepalese from Southern Bhutan. Prior to this resettlement program, the Bhutanese community was not well established with only 1.2% of Bhutan-born people in Australia arriving prior to 2007 as of the 2016 Census. While resettled groups from other regions gravitated towards pre-established communities in Victoria and New South Wales, the Bhutanese people are more centralised in South Australia and Western Australia; 23.7% and 21.9% of Bhutan-born people in Australia respectively (DHA, 2018a).

#### **The Role of Education in the Origin Country**

One important factor when examining the integration of refugee children is the educational context in their country of origin. In theory, public education is free to all children up to secondary level in many MENA countries but universal enrolment in schools remains a major challenge (UNICEF, 2020a). Iraq is known to have one of the best education system among the MENA countries where free education for all is provided from primary to PhD level. Yet decades of confict has led to severe damage to major infrastructures and school closures, leaving hundreds of thousands of children out of school. Between 1960 and 2009, Iraq and Iran spent a signifcant amount of their modern history in confict: 52 and 48 years respectively (UNICEF, 2020b). Many children were forced to fght and become child soldiers. The risk of dropping out of primary and lower secondary education is extremely high, especially among girls and those with special learning needs. Children in Iran are slightly better off following major education reforms in 2011 with 95% of primary school enrolment. However, over three quarters of a million children remained out of school in Iran in 2015–16. The situation in Afghanistan is probably the worst where the Taliban imposed an outright ban on schooling for girls. Threats and intimidation of teachers and healthcare workers, and school closure are also widely documented by United Nations Assistance Mission in Afghanistan (UNICEF, 2016).

Consequently, on arrival, many refugee children will have experienced years of being out of school for different reasons depending on their country of origin. Religious identity as well as observance and cultural expectations of gender roles further exacerbate the educational and integration experience of boys and girls from MENA countries. It is therefore imperative to examine integration outcomes of refugee children by national origin and gender.

#### **Measuring Refugee Integration**

In general, studies investigating integration outcomes of refugee children have focused on wellbeing, education, and health. Among these, refugees tend to show more mental health diffculties than their native-born counterparts, including higher rates of post-traumatic stress disorder (PTSD) and problem behaviours (Hadfeld et al., 2017). This is likely to differ by national origin and understanding which groups may experience greater diffculties with wellbeing can inform government policies and programmes.

Successful integration of humanitarian migrants into Australian society can be defned through a variety of measures including education, health, housing, employment, language training, and social connections (Ager & Strang, 2008). These measures of integration are often associated with the particular contexts the migrant groups are leaving behind. For example, pre-migration education, while not improving labour participation in the short-term following migration, does increase access to employment after 2 years (Delaporte & Piracha, 2018). Social integration measures such as English profciency and self-suffciency, diverse friendships networks, and having a sense of belonging in Australia are associated with better physical and mental health outcomes (Chen et al., 2019). However, barriers often exist in the cultural adjustments surrounding gender roles and can further entrench disadvantage for refugee women in accessing education opportunities, particularly for those from more traditional societies (Hatoss & Huijser, 2010; Cheung & Phillimore, 2017). This chapter draws on longitudinal data from *Building a New Life in Australia* (BNLA) to offer new evidence in our understanding of the integration and wellbeing of refugee children in Australia and policy recommendations to address the social disadvantages facing this population. We examine language profciency, household characteristics, and mental health and wellbeing of refugee youth in Australia.

#### **Data and Methods**

We analyse data from BNLA, a nationally-representative, household panel study of humanitarian immigrants in Australia, aimed at assisting the government in policy development and programme improvement in a bid to identify barriers to successful integration. The study includes around 1500 refugee migrating units and over 2000 individuals within these units. The sample encompasses individuals who received a permanent humanitarian visa either onshore (asylees) or offshore (refugees) between May and December 2013. Each migrating unit contains a primary applicant and secondary applicants such as spouse or children. All individuals in BNLA must be aged 15 or older and living with an adult primary applicant at the time of the interview. Secondary applicant adolescents aged between 15 and 17 are of particular interest. BNLA collects information on the same respondents since 2013 using faceto-face and telephone interviews. In Wave 3, BNLA conducted a child module that focuses explicitly on the wellbeing and social outcomes of children in the household, with a goal of exploring intergenerational transmission of trauma among refugee communities. The child module consists of two questionnaires completed by the primary caregiver on up to two children aged 5–17 years and a self-report questionnaire by children aged 11–17 using a short Pen and Paper Instrument (PAPI). Our analyses draw on both the longitudinal household survey data and the crosssectional child module as the two samples include different information about refugee children's psychosocial outcomes and family background. Our analytical sample contains refugee children from the longitudinal sample (n = 596) and from the Wave 3 child module (n = 159).

#### **A Portrait of Refugee Children**

Table 5.1 provides descriptive statistics on key characteristics of our sample of refugee children using the longitudinal household data fle of BNLA and the child module. From the literature we know that refugee integration outcomes of women and men vary enormously (see for example Cheung & Phillimore, 2017), so we present these statistics by gender. We discuss the national origin background of refugee children, age at migration, premigration experiences, including whether they spent any time in refugee camps, detention, their migration pathway (on-shore or offshore) and their English fuency and literacy. We also report on the socioeconomic characteristics of their migrating unit. On outcomes, we focus on refugee children's health, school achievement and behaviour using data from the child module in Wave 3.

Broadly speaking, our sample is slightly skewed towards female comprising approximately 59%. The average age of the respondents is 18 and their average age at arrival is 15.6. We observe few differences in age and age at arrival by gender though females tend to be slightly older. The higher age at arrival among the sample is likely due to the fact that the longitudinal sample of the BNLA was restricted to persons aged 15 and above. The older age at arrival among our sample suggests that this group may experience additional challenges in the integration process, especially in educational attainment. Children who arrive at later ages are often further behind in school (Busby & Corak, 2014; Corak, 2011). In Australia, the minimum school leaving age is 17. It is possible that children who arrive close to that age may not be accepted for regular school (Corak, 2011). Even if they are accepted, they may experience additional challenges adapting to a completely new academic system and a western-oriented curriculum taught in a language that is not their mother tongue. Many refugee children on arrival may not speak any English at all or are likely to still be developing their English language skills. A later age at arrival for refugee children from non-Anglophone backgrounds is strongly associated with lower educational attainment, which has long-term implications for their adult behaviours and socioeconomic and wellbeing outcomes (Beck et al., 2012).

In terms of national origin, 31% of refugee children in our sample come from Iraq, 22% are from Afghanistan and a further 14% from Iran. We also have a substantial proportion from Nepal (16%), Myanmar (5%), Pakistan (3%), and Sri Lanka (1.6%). Given the smaller size of these groups, we have combined them in one group as South and Central Asia. Likewise, we combined national origin groups from Egypt (0.4%) and Syria (3.1%) together as MENA (Middle East/North Africa) countries, and Congo (4.9%) is included in 'Other Africa'.


**Table 5.1** Descriptive statistics for refugee children by gender

Source: Building a New Life in Australia survey, waves 1–5

Note: Values presented are percentages unless otherwise specifed

The overwhelming majority of our sample (98%) completed their application off-shore so their refugee status was confrmed prior to arriving in Australia. Among the refugee children in our sample, all females were off-shore applicants compared with 94.4% of males. Arriving in Australia with a confrmed refugee status is advantageous as it allows refugee children full access to refugee resettlement services and resources. For instance, Australia's resettlement program for humanitarian migrants has been well-regarded (Correa-Velez et al., 2010). In contrast, only 2% were in Australia (onshore) when they submitted their application for a humanitarian visa. Though the rules for asylum seekers have changed over time, this suggests that a small proportion of refugee children were on any kind of temporary status while awaiting for their refugee status to be determined.

Relatedly, a substantial proportion of children spent time in a refugee camp before they arrive in Australia, with approximately 30% of females and 21.7% of male children experiencing this transition, which is likely to have an impact on their integration. By contrast, our data show that only 2% of refugee children were in detention before their arrival. There is mandatory detention for individuals who arrive in Australia without a visa and seeking asylum (Steel et al., 2011). Prolonged detention is associated with lower mental health among refugees (Steel et al., 2006).

Respondents in the longitudinal sample were asked eight questions at each wave about their post traumatic experiences. We subsequently created a dichotomous measure to assess whether they were likely to have post-traumatic stress disorder (PTSD) or not. Our data show that almost half of refugee children in our sample (44.6%) experienced PTSD. This is unsurprising as experiences of trauma and confict can have long lasting impacts and impede children's ability to learn in school and to adapting to a new life in Australia. Table 5.1 shows that female children show signifcantly higher rates of PTSD (47.2%) relative to their male counterparts (40.9%). This is consistent with Bean et al. (2007) and Phillimore and Cheung (2021) who found that female refugees typically exhibit worse mental health outcomes.

The overwhelming majority (94.5%) of refugee children continue to live with the primary applicant whom they migrated with, though males were more likely to do so (98.4%) relative to females (91.8%). Most refugee children live in households with around 5–6 persons, suggesting that refugee children belong to larger families or live with household members outside of the nuclear family.

#### **The Importance of Home Environment**

To fully understand the home environment refugee children are being brought up in, it is important to consider the background of the migrating unit, which consists of all persons who migrated to Australia under the same migration application. The migrating unit is comprised of a primary applicant who sponsors refugee children as dependants. While the migrating unit is not necessarily the parents of refugee children, it is likely to be the case given how Australian immigration policy acknowledges family relationships; spouses and children are the main relationships recognized as secondary applicants. In turn, it is unlikely that refugee children will have migrated with family members outside of their nuclear family. Therefore, we use migrating unit as a proxy for family background.

A risk factor for refugee children's integration may be the overall level of disadvantage in the household. Some ways to understand this include the average education and the labour market participation of the household. Educational attainment is measured as a categorical variable with a value of "0" representing no formal education and a "4" representing a university degree. As indicated in Table 5.1, refugee children live in households where the average highest education has a value of 1.8, which roughly approximates to less than 7 years of formal schooling. In most educational systems this would be below lower secondary level. It is well established in the cultural reproduction literature that parental education is a strong predictor of children's educational attainment (Evans & Kelley, 2002). Low educational resources in the household means that refugee children are unlikely to beneft from any cultural and social capital in the home environment for their education.

The overall level of English profciency of the household is also another important aspect shaping the integration of refugee children. A potential risk factor for many refugee children is low English profciency among co-resident adults. Four questions on English language ability were asked in the BNLA survey for all adult and children respondents: how well they understand, speak, read and write English on a scale from 1 (not at all) to 4 (very well). We developed two measures of English profciency: fuency and literacy. We took the average scores of understanding and speaking to measure English fuency, and those of reading and writing for English literacy. Therefore, a higher score indicates a higher level of English fuency and/or literacy.

Table 5.1 shows that on average, refugee children live in households where the average English profciency is about 1.8 or "not well". The low English profciency of refugee parents (or co-resident adults) in addition to having low levels of formal education, will make it diffcult to for them to succeed in education and subsequently in the labour market. This is because these adults will not be able to provide assistance with refugee children's school work. Taken together, these household characteristics will have profound implications for the integration of refugee children.

A further risk factor for refugee children is the low level of labour market participation of household members such as the refugee adults. In our sample, refugee children lived in households where adults were employed only 12.9% of the time across the survey years. There is also a small gender difference as girls tend to live in households with slightly lower labour market participation levels (12.2%) than boys (14.1%). Socioeconomic disparities in households could have long term implications for children's economic wellbeing.

It is likely that low education compounded with poor English profciency contribute to the low levels of labour force participation among co-resident adults. This 'domino' effect of multiple disadvantages poses enormous challenges for refugee parents to integrate into the Australian society themselves. This may indicate that adult refugees have limited ability to seek professional help. They are also unlikely to have informal social networks that can provide information about schools for their children. Given the low rates of paid employment among refugee adults, they will have little choice but to resort to welfare and state benefts. Unemployment and low income, in turn, could lead to poor integration for the entire family and above all no or few resources to support their children's education and healthcare needs.

#### **Gender and Refugee Children's Integration and Wellbeing**

Given the diverse backgrounds of refugee children, their integration pathways are likely to vary by subgroup. In this section, we will examine a number of indicators of refugee children's wellbeing and integration by gender. We focus on children's English language fuency and literacy and indicators of their mental health and wellbeing, such as SDQs (Strength and Diffculty Questions), and life satisfaction.

#### *Language Profciency: Fluency and Literacy by Gender*

Refugee children in the BNLA sample have far higher English profciency than their parents or co-resident adults, with an average score of 3 for fuency and 3.1 for literacy, compared to 1.8 for their parents. Girls also have slightly higher fuency and literacy scores achieving 3 and 3.1 compared to boys with a score of 3 for both outcomes. Despite some gender differences, we fnd that all refugee children surpass the English profciency of their migrating unit, showing clear signs of intergenerational progress in English language profciency.

#### *Socioemotional Wellbeing by Gender*

To understand refugee children's social-emotional wellbeing, BNLA asks a series of questions using the standard Strengths and Diffculties Questionnaire (SDQ) and Total Diffculties Score for Children. SQD is a multidimensional behavioural

screening questionnaire that asks a series of questions about a child's behaviour for individuals aged 3–16. The questions cover emotional symptoms, inattention, peer relationship problems, prosocial behaviour, and conduct problems and are ranked in a score. BNLA contains both a self-assessed SDQ completed by the child and a separate SDQ questionnaire completed by the parent based on their assessment of the child. Total Diffculties is measured using a score ranging from 1 to 40 with a higher score indicating greater diffculty in their adjustment.

Table 5.1 indicates that the average SDQ score for all refugee children in our sample is 10.1. However, when we disaggregate this by gender, we fnd that refugee girls have a much higher SDQ score (13.5) compared with boys (7.7). When we examine children's SDQ scores as assessed by their parents, we see a similar gendered pattern with parents reporting higher SDQ scores for their daughters (10.4) compared with sons (8.8). While a SDQ score under 15 is considered "normal", higher scores among refugee girls would suggest they need more support for their developmental needs and social-emotional wellbeing.

Another aspect of children's wellbeing is life satisfaction. When children were asked "how happy with how things are in life for me right now", 24% of refugee girls strongly agreed compared with 38.2% of boys. By contrast, 11.9% of females strongly disagreed with the statement compared with only 4.2% of males. In general, refugee girls appear to show worse wellbeing than their male counterparts. It is possible that adolescent girls in general experience more psychosocial and socialemotional problems. It could also be that refugee girls fnd it more challenging growing up in a new country and adapting to western cultural norms and gendered roles.

#### **The Enduring Effects of National Origin and Children's Outcomes**

Thus far, we have considered the characteristics of refugee children and their migrating units and found notable differences by gender. To further understand the diversity of this group, we examine refugee children's outcomes by national origin.

#### *Premigration Experiences*

To begin, we examine how premigration experiences differ by national origin to understand refugees' different premigration journeys and pathways prior to their arrival in Australia. Australia's Humanitarian program comprises an on-shore (asylum seekers/asylees) and an off-shore (refugees) component, depending on where the application for a humanitarian visa was processed. Whether immigrants arrive on-shore or off-shore will structure their postmigration experiences, including whether they spend time in detention or in a refugee camp, whether they are on temporary visas, and the level of uncertainty and precarity they experience.

Our data show notable national origin differences in refugee children's migration pathway. In particular, all Iraqi and Afghani refugee children in our sample arrived via the on-shore pathways or were asylum seekers. In contrast, Iranian (97.3%), Central Asian (93.5%), and Middle East/North Africans (89.3%) are more likely to secure their refugee status prior to arriving in Australia. This suggests that Iraqi and Afghani refugees were likely to have endured more precarious statuses and limited access to resources while awaiting confrmation of their refugee status. Asylum seekers often fnd it impossible to make plans for the future given their temporary status. This will negatively affect their ability to invest in human capital and their labour market mobility. Not only has research found a negative association between lengthy asylum processes and psychopathology (Laban et al., 2005), uncertainty can have profound detrimental effects on the psychosocial development of refugee children.

We fnd that some groups are much more likely to spend time in a refugee camp prior to arrival. Consistent with their premigration pathway, those who arrived in Australia as refugees are also more likely to have spent time in a refugee camp. For instance, 97.3% of Iranians arrived with their refugee visas and 29% spent time in a refugee camp before they arrived. Likewise, 93.5% of Central Asians arrived with their refugee status and 63% of them had spent time in refugee camps before they arrived in Australia. In contrast, 100% of Iraqi and Afghani were asylees and only 1.4% and 2.7% spent time in a refugee camp prior to their arrival in Australia. In sum, Afghani and Iraqi refugee children will have likely experienced more uncertainty than refugee children from MENA countries, Asia, and Iran. The BNLA does not record the length of time spent in refugee camps. Whether in Kenya, Jordon, Pakistan or Greece, most camps are run by NGOs funded by UNHCR. Many of these temporary accommodations are basic tarpaulin tents or makeshift huts with poor insulation. Inadequate access to running water, no electricity and unhygienic toilets, compound with limited onsite healthcare facilities make this transition a 'living hell' (Guardian, 2020). Compared to the uncertainty onshore asylum seekers have to endure, refugees who have spent time in camps are also subject to considerable precarity. The tolls of living in camps on their physical and mental health can be signifcant and far-reaching (Fig. 5.1).

#### *English Language Profciency by National Origin*

Most refugee children arrive from non-English speaking countries. As reported above in the analysis by gender, despite the low English profciency of household adults, most refugee children understand and speak English "well". Our results disaggregated by country of origin also show similar patterns of English profciency among migrating units with immigrant adults from Middle East/North Africa showing the highest English profciency and those from Central Asia showing the lowest.

**Fig. 5.1** Migration pathway by country of origin. (Source: Building a New Life in Australia survey, waves 1–5)

**Fig. 5.2** Children's and migrating Unit's English language literacy by gender and country of birth Note: English literacy scores range from 1 to 4 with a higher value indicating greater English literacy. Source: Building a New Life in Australia survey, waves 1–5

This may be due to the fact that English is widely spoken in Egypt and Syria but not in Myanmar or Nepal. Nonetheless, Fig. 5.2 shows lower group variation by origin country in English profciency among refugee children compared with coresident adults.

Overall, we fnd that refugee children have a solid command of English profciency. Most groups, with the exception of Central Asians, have an average English profciency score of 3 or higher. Despite the fact that in most cases English is not their frst language, this high level of English profciency among refugee children is

**Fig. 5.3** Average education and % employment by national origin Source: Building a New Life in Australia survey, waves 1–5 Note: Education scores range from 0 to 4 with a higher value indicating higher average education

encouraging. It bodes well for their integration given the importance of English skills in progressing in the Australian schooling system and the labour market.

As noted in Table 5.1, girls have slightly higher English language fuency and literacy than boys. However, when we disaggregate language fuency and literacy by gender and country of birth, the gender differences appear smaller and are more dependent on the country of origin. For instance, in Fig. 5.6, among Afghani children, girls shower higher English literacy (3.1) than boys (2.9), but among Middle Eastern/North African refugees, the reverse is true, with girls showing lower English literacy (2.9) than their male counterparts (3.1). Our data show that the gap between refugee children and their migrating unit is far wider in literacy, indicating that refugee children have been making great strides in English language since arriving in Australia.

#### *Household Structure and Family Size by National Origin*

It has long been established that family resources are crucial for the educational success of children especially when they are younger (Mare, 1980). In this section we examine how refugee children's family background and household characteristics differ by national origin. We focus on the size, human capital resources, and socioeconomic and linguistic characteristics of their household adults.

We fnd that family and household resources differ by national origin. The frst key difference is in the level of human capital that they arrive with. In Fig. 5.3, we examine the average education level of the migrating unit by national origin. Overall, the average education of the migrating unit varies widely by national origin though all are quite low. No group shows an average education with an equivalent of a high school certifcate. Refugee children from Afghanistan and Central Asia grow up among co-resident adults in the migrating units with the lowest average education. The highest average education of Afghani and Central Asian migrating units is less than elementary school. The migrating units with the highest education are from Iraq and the Middle East/North Africa where adults achieved somewhere between 7 and 11 years of schooling.

Figure 5.3 also shows the employment profles of co-resident adults, measured by the percent of time in paid employment of adults in the migrating unit. We created a summary measure by using the current employment question ("are you employed in paid work?") for each wave and averaging the time that adults were employed over the 5 survey years. Overall, we fnd that refugee children from Middle East/North Africa live in migrating units with the highest levels of adult employment where co-resident adults were employed about 58% of the time. In contrast, this is much lower among refugees from Iran (16.7%), Central Asia (14.7%), Iraq (6%), and Afghanistan (6%). The low levels of adult employment from migrating units from Iraq and Afghanistan suggest that refugee children from these groups may be particularly disadvantaged from their family's lower employment level and lower socio-economic resources. Being dependent on state benefts and the struggle to make ends meet could have a detrimental impact on refugee children's psycho-social developments as well as educational attainment. In sum, our fndings indicate that refugee children from Afghanistan and Central Asia are particularly disadvantaged as indicated by their households with low levels of employment and formal education.

#### *Mental Health and Wellbeing by National Origin*

We also consider how children's socioemotional characteristics differ by national origin. In Fig. 5.4 below, we show the SDQ scores reported by child respondents and their parents. Except for Iranian refugee children, we fnd that parents typically rate their children's SDQ as lower, indicating fewer socioemotional problems, than children's self-assessment. This may suggest that refugee children experienced greater diffculties than their parents were aware of. We also fnd variation in SDQ scores by national origin. Overall, we fnd that MENA refugee children show the lowest SDQ scores, which is confrmed by both the self-assessed (6) and parental assessment scores (5.7). We fnd that Afghani (9.9) and Iraqi (9.2) refugee children show the highest self-assessed SDQ scores whereas Irani (9.2) and Iraqi (8.6) refugee children show the highest parent-assessed SDQ scores. The disparities between parental- and self-assessed SDQ scores may be causes of concern. This may be

**Fig. 5.4** Strength and diffculties questionnaire total diffculties scores (Self-assessment and parental assessment) by Country of Birth. (Source: Building a New Life in Australia survey, waves 1–5)

**Fig. 5.5** Percent excellent self rated health by national origin. (Source: Building a new life in Australia survey, waves 1–5)

indicative of the challenges that refugee children encounter during their integration progress and the extent to which this is evident to their parents. Additionally, while it is common for children to experience intergenerational conficts with their immigrant parents, this may be exacerbated in refugee families who on average have endured greater hardships during their migration journey (Portes & Rumabut, 2014).

Another measure of subjective wellbeing in the BNLA is refugee children's selfrated health. Again, we examine the details by national origin. Figure 5.5 shows large between-group variations in self-rated health. Among refugee children, those from MENA countries rank the highest in self-rated health, with 75% rating their health as 'excellent'. We fnd a large drop off though, as only 35% of Afghani children rank their health as 'excellent' followed by 42.7% of Iranian children. Central Asians show the lowest as only 29% rank their health as 'excellent.' Both groups of refugee children reporting excellent health are from other Africa and MENA countries yet the former are all asylees and 90% of the latter obtained refugee status before arriving. It is unlikely that migration route or having spent any time in refugee camps is associated with self-rated health.

#### *How National Origin and Gender Matter for Refugee Children's Outcomes*

In this fnal section, we consider the extent to which national origin and gender together may shape wellbeing. It would be ideal to examine how refugee children's outcomes vary by gender and national origin but small cell sizes on many of the outcomes prevent us from doing so. Our descriptive results in Table 5.1 showed some disadvantages for girls with a higher proportion of refugee girls living in refugee camps and reporting PTSD than refugee boys. However, disaggregating by gender conceals some interesting patterns of disadvantage for some groups. Even by simply disaggregating the fgures by gender and national origin, we can see refugee children from Iraq suffered from a much higher level of PTSD than their peers in other countries. Equally, a staggeringly high proportion (83.4%) of refugee girls from MENA countries (Syria and Egypt) have reported suffering PTSD. The same pattern is for refugee boys in other African countries. These conditions are likely to be triggered by deeply traumatic experiences associated with violence and conficts in both the Middle East and certain African countries (Fig. 5.6).

#### **Discussion and Conclusion**

Overall, we fnd that refugee children are outperforming their parents. Despite living in disadvantaged households, refugee children are doing well and making intergenerational progress. However, we fnd some major differences by gender and national origin. Our analyses emphasise the importance of national origin, gender, and pre-migration characteristics in understanding the settlement experience and integration of refugee children in Australia. Our fndings show signifcant differences between boys and girls as well as origin country differences in the range of outcomes examined.

In terms of health and mental wellbeing, girls have higher SDQs and lower selfrated health than boys. Refugee children from MENA countries appear to fare better

**Fig. 5.6** Percentage with PTSD by gender and country of birth. (Source: Building a new life in Australia survey, waves 1–5)

compared to their peers from Central Asia, Iraq or Afghanistan. They tend to report lower SDQ, as do their parents. MENA refugee children also report better health status and higher levels of English Fluency, and to a lesser degree English literacy. The evidence thus far seems to support the notion that prolonged confict is associated with long-term consequences among refugee children. The MENA countries in our sample are Egypt and Syria were in confict for 5–6 years during the period of 1960–2009. This stands in stark contrast with Iraq and Iran where most of the population were subject to decades of war and atrocities.

Since most of the refugee children in BNLA have not completed their schooling, we are unable to assess their educational attainment. After linking the child questionnaire to the main survey, particularly to key variables such as national origin, the large amount of missing data prevented further regression analysis of any particular outcomes of interest.

#### **Policy Recommendations**

Building a new life in a new country requires resilience and resources. Refugee children are not a problem that needs to be fxed. We should also move away from the discourse of identifying the 'good' or 'deserving' refugees so that we can only 'let the right ones in'. Most refugee children settle in the neighbouring country next to their home country because the majority are far too poor or lack the resources to travel to wealthy countries in the west. Considering nearly half of the refugee children in BNLA experienced PTSD, it would be diffcult to overstate the scale of the

ordeal they have been subjected to throughout their treacherous journey getting to Australia. They may have lost their parents, been out of school for years, subjected to sexual violence, forced to become child soldiers in sectarian conficts or simply displaced. Integrating these young people into the wider society requires a better understanding of the ways in which their pre-arrival experiences shape their future. Refugee children come from extremely diverse backgrounds and we argue that understanding the national context of their country of origin would greatly help towards a more targeted approach in policy intervention.

First, we know that girls have disproportionately suffered from being out-ofschool in Afghanistan, Iran, and Iraq. It would be diffcult to quantify the amount of lost schooling due to the extremely volatile conficts these children experienced. Refugee children regardless of their gender, would need extra resources in schools to support their learning and integrate them into a new educational environment. Importantly, resources would also be needed to support families and migrating units where, as our fndings show, refugee children are living with poorly educated coresident adults with poor command of English who are largely not in paid employment. Targeted policy initiatives aimed at supporting co-national social networks may strengthen the human capital outside the households of these refugee migration units. Many highly-educated Iranians and Egyptians migrated to Australia but not via humanitarian routes. These co-national communities serve as a readily available resource that government and NGOS can harness to provide the much-needed support for refugee children and their families.

Second, concerning mental health and wellbeing, evidence has shown that women are more likely to suffer from depression than men in the general population and especially among refugee women (Kuehner, 2016). They are also signifcantly more likely to report poorer physical as well as mental health (Cheung & Phillimore, 2017). Our fndings on refugee children support this. Not only were refugee girls more likely to have been in refugee camps and to report PTSD, they were also less likely to rate their health as excellent and report higher levels of diffculties compared to refugee boys. This could be due to the fact that women are more likely to suffer from sexual violence while feeing their countries. The renewed conficts between Ethiopia and Eritrea are a chilling reminder of the horrifc sexual violence suffered by displaced women and girls. Without early interventions and appropriate programmes of support between schools and healthcare providers, refugee girls may experience poorer mental health and wellbeing well into their adulthood and for years to come.

Third, we expect refugee girls from Muslim countries will face additional barriers navigating their way through their new life in Australia, where cultural expectations are radically different from those in their home country. Unlike single-sex schools in some Muslim countries, most schools in Australia are co-educational where boys and girls study together. Yet their refugee parents may have a very different set of cultural expectations regarding gender roles in the home and in public spaces. Refugee girls would need to adapt in a very short space of time to negotiate conficting expectations in the new society. While research has found that social networks such as religious groups and co-national organisations can act as a source of support for better mental health (Bakker et al., 2016), they can also act as the "guardian" of more conservative gender roles and in turn cause additional tension to the integration of adolescent refugee girls.

Fourth, we know that Iraqis tend to settle with their established communities. Co-nationals tend to look out for each other and share information from schools to employment opportunities. It would make sense for integration programmes to target intervention in supporting and strengthening these community organisations. Our evidence points to a persistent disadvantage of refugee children from Central Asia, Iraq, and Afghanistan.

Policies and programmes need to be comprehensive enough to promote structural integration to enable refugee children to transition to a full member of the Australian society in all aspects: stable and quality housing, accessible and affordable healthcare, culturally sensitive education, training and pathways to employment including language programmes. At the same time, policy intervention needs to be targeting resources to the groups with most acute disadvantages: girls from Afghanistan, Iraq, and Central Asia.

Refugee children in the BNLA sample are the 'lucky ones'. Babies who were in offshore processing centres on Pacifc Islands Nauru or Manus in Papua New Guinea do not have any rights of entry because they were not born in Australia even though their parents were granted refugee status. These offshore processing centres are considered by many as detention centres. Since July 2013, the Australian Government no longer accepts asylum seekers by boat from offshore processing even though they are recognised as refugees. The destiny of these stateless refugee children are likely to be much grimmer than the ones who made it onshore. Much more needs to be done to urgently review such polices so that hundreds and thousands of these children can also have a chance to build a new life in Australia.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 6 Adolescence a Period of Vulnerability and Risk for Adverse Outcomes across the Life Course: The Role of Parent Engagement in Learning**

**Jenny Povey, Stefanie Plage, Yanshu Huang, Alexandra Gramotnev, Stephanie Cook, Sophie Austerberry, and Mark Western**

Deep and persistent disadvantage in the educational outcomes of students continues to concern developed nations such as Australia, the United States, and the United Kingdom (Brownstein, 2016; Thomson, 2013; Weale, 2016; Wilson et al., 2015). Students from socio-economically disadvantaged backgrounds are more likely to underachieve in school and face an elevated risk for a variety of adverse outcomes across their life course such as delinquent activity and offending, substance misuse, social exclusion and isolation, early parenthood, mental health illnesses, and underemployment (Hatch et al., 2007; Henry & Huizinga, 2007a, b; OECD, 2013; Rocque et al., 2017; Stranger, 2002). In this chapter we apply the life course principle of 'linked lives' to investigate how parents' engagement in their adolescent child's education and the school environment are related to mental health, specifcally anxiety and depression, self-concept, and educational expectations. We draw on data from a general population study and qualitative interviews to explore the interconnected nature of parents, schools, and socio-economic context in infuencing adolescents' aspirations and psychological wellbeing across early, middle, and late adolescence.

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia

e-mail: j.povey@uq.edu.au; yanshu.huang@uq.edu.au; m.western@uq.edu.au

S. Plage

A. Gramotnev · S. Cook · S. Austerberry

J. Povey (\*) · Y. Huang · M. Western

Australian Research Council Centre of Excellence for Children and Families over the Life Course, School of Social Science, The University of Queensland, Brisbane, QLD, Australia e-mail: s.plage@uq.edu.au

Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia e-mail: a.gramotnev@uq.edu.au; s.cook@uq.edu.au; s.austerberry@uq.edu.au; m.western@uq.edu.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_6

#### **Background**

There is considerable evidence that parent engagement in their child's learning has positive effects on student achievement (Benner et al., 2016; Castro et al., 2015; Hill & Tyson, 2009). Parent engagement may be defned as the engagement of parents or primary carers in education-related activities that are expected to foster academic achievement and the social and emotional wellbeing of children (Fishel & Ramirez, 2005). Parent engagement is a multidimensional construct including home-based parent engagement, school-based parent-engagement and academic socialisation (Epstein & Sanders, 2002a, b; Fan & Chen, 2001; Grolnick & Slowiaczek, 1994; Hill & Tyson, 2009; Wang & Sheikh-Khalil, 2014). School-based engagement includes drawing on parent expertise and two-way communication with teachers. Home-based engagement includes assisting with schoolwork and fostering other learning opportunities in everyday activities. Academic socialisation is the process through which parents' foster high educational aspirations and expectations in their child. It involves communicating the importance and value of education (including showing interest in their child's learning and education) and scaffolding a child's decision-making and future planning capabilities. A meta-analysis conducted by Hill and Tyson (2009) found that academic socialisation had the strongest positive association with student achievement compared to the other two components of parent engagement. While most research focusses on parent engagement during the primary school years, its positive effect on student achievement continues during adolescence (Hill & Tyson, 2009; Hill & Wang, 2015; Hill et al., 2018; Gordon, 2016; Gordon & Cui, 2012; Wilder, 2014).

Adolescence is a critical developmental stage where physiological changes, changes in the school environment, and changes in relationships with peers and parents occur. These transitions may challenge adolescents' sense of identity, and thus adolescence is a vulnerable period (Alsaker & Kroger, 2006). The impact of adolescent challenges on adult outcomes has only been recently noted (Due et al., 2011), with many scholars acknowledging adolescence as a 'turning point' where life trajectories may be especially malleable (Johnson et al., 2011), and where the impact of previous life stages may combine with adolescent experiences to infuence outcomes in adulthood (Pollitt et al., 2005). While previous life stages infuence identity formation, in adolescence, education and occupational goals are integral in shaping who adolescents are and would like to be (Flammer & Alsaker, 2006).

Adolescence is also where, for some students, we observe a decline in student engagement with education (Archambault et al., 2009; Fredricks & Eccles, 2002; Lam et al., 2016; Trautwein et al., 2006) and academic performance (Hill & Tyson, 2009). Many adolescents also experience declining self-concept (Nagy et al., 2010) and increased anxiety and depression (Merikangas et al., 2010; Paus et al., 2008; Twenge et al., 2019) linked to extensive physiological and psychosocial transformations and vulnerabilities (Blakemore, 2012; Casey et al., 2008). School (Verhoeven et al., 2019) and family (Beyers & Goossens, 2008; Schachter & Ventura, 2008) contexts play important roles in mitigating these challenges and shaping identity formation.

Key aspects of school context that matter for adolescent identity formation and academic and wellbeing outcomes include school belonging, peer connections and bullying (Gillen-O'Neel & Fuligni, 2013; Bond et al., 2007; Moore et al., 2015). School belonging is 'the extent to which students feel personally accepted, respected, included, and supported by others in the school social environment' (Goodenow & Grady 1993, p. 80). Higher levels of school belonging are associated with lower levels of school-dropout (Gillen-O'Neel and Fuligni (2013). Peer connections at school are positively associated with lower rates of anxiety and depressive symptoms, drug use, and disruptive behaviour in adolescents (Bond et al., 2007) while, bullying or peer victimisation amongst school peers adversely impacts outcomes such as school completion and employment (McDougall & Vaillancourt, 2015; Moore et al., 2015).

Parenting styles play an important role in adolescence, bolstering psychological wellbeing and academic success. Parenting styles such as authoritative parenting (Baumrind, 2005) or autonomy-granting parenting (Darling & Toyokawa, 1997), that foster greater independence while setting and maintaining realistic discipline and rules, are related to adolescents' positive self-evaluations and selfconcept (Cripps & Zyromski, 2009) and lower depression (Yap et al., 2014). In contrast, parenting styles more focused on enforcing higher expectations, sometimes with punitive approaches with a lessened focus on independence, such as authoritarian or demanding parenting, are related to increased anxiety (Yap et al., 2014) and poorer school achievement (Pinquart, 2016). Finally, parenting styles that are sensitive to, and assist with, the emotional needs of adolescents—such as warm or responsive parenting—are related to lower anxiety and depression (Yap et al., 2014).

Despite the positive benefts of parents engaging in their child's education, such as improved academic achievement and wellbeing, parent engagement in learning progressively declines during primary school (Cheung & Pomerantz, 2011; Shumow & Schmidt, 2014). This may be due to a growing emphasis on student autonomy as students progress in school, and also because parents may feel less able to support their child academically in secondary school (Jensen & Minke, 2017). Parent engagement declines substantially in secondary school and for some students may be non-existent (Daniel, 2015; Green et al., 2007; Napolitano, 2013; Seginer, 2006; Spera, 2005). These declines are steeper among families from socio-economically disadvantaged backgrounds (Bridgeland et al., 2008). However, parent engagement in adolescence is critical for educational success and long-term outcomes. Kim and Hill (2015) found mothers' engagement in their child's learning during secondary schooling to be more strongly associated with academic performance than parent engagement during primary school. A lack of parent engagement in their child's learning and education during adolescence is also associated with poorer mental health outcomes in adulthood (Bakoula et al., 2009; Mensah & Hobcraft, 2008; Westerlund et al., 2015).

Notwithstanding the evidence on the positive effects of parent engagement, signifcant gaps in our knowledge remain. Most studies focus on early childhood or use cross-sectional designs that do not account for the high variability in both the child's development and the parent-child relationship over time. Adolescents undergo many developmental changes and longitudinal methods better account for the within- and between-individual variability that this unpredictable and turbulent period brings. Very few studies also examine how parent engagement with their adolescent child's education particularly in early and middle adolescence infuences mental health, self-concept and educational aspirations or expectations. Further, most studies examining the role of parent engagement on student's educational outcomes use samples from general populations. Very little research examines how parent engagement in learning can support students from disadvantaged backgrounds. Finally, almost no research examines parent engagement and aspirations of school students in late adolescence who are living in socio-economically disadvantaged circumstances with complex needs. In addition to experiencing a combination of economic hardship, housing instability, physical or mental health conditions, learning diffculties, and language diffculties, some of these young people experience domestic and family violence and may be living away from their nuclear families.

We use the 'linked lives' principle to explore each aspect of parenting and engagement in their adolescent's education and dimensions of the school context to explore how adolescents draw upon their social environment to develop future aspirations and maintain their mental wellbeing. The chapter is in two parts. Part 1 uses a general population sample to examine parent engagement in early and middle adolescence drawing on analyses using the *Longitudinal Study of Australian Children* (LSAC) K cohort sample. Part 2 explores the complexities of parent engagement in student learning in late adolescence and how this infuences educational and occupational aspirations and the stressors for mental health among students from highly disadvantaged backgrounds. This section draws on data from qualitative interviews with senior secondary school students (grades 10–12) who experience social disadvantage across multiple dimensions, including housing instability, contact with the child protection system, contact with the youth justice system, early parenthood, learning disability, or being part of a cultural or linguistic minority. These interviews come from the *Learning through COVID-19* study funded by the Paul Ramsay Foundation (McDaid et al., 2021). This approach allows us to extend generalizable fndings about associations between parental engagement and student outcomes with new insights about how disadvantaged students experience parent engagement during remote learning and how this infuences their aspirations and wellbeing.

#### **Early and Middle Adolescence**

In early and middle adolescence, adolescents are still strongly infuenced by their family context, including parent engagement in their education and parenting styles. Adolescents are also affected by their school setting through peer connections, feelings of school belonging and bullying.

#### *The Study*

Using the *Longitudinal Study of Australian Children* (LSAC) K cohort sample, we explore the associations of parent engagement in their child's learning and education with ability self-concept, mental health, and educational aspirations. These analyses allow for an examination of parent engagement and key family and school factors across two time points in early and middle adolescence. The wave fve data were collected in 2012 and 3956 adolescents aged 12–13 years took part in the study. The wave six data were collected in 2014 and 3537 adolescents aged 14–15 years took part in the study. Only adolescents who answered the relevant questions were included in the sample. Forty-one parents from wave fve and 162 from wave six, did not give consent for their child to answer questions about stress, anxiety, and emotions and 9 adolescents from wave 5 and 24 from wave 6 chose not to answer these questions.

For our analyses examining parent factors, we use measures of parent engagement in their adolescent's education and adolescents' perceptions of their parenting styles. Combined parent engagement in their adolescents' education is measured using the combination of adolescents' ratings of their mothers' and fathers' interest in their education, rated from "no interest in [their education] at all" to "a lot of interest." Parenting styles were rated by adolescents using the Parenting Style Inventory II (Darling & Toyokawa, 1997) and scale means of responsive parenting, demanding parenting, and autonomy-granting parenting were constructed.

To examine the school environment, we use measures of peer connection problems, experiences of bullying in the last month, and feelings of school membership. Peer connection problems—i.e., diffculty getting along with and connecting with peers—uses fve items from the peer problems subscale of the Strengths and Diffculties Questionnaire (Goodman, 1997) and was reported by the adolescent themselves. Example items include "picked on/bullied by children" and "has been solitary". Experiences of bullying in the last month was reported by the adolescent and constructed into a binary categorical variable of any experience of bullying, using six categorical questions asking about specifc acts of bullying (e.g. "hit/ kicked", "threatened to take my things", "said mean things/called me names"). School membership was measured using the Psychological Sense of School Membership (Goodenow, 1993) scale, comprising twelve items assessing adolescents' experiences of peer and teacher relationships at school. Example items

included "teachers are interested in me" and "other students accept my opinions" and adolescents rated how accurate each statement was to their experience at school.

#### *Self-Concept*

In adolescence thinking becomes more abstract, and self-refection and selfawareness increase. Adolescence is potentially a time of self-concept diffculties, where one's sense of self develops substantially (Sebastian et al., 2008). Adolescent development also includes adapting to or coping with new tasks including becoming more emotionally independent of parents and other adults, making plans and preparations for a future occupation, and developing values to guide behaviours (Seiffge-Krenke & Gelhaar, 2008).

While related to self-esteem and identity, the construct 'self-concept', measures 'the perception of oneself, including one's attitudes, knowledge, and feelings regarding abilities, appearance, and social relationships' (Reynolds, 1993, p. 20). Self-concept is multidimensional meaning an individual can have 'multiple interrelated self-concepts in a range of domains' (Mercer, 2012, p. 11). For example, ability or academic self-concept is a broad domain that includes perceptions of abilities more generally as well as specifc learning areas (e.g., reading, math, music). The effects of self-concept are stronger when specifc to the learning domain, meaning that ability or academic self-concept may be different for the domains of mathematics and reading skills (Arens et al., 2011; Valentine et al., 2004).

To assess adolescent self-concept, we use a short-form measure of academic ability self-concept (Marsh, 1990), measured at waves fve and six of LSAC. This measure comprised of three items about their academic ability at school and adolescents were asked to rate their agreement with each statement: "I learn things quickly in most school subjects", "I'm good at most school subjects", and "I do well in tests in most school subjects." We used the scale mean to create a composite measure of academic ability self-concept, with higher values indicating more positive beliefs about academic ability.

#### **Why Self-Concept Is Important for Student Outcomes**

Numerous studies have found that self-concept in adolescence predicts higher levels of educational attainment (Guay et al., 2004; Valentine et al., 2004; Vargas et al., 2015; Wigfeld et al., 2006) and academic achievement (Susperreguy et al., 2018). Eccles (1987) theorised that positive self-evaluations can foster children's expectations of future success in academics. The 'reciprocal effects' model acknowledges that academic self-concept and achievement both affect each other, or are mutually reinforcing (Seaton et al., 2015). Positive academic self-concept has been found to predict educational attainment level 10 years later, while controlling for academic achievement, family structure and family socio-economic status (SES; Guay et al.,

2004). A meta-review of longitudinal studies further demonstrated a signifcant, although small, effect size between positive self-belief and academic achievement when controlling for prior achievement (Valentine et al., 2004). This relationship was stronger when assessing self-belief within the academic domain, and when measures corresponded to this domain (e.g., by subject area) (Valentine et al., 2004). In addition to academic performance, academic self-concept also infuences career related choices (Wigfeld et al., 2006).

However, academic self-concept may be more likely to decline during adolescence than other domains of self-concept. Shapka and Keating (2005) measured the differences in domains of self-concept between grade 9 and grade 10, using the Harter Self-Perception Profle for College Students which measures general selfworth, and 12 sub-scales of self-perceived competence in various physical, social and cognitive domains. They found that most measures of self-concept increased with age, except for scholastic competence, which 'declined over the course of high school' (p. 88).

#### **Parent Factors and Self-Concept**

Given that adolescents typically experience a decrease in their academic selfconcept, and that academic self-concept is linked to long-term educational attainment, it is important to consider what may act as a protective factor to maintain academic self-concept during adolescence. Parent engagement in education improves student perceptions of their self-worth and self-effcacy (Cripps & Zyromski, 2009). One pathway that may account for the positive effect of parent engagement on student outcomes is the effect of parent engagement on self-concept as self-concept is often framed with reference to parents and peers.

Examining various dimensions of self-concept, Cripps and Zyromski's (2009) review found that parent involvement in their child's education can positively or negatively affect adolescents' self-concept depending on parenting style. Authoritative parenting led to more positive self-concept and higher levels of intrinsic motivation for learning (Cripps & Zyromski, 2009) than authoritarian and permissive parenting styles. In other words, parents who provided both structure and opportunity for independence fostered positive self-concept and greater internal motivation for learning. Similarly, in a study that examined the relationship between parental involvement, growth-fostering relationships, and self-concept in adolescents, perceived parental involvement signifcantly contributed to a positive selfconcept (Gibson & Jefferson, 2006).

We utilise a panel fxed-effects regression model to account for the high variability in both the child's development and the parent-child relationship over time when examining parent effects on adolescent academic ability self-concept. As shown in Fig. 6.1, several parental factors are signifcant predictors of academic ability self-concept. Firstly, parent interest in education—namely the parent engagement dimension relating to home-based engagement—had a positive association with student academic ability self-concept, even when accounting for other

**Fig. 6.1** Relationship between self-concept and factors in the family and school contexts Notes: Coeffcients from a panel fxed effects regression model adjusted for gender and Indigeneity. All black bars in the fgure are statistically signifcant at p < 0.05 and grey bars are not statistically signifcant. (Source: The Longitudinal Survey of Australian Children (Waves 5 and 6) sample (observations n = 6585) (individuals n = 3727))

parenting, school, and sociodemographic factors. Turning to other parenting factors, responsive and autonomy-granting parenting were both associated with signifcant increases in positive academic ability self-concept. These fndings resonate with past research which suggests that adolescents receiving greater engagement with their education from their parents and experiencing parenting styles that fosters greater self-reliance as well as providing positive feedback on academic performance (Cripps & Zyromski, 2009; Putnick et al., 2008) have more positive beliefs about their abilities.

In addition to psychosocial parenting factors, the fndings also highlight how family composition can infuence adolescent self-concept. Specifcally, the results presented in Fig. 6.1 shows that living in a blended family is associated with a lower academic ability self-concept than living with two biological parents. Previous research with US adolescents has found that, after controlling for key demographics such as gender, race and socio-economic status, family structure is the strongest predictor of academic achievement (Jeynes, 2005). This is the strongest predictor in the model and suggests the infuence of family structure extends to academic ability self-concept as well as academic achievement.

#### **School Factors and Self-Concept**

Self-concept has been theorised as a mechanism that explains the relationship between peer bullying and academic achievement, given its strong association with both academic outcomes and bullying (Roeleveld, 2011; Valentine et al., 2004). Cross-sectional research suggests that bullying victimisation infuences academic achievement indirectly through academic self-concept for adolescent girls (Jenkins & Demaray, 2015). Furthermore, a meta-analysis examining the associations between self-concept and academic achievement found a small effect of positive self-belief on academic achievement that was stronger for self-beliefs specifc to the academic domain (Valentine et al., 2004). Quality relationships with peers also exerts an important positive infuence on adolescent's general self-concept (Hay & Ashman, 2003). While there is scant research on the relationship between peer connections and academic ability self-concept, it seems plausible that positive peer connection would contribute to strengthened academic self-concept through positive environmental factors, such as social networks, in the school setting (Jenkins & Demaray, 2015). Finally, in terms of the overall social environment, collegial school contexts may facilitate positive self-beliefs about academic ability. Given that selfconcept is infuenced by social factors, such as social comparisons, sense of safety at school and teacher-student relationships, school belonging may also infuence academic or ability self-concept. Indeed, one study found a moderate correlation between school belonging and academic self-concept (Singh et al., 2010).

Turning to our fndings on school factors, as shown in Fig. 6.1, we found that students were more likely to have a poorer academic ability self-concept if they reported more peer connection problems. Unexpectedly, we also found that students who experienced bullying recently had better academic ability self-concept. That is, more positive beliefs about their abilities. This is contrary to fndings which suggest that having a more positive self-concept is related to lower reports of peer victimisation (e.g., Jenkins & Demaray, 2012). It may be possible that those with higher academic ability self-concept are targets for bullying, and that is what is driving this result. Nonetheless, this fnding warrants further investigation. Finally, increases in feelings of school belonging had a small but signifcant positive effect on academic ability self-concept.

#### *Mental Health*

Adolescent mental health is becoming a global concern, with adolescents having the highest prevalence of mental disorders of all age groups (American Psychological Association, 2018; Australian Government, 2020). The global prevalence of mental health disorders affecting adolescents is 10–20% (World Health Organisation, 2020). In Australia 14% of those aged between 12 and 17 suffer from mental disorders, including depression, anxiety, ADHD, and conduct disorder (Lawrence et al., 2015). The prevalence of anxiety disorders is the highest (7%), followed by ADHD (6%) and major depressive disorders (5%) (Lawrence et al., 2015).

Internationally, rates of internalising mental health disorders, such as anxiety and depression are rising (Bor et al., 2014; Gunnell et al., 2018; Mojtabai et al., 2016; Twenge et al., 2019). For example, large and nationally representative studies of US adolescents have shown an increased rate of major depressive episode of those aged 12–17, from 9% in 2005 to 11% in 2014 and to 13% in 2017 (Mojtabai et al., 2016; Twenge et al., 2019). The actual prevalence of mental health disorders is likely to be higher still as most cases may go unnoticed, because adolescents are less likely to seek help or they or their parents may not know how to recognise symptoms (Kessler et al., 2007; Lawrence et al., 2015).

We examine adolescent mental health through experiences of depression and anxiety, measured at waves fve and six of LSAC. To measure depression, we use the Short Mood and Feelings Questionnaire (Angold et al., 1995) which comprise thirteen statements. Example statements include "I felt miserable or unhappy" and "I found it hard to think properly or concentrate." Adolescents were asked to evaluate whether each statement was "not true", "sometimes" true, or "true." The items were summed to create a measure of depression. For anxiety, we use the short-form Spence Anxiety Scale (Spence, 1998), comprising eight items. Example items include "I worry about things" and "I wake up feeling scared." These statements were evaluated by adolescents on a scale from "never" to "always". All items were summed to create a composite measure of anxiety.

#### **Why Mental Health Is Important for Student Outcomes**

The impacts of anxiety and depression on adolescent outcomes across the life course are numerous and include low academic achievement, increased risk of dropping out of school, increased risk of anti-social behaviour and diminished employment prospects (Bernal-Morales et al., 2015; Hatch et al., 2007; OECD, 2013). Leaving school before graduating is associated with precarious employment and a reduction in lifetime earning capacity (Lamb & Huo, 2017). A 25-year longitudinal study of New Zealand children found that the impact of poor mental health in adolescence persisted over the life course and was associated with higher rates of welfare dependence and unemployment (Fergusson et al., 2007). Mental health disorders also put individuals at greater risk of intentional self-harm and suicide (Daraganova, 2017; Skegg, 2005). According to the WHO, suicide is the third leading cause of death in adolescents aged 15–19 years old (World Health Organization, 2020). In Australia over the period 2017–2019, suicide accounted for 37% of all deaths among young people aged 15–24 years (Australian Institute of Health and Welfare [AIHW], 2021a).

#### **Parent Factors and Mental Health**

Numerous studies report that parent engagement in children's education is a protective factor in reducing suicidal thoughts and behaviours (Kang et al., 2017; Kim, 2016; Madjar et al., 2018; Piña-Watson et al., 2014; Tammariello et al., 2012; Wang et al., 2019) and improving adolescent socioemotional functioning (Garbacz et al., 2018). However, few studies have shown how parent engagement in their child's education can improve emotional functioning, for example decreasing anxiety and

depressive symptoms in adolescence (Bireda & Pillay, 2017; Matos et al., 2006; Tammariello et al., 2012; Wang & Sheikh-Khalil, 2014; Wang et al., 2019; Westerlund et al., 2015).

We used panel fxed-effects regressions to account for the high variability in both the child's development and the parent-child relationship over time to examine parent factors and adolescent mental health. For depression in adolescents, we see that parent engagement in their child's education decreases depressive symptoms in adolescence even when accounting for other factors such as school environment and sociodemographic characteristics (see Fig. 6.2). When examining parent factors for anxiety, parent engagement in their child's education decreases anxiety symptoms in adolescence. However, when school factors like school belonging and experiences of bullying are taken into account (as shown in Fig. 6.3), parent engagement in their child's education is no longer statistically signifcant.

Parenting styles are also signifcantly associated with adolescent mental health. Adolescents who reported more responsive parenting also reported lower depression (Fig. 6.2) and anxiety (Fig. 6.3). While depressive symptoms were impacted by parental interest in an adolescents' education, the key parental factor for decreased anxiety was a more general parenting approach that responds more actively towards emotional needs. These fndings partially resonate with meta-analyses on parenting styles and adolescent mental health, particularly for warm parenting. Adolescents who report a positive relationship with their parents and where their parents are engaged in their lives, report lower anxiety and depression, with these benefts to

**Fig. 6.2** Relationship between depressive symptoms and factors in the family and school contexts Notes: Coeffcients from a panel fxed effects regression model adjusted for gender and Indigeneity. All black bars in the fgure are statistically signifcant at p < 0.05 and grey bars are not statistically signifcant. (Source: The Longitudinal Survey of Australian Children (Waves 5 and 6) sample (observations n = 6585) (individuals n = 3727))

**Fig. 6.3** Relationship between anxiety symptoms and factors in the family and school contexts Notes: Coeffcients from a panel fxed effects regression model adjusted for gender and Indigeneity. All black bars in the fgure are statistically signifcant at p < 0.05 and grey bars are not statistically signifcant. (Source: The Longitudinal Survey of Australian Children (Waves 5 and 6) sample (observations n = 6585) (individuals n = 3727))

mental health extending into young adulthood (Clayborne et al. 2021; Gorostiaga et al., 2019; Yap et al., 2014). However, although previous studies found relatively consistent associations between autonomy-granting parenting and youth depression and anxiety, we did not fnd a similar association in this study. Given that depression and anxiety are primarily mood-based health factors, responsiveness to emotional needs may have an immediate restorative impact on mental health, as compared to parenting that fosters greater independence.

#### **School Factors and Mental Health**

In terms of school factors such as diminished peer connections, past research has found peer confict, particularly bullying victimisation, leads to both higher depression and anxiety (Halliday et al., 2021; Hawker & Boulton, 2000; Moore et al., 2017). In contrast, good peer connectedness, such as feelings of belongingness with ones' school, can be associated with improved mental health amongst adolescents. Adolescence is a developmental period marked by transitions away from relying on adults toward peers (Fuhrman & Buhrmester, 1992)—such as classmates—for social support and esteem (McLaughlin & Clarke, 2010). Indeed, a systematic review of longitudinal studies of the impact of school factors on the emotional health of adolescents found that school connectedness was associated with a lower

risk of depression for adolescents, regardless of gender, and anxiety for girls (Kidger et al., 2012).

In our models for adolescent depression (see Fig. 6.2) and anxiety (see Fig. 6.3), consistent with past research, adolescents with greater diffculty connecting with peers had higher depression and anxiety. Adolescents who reported recent experiences of bullying had higher anxiety (see Fig. 6.3) but bullying victimisation was not associated with changes in depression (see Fig. 6.2). Regardless, depression is nonetheless linked to diffculty with connecting with peers, highlighting the importance of positive peer relationships in adolescence for mental health outcomes. Indeed, adolescents who experienced increased feelings of school connectedness also reported lower depression (see Fig. 6.2) and anxiety (see Fig. 6.3). In other words, as adolescents progress through secondary school and become more integrated in the wider student body, these social networks subsequently provide benefts to their mental health.

In sum, adolescent mental health is impacted simultaneously by a multitude of social relationships across two central domains: parental factors and school factors. Parental engagement in their child's education can communicate commitment in an adolescent's future, leading to direct decreases in depression even when accounting for important school factors. In addition, as adolescents expand their available social resources for support, their mental health still benefts from parenting that is responsive to their emotional needs. School factors such as experiencing confict with peers are associated with increased depression and anxiety. This may be mitigated by widening social circles, particularly at school, which also provides an avenue through which youth can access social support. Adolescents who report greater belongingness to their student body see improvements in their mental health as they progress through secondary school education.

#### *Educational Aspirations*

Adolescence is often the time young people start to consider major future goals. A key decision is whether to continue on to further education or to enter the workforce. Across OECD countries, about half of young people transition out of the education system, with the majority of these individuals entering employment (OECD, 2020). Of those who continue onto further education, most will pursue a tertiary qualifcation, but others may choose another form of post-secondary nontertiary option. In Australia, 81% of school leavers completed some form of secondary-level education with 60% enrolled in a form of further study (Australian Bureau of Statistics, 2020). Of our sample from the K cohort in LSAC, approximately 71% anticipate seeking a university degree, with 14% seeking to enter trade or vocational training and 15% not considering further education.

Educational aspirations was measured at wave six of LSAC using a re-coded categorical variable where adolescents were asked "Looking ahead how far do you think you will go with your education?" with the categories "complete secondary school" (reference group), "complete a trade or vocational training course", or "complete a university degree." Adolescents also responded to a binary question where they indicated if they discussed their futures with their parents. This measure was used as a component of academic socialisation.

#### **Why Educational Aspirations Are Important for Student Outcomes**

A wealth of evidence suggests that high educational and career aspirations during adolescence leads to increased educational attainment (Beal & Crockett, 2010; Rojewski, 1999), occupational prestige, and wage attainment in adulthood (Pinquart et al., 2003; Schoon & Parsons, 2002). Furthermore, OECD data suggests that completing post-secondary education, particularly at a tertiary level, is related to higher employment rates, wages, and wage growth later in life (OECD, 2020). These trends highlight the need to understand the motivations and contributing factors to adolescent aspirations for the future, especially as adolescence is a period in which young people develop meaningful understandings of education and future-oriented goals (Anders & Micklewright, 2013; Croll & Attwood, 2013). These educational and occupational goals not only shape who they would like to be, these educational plans have long-term consequences for later life outcomes.

#### **Parent Factors and Educational Aspirations**

Parents can play a substantial role in shaping adolescents' goals for the future. Parenting styles that focus on imposing high expectations on children may lead to increased stress in adolescents, and in turn poorer academic achievement. Alternatively, parenting styles that foster greater independence and trust may be associated with better academic achievement. Indeed, according to a meta-analysis, autonomy-granting and warm parenting was associated with greater academic performance over time (Pinquart, 2016). In contrast, this meta-analysis also found that more controlling or demanding parenting styles were associated with poorer achievement in school.

We used a multinomial logistic regression to assess how parent factors infuence educational aspirations (see Fig. 6.4). When accounting for schooling and educationspecifc parenting factors, experiencing more demanding parenting was the only signifcant parenting style related to aspiring to attend university (Fig. 6.4). In other words, experiencing high expectations from parents, rather than receiving more independence fostering and emotionally responsive parenting, is a driver of pursuing higher education.

In addition to general parenting styles, parents with higher educational attainment may be more engaged in their children's education at home, likely because they have directly experienced the benefts of further education themselves (Tan et al., 2020). This increased engagement may be associated with adolescents' educational expectations, and with being socialised towards a more education-oriented

**Fig. 6.4** Relative risk ratios with 95% confdence intervals of factors in the family and school contexts for adolescents aspiring to pursue tertiary education. (Reference group: Only complete secondary school)

Notes: Relative risk ratios from a multinomial logistic regression model. All black bars in the fgure are statistically signifcant at p < 0.05 and grey bars are not statistically signifcant. (Source: The Longitudinal Survey of Australian Children (Wave 6) sample (n = 2566))

direction, an important factor related to student achievement (Hill & Tyson, 2009). Indeed, extant research has found that adolescents whose parents have higher expectations for their education may have more substantial future aspirations (Hill et al., 2004) and higher academic achievement (Pinquart & Ebeling, 2020; Strom & Boster, 2007). Given that greater academic achievement in secondary school also leads to improved economic outcomes in adulthood (Ashby & Schoon, 2010; Schoon & Parsons, 2002), parents thus play a critical role in the future economic outcomes of their children.

Parental interest in children's learning can also take the form of direct interest in the education itself, such as children's academic performance or school activities. This form of parental interest is associated with increased academic success in school (Pinquart, 2016; Pinquart & Ebeling, 2020). In contrast, academic outcomes including educational aspirations post-secondary school may be more infuenced by more future-oriented parenting factors. Academic socialisation—the process through which parents foster greater understanding in the value of education, while providing helpful strategies for scholastic success as well as cultivating future plans with their children—is strongly related to academic achievement (Hill & Tyson, 2009; Hill & Wang, 2015; Hill et al., 2018; Gordon, 2016; Gordon & Cui, 2012; Wilder, 2014).

As shown in Fig. 6.4, in contrast to research examining academic achievement, parental interest in learning may not be as strong for infuencing aspirations. Specifcally, results suggest that parent interest in learning itself is not enough to encourage adolescents to aspire for higher education (see Fig. 6.4). However, our results suggest that academic socialisation, a more future-oriented and values driven parenting factor, is related to higher odds of aspiring to pursue a university degree relative to only completing secondary school (see Fig. 6.4). Specifcally, adolescents who report discussing their future plans with their parents have higher odds of aspiring to go to university relative to preferring to only complete secondary school. Together, these results suggest that general interest in an adolescents' schooling is not enough to foster further educational aspirations. Instead, parent engagement through academic socialisation, specifcally through active discussions about the future, is critical in promoting adolescents' goals to pursue tertiary education. These results hold even when adjusting for living in the most disadvantaged regions.

Finally, it is important to consider that the relationship between discussing future plans with one's parents and aspiring to go to university (compared to adolescents who have no plans for further education) may refect having suffcient socioeconomic resources that facilitate viewing university as a viable future option. Adolescents from low socio-economic status backgrounds may be averse to aspiring to further education due to the risk of signifcant downward mobility if they are unsuccessful in completing additional schooling (van de Werfhorst & Hoftstede, 2007). Families who are markedly disadvantaged also experience greater diffculty engaging with their children's education (Baquedano-López et al., 2013). Contributing factors to this diffculty can range from the structure and culture of schools which better facilitate parental involvement for more advantaged families, to limited time and resources to contribute to their children's school activities. Thus, another way of interpreting our results is that for adolescents not intending to pursue tertiary education, the role of academic socialisation may be less important in these families, as their resources to engage with their children's education are far more limited. Despite the fact that pursuing further education leads to better economic outcomes in adulthood (Schoon & Parsons, 2002), familial socio-economic factors during adolescence may nonetheless limit the options that could be considered following secondary school, leading to persistent social disadvantage.

#### **School Factors and Educational Aspirations**

Peer problems during adolescence are correlated with poorer employment rates at young adulthood (Bania et al., 2019; Moore et al., 2015), highlighting the need to examine peer and school factors associated with adolescent educational aspirations. Disruptive school environments and greater peer confict are related to students having more behavioural problems, disengagement from school, and lower educational aspirations (Metsäpelto et al., 2017; Wang et al., 2010). Additionally, having diffculties getting along with peers is related to lower odds of aspiring for full-time education post-secondary school relative to aspiring to enter full-time employment

or vocational training (Hartas, 2016). Furthermore, adolescents who have experienced bullying have lower odds of considering completing a secondary school qualifcation as important (Hartas, 2016; Pinquart et al., 2003). In terms of positive peer and school factors, having a greater sense of school belonging may be associated with better school motivation (Gillen-O'Neel & Fuligni, 2013), having more positive future-orientations (Crespo et al., 2013), and a lower likelihood of not completing secondary school (Bond et al., 2007).

As seen in Fig. 6.4, the results from our multinomial logistic regression suggest that having a greater sense of school belongingness is related to higher odds of aspiring to go to university, relative to only seeking to complete secondary school. However, in contrast to past research, we did not fnd a signifcant relationship between experiences of peer disconnection or bullying victimisation with educational aspirations. Furthermore, although school belongingness is statistically signifcantly related to higher odds of aspiring to attend university relative to no further education aspirations, the effect size is also quite small, suggesting that parent infuences may be more important for seeking to enter higher education.

Finally, turning to decisions to pursue further education through vocational study, parent and school factors were not signifcant correlates (see Fig. 6.5). Instead, the only signifcant correlate is gender, with girls having a lower odds of aspiring to trade and vocational study relative to boys when compared to having no aspirations for further study following school completion. Trade occupations remain

**Fig. 6.5** Relative risk ratios with 95% confdence intervals of factors in the family and school contexts for adolescents aspiring to complete trade or vocational study (Reference group: Only complete secondary school)

Notes: Relative risk ratios from a multinomial logistic regression model. All black bars in the fgure are statistically signifcant at p < 0.05 and grey bars are not statistically signifcant. (Source: The Longitudinal Survey of Australian Children (Wave 6) sample (n = 2566))

male-dominated in Australian society (WGEA, 2019), suggesting that aspirations towards trades is based on gendered norms. In other words, parental and school factors may infuence attitudes towards further education, however, entering the workforce following a non-tertiary education route post-secondary school was found to be only related to gender.

Together, these results suggest that parents, rather than peers or the school environment have the most impact on adolescent aspirations for pursuing tertiary education. This is in contrast to research that suggests that peers, rather than parents, have greater infuence on adolescents engaging in risk behaviours (Dafoe et al., 2018). Interestingly, parent interest in education was not signifcantly correlated with postsecondary aspirations. Instead, higher parental expectations and the prioritisation of future goals emerged as important determinants of adolescent aspirations. Specifcally, greater academic socialisation, through active discussion of plans for the future with parents, plays a critical role in developing future-oriented goals in adolescents such as aspiring to complete a university degree. Furthermore, receiving more demanding parenting from parents was also linked to aspiring to pursue higher education. Collectively, these results suggest that future-oriented parent engagement in the form of academic socialisation, rather than an immediate interest in a child's school performance (such as school-based parent engagement), is essential for aspiring to pursue further education.

#### **Late Adolescence**

One important theme in late adolescence, from a life course perspective, is that of 'continuity and discontinuity in life pathways'. Late adolescence is a time where individual development can either continue a stable trajectory across life stages or be redirected (Johnson et al., 2011, p. 273). Neither transition is objectively good or bad, but educational plans during late adolescence can have long term consequences for later life outcomes. Educational plans and aspirations can be impacted by familial context and parental-adolescent relationships (Whiston & Keller, 2004). Parenting can infuence adolescents' educational aspirations through socialisation of educational values (Spera, 2005), with children whose parents express higher educational aspirations for them, having higher academic and educational goals than children whose parents have lower educational aspirations (Sawitri et al., 2015).

#### *The Study*

In the second part of this chapter, we present new insights into the complexities of parent engagement in late adolescence for students from disadvantaged backgrounds. We also provide a brief overview of how these students perceived support from their parents with their mental health and fnally, how they felt about the future and what aspirations they held.

We draw on data from a larger study (*Learning through COVID-19* funded by the Paul Ramsay Foundation) consisting of qualitative interviews with 29 senior secondary school students (grades 10–121 ) from socio-economically disadvantaged backgrounds from three states in Australia. The interviews were undertaken between September and November 2020 and examined how COVID-19 lockdowns impacted student learning and educational support. Some students experienced high levels of socio-economic disadvantage across multiple dimensions, including housing instability (n = 5), contact with the child protection system (n = 6), contact with the youth justice system (n = 1) and parenting at least one child (n = 6). Nine students were part of a cultural or linguistic minority, and one student identifed as being of Aboriginal or Torres Strait Islander descent. All participants were recruited through service providers who were briefed to approach students living in socially disadvantaged circumstances who were struggling at school and at risk of disengaging.

The participants were predominantly female (6 male/23 female) and lived with at least one parent (n = 19). One third (n = 10) lived either independently, with extended kinship, in refuges or in residential care, having little or no contact with their parents. The sample composition provided analytical leverage for exploring parent engagement beyond the nuclear family offering potential to further conceptualise what parent engagement entails for students in late adolescence. Some participants were enrolled in mainstream (n = 13) and others in fexi-schools (n = 16). Flexi-schools are an alternative option to mainstream schooling in Australia for students, often from disadvantaged backgrounds, who have struggled with the mainstream schooling system. They come in various forms—some are attached to a mainstream school, while others are completely separate institutions. They are generally characterised by the provision of extra support to help students with their life outside of the classroom (such as counselling or housing support) as well as fexible ways of learning, smaller class sizes and increased access to teachers.

#### *What Does Parent Engagement in Learning Look Like in Late Adolescence in Socio-Economically Disadvantaged Households During Remote Learning?*

Parent engagement in student learning decreases as students move up year levels (Daniel, 2015; Murray et al., 2015). Given the increasing expectations placed on maturing students, it is to be expected that this trend is further exacerbated in secondary school. The experiences of the participants in this study illustrate this

<sup>1</sup>For some participants, exact grade information was not available, however they were all aged 15 or above and this cohort specifcally targeted adolescents in grades 10–12.

tendency from their perspective, for example, feeling less supported compared with younger siblings:

Participant: My parents […] don't really care as much because I'm older. I don't know. Interviewer: So, when you compare with your younger brothers, did you feel that they were getting more from your parents to help them out?

Participant: Well probably, because, I don't know, they need more, I don't know, attention and learning and everything. Because I already know most things. (#3, female, 18, grade 12)

However, older participants did not necessarily feel they were being overlooked or missing out unfairly. There was a shared understanding that greater academic capacity also justifed taking on more responsibilities for keeping up with schoolwork. This understanding is potentially problematic if secondary students see themselves being less able to ask for help when needed, or if support is reductively thought about as 'assistance with homework'. Indeed, many participants with younger siblings took on responsibilities for supporting their siblings' learning as well as their own. This was particularly pronounced for participants from a CALD background.

I help my mum raising all of the children, helping out with the cares and also helping the – my younger siblings with homework and things like that. (#18, female, 17, grade 11)

Students from a CALD background were also less likely to draw on their parents for support, due to language barriers or subject knowledge, as explained by this student:

My mum doesn't understand English, my dad's [and] my brothers, their English is really, really poor, and my oldest brother, I did ask him for my maths help, but he's terrible at his maths. (#20, female, 17, grade 11)

This left a gap not easily flled by overworked teachers. This student went on to explain:

They [my teachers] said, 'Is there any people who could help you with your English?' Yeah, and I respond, 'There is nobody.' Because I live with my mum and dad and I have an older brother and the others are young. They go to primary school, and my youngest brother is age four. Yeah, so this was really hard. (#20, female, 17, grade 11)

In circumstances where parents were available to engage in their children's learning, mostly at home, participants expressed their appreciation and the positive impact this had on them.

I only live with my mum. I don't have much to do with my dad […] she was really helpful […] especially because she was doing a lot of things online too, so we could kind of both help each other to stay motivated. And she was a big help in just making sure I was staying on track. She would best try to explain the different things. She's a primary school teacher, so she couldn't help as much as my high school teachers would, of course. But she could still try and explain things to me the best she could and, yeah, she was a big, big help. (#29, female, 17, grade 11)

In this excerpt, help is framed as mutual within the parent-adolescent relationship, rather than positioning the participant as a passive recipient of educational support from an adult. Parent engagement needs to account for competence and autonomy within the way that it is implemented (Li et al., 2020; Raftery et al., 2012). These excerpts also illustrate the need to engage with parent engagement practices in contexts where students have limited contact or no parents at all in their lives.

#### *What Does Parent Engagement Look Like in Non-traditional Socio-Economically Disadvantaged Households During Remote Learning?*

Ten participants in this study were living independently, had limited or no contact with their nuclear family, or had living arrangements with extended kinship members. Not all adolescents have a connection with their biological parents. According to the Child Protection National Minimum Data Set, at 30 June 2020, 7286 adolescents aged 15–17 were living in out of home care in all of Australia (8.3 per 1000 children). For the states included in our sample, NSW had the highest rate of adolescents aged 15–17 in out of home care (10.7 per 1000 children), followed by Tasmania (8.6 per 1000 children) and Queensland (6.5 per 1000 children) (AIHW, 2021b). This is an important opportunity to gain a better understanding of how adolescents whose experiences do not ft neatly into narrow defnitions of parental engagement have their support needs met by drawing on different available sources to enhance their educational engagement. Here, who provides support for learning is less important than the timely access to and level of trust developed in relations with school staff, social workers, service providers or other members of the community. In particular, fexi-school students commented extensively on having access to these sources of support, including crucial emotional support:

I was just sitting in class and [school staff member] just came up and he just put a glass of water and goes, 'If you need to come and have a chat, just come and chat whenever you need.' It was just good knowing that they could see that I was upset. I didn't have to say I was upset, they could just see it, and then they just knew just to not push me. […] It is pretty stressful, but it's good to know that I am able to do some at home and that if I need help, that I am able just to talk to [school staff member …] even after school hours […] because they know how important this job is to me, and they know how important my Year 10 is to me, and they know how much stressed I am. I feel I could just explode. (#39, female, 16, grade 10)

This experience indicates how needs for social, educational, and emotional support overlap and bear upon educational and occupational aspirations and possibilities. This student was living independently with her partner and his grandmother, after having spent some time homeless and in a refuge. She had no contact with her family of origin and at the time of the interview, was juggling schoolwork with a work placement which she hoped would lead into a career after completing secondary school. She also struggled with anxiety, and the wraparound support provided by the school was crucial for meeting needs that would in other circumstances be covered within the nuclear family.

In the absence of parent caregivers, students turned to the relationships with their school or others in their families. Further strategies to enhance participants' capacity during remote learning included meeting basic needs for personal space, food, and drink as well as safety. For the participant below and her siblings, while their mother was working, this support was provided by:

My grandma […]. Checking up on us, if we needed anything, like waters, drinks, food while we were working […]. As much as she could. (#2, female, 15, grade unspecifed)

Extended kinship support was also signifcant for academic socialisation, as described by this participant:

My aunty […] I was living with her when my grandma passed away, but then I decided to move out and do my own thing. […] She's going to help me fnish school and then she's going to try and help me get into TAFE and probably do a few more courses and TAFE and then, yeah, look for a job then. (#21, female, 18, grade unspecifed)

This student faced many barriers progressing through school and on to TAFE – she had a young child, had been in the Child Safety system and suffered with multiple mental health disorders. Child Safety played a role in forcing her to attend school again, which she spoke of with resentment, but it was the support of her family member that had helped her to plan for, and have expectations of, a sustainable future.

The temporary involvement of a parent, even when that involvement occurred amidst changing home environments, was helpful for one of the adolescents interviewed in this study. He resided in a youth housing facility, had signifcant mental health concerns and was in contact with the Child Safety and Justice systems. During some of the time that COVID-19 lockdowns were in place, he lived with his mother, and spoke of her assistance with grateful appreciation:

Interviewer: And you said that your mum helped you. How did – how did she help you during that time?

Participant: She was just reading the questions out for me, like the ones I couldn't read. Yeah, like sort of like helping me, like she was doing the sounds and like trying to […] break it down into like two […] and then try to put it together yourself? […] Well, that's what mum was doing.

Interviewer: And had she ever helped you before or was it just during COVID […]? Participant: It was just during COVID […] Like, she didn't have to, you know. (#38, male, 17, grade 10)

Even though this participant was receiving support from the education and Child Safety systems, the temporary engagement of his biological mother in his learning was meaningful to him. The range of sources and practices of support for student engagement in learning during COVID-19 lockdowns highlights how context matters for secondary school students' educational experiences. Practices included multiple stakeholders and caregivers, temporary or more sustained engagement that often extended beyond simply offering home-work assistance, for example, liaising with school and other partners in the educational system in addition to providing a supportive home environment in which learning takes place.

#### *How Does Parent Engagement in Learning Contribute to or Mitigate School Related Stress During Remote Learning in Socioeconomically Disadvantaged Households?*

Our qualitative interviews also highlighted some complexities in relationships between parental engagement and student mental health and wellbeing. As this student from a CALD background shared:


This students' parents offered (well-meaning) encouragement to persist with diffcult tasks while learning from home during COVID-19 restrictions. However, their daughter perceived this as misaligned with her lived experience and interpreted it as a lack of family capacity to empathize with her. This was compounded by her parents' desire for her to follow a specifc occupational pathway which required a high level of academic success in secondary school. Parent (and teacher) engagement generated undue pressure and feelings of isolation, triggering anxieties around disappointing those who had expressed a belief in this student's abilities. In this way, the responsibility to perform well was felt even more acutely. In contrast, we also saw the potential for parent engagement to mitigate performance pressure when attuned to the secondary school students' experiences. These students below refected on learning during COVID-19 and the role of their parent:


Parents like these engaged in their children's learning as much through encouragement to persist as they did in encouragement to take time out. Some participants mentioned support from parents and other caregivers in pacing themselves, taking regular breaks and moderating (unrealistic) expectations. Some participants who were enrolled in fexi-schools commented that their parents' ability to identify when they were not thriving in the mainstream system, and support them in enrolling in alternative school models, was a signifcant factor in their continued educational participation and had resulted in improvements in their mental health.

#### *What Were the Educational Aspirations and Expectations of Senior Students from Socioeconomically Disadvantaged Households During Remote Learning?*

During late adolescence exploring possible future life plans such as studying further or preparing for a job or career is an important task in the development of identity. These choices shape life course trajectories and may be a source of stress, given that the stakes are perceived as high. Decisions for an educational or occupational pathway often have consequences for many years to come, incur fnancial and other costs, and may require commitments from immediate family or caregivers for their successful implementation. Such decisions are based on pragmatic constraints (i.e., the resources available to the student and their family) and individual aspirations (i.e., the disposition to pursue an educational or occupational goal over another). We defne aspirations as future orientation, planning for a career or educational pathway. Aspirations here are considered separately from expectations, with the former referring to what the individual would like to do in the future, and the latter to what they expect they will do (Almroth et al., 2018). Aspirations are sensitive to socioeconomic status (Goldthorpe, 1996; Coates, 2014) with students from lower socioeconomic backgrounds having lower aspirations for educational attainment than students from more advantaged backgrounds.

All adolescents in our sample planned to complete year 12, with most intending to go on to further study at TAFE or university. This was quite remarkable, given the signifcant level of disadvantage these students were experiencing in their lives. Some expressed their future plans in terms of expectations rather than aspirations, particularly those who were uncertain of what they wished to do but expected to go on to TAFE and hopefully discover an occupational pathway.

I'm thinking of getting a job and then like getting myself into a TAFE course. (#24, female, 16, grade 11).

Participants rarely framed their discussion of aspirations and expectations in terms of support received from their teachers or parents. This may refect the desire for independence and autonomy that characterises adolescence, although participants also struggled with the responsibilities, they perceived that the next stage of their life would bring. Discussions about educational and occupational futures revealed that for many participants these were areas of considerable uncertainty and, in some cases, worry:

Quite a big worry has been that […] I'm not going to be quite […] smart enough to go to uni or become an adult. I guess that's quite scary to me, having responsibilities. That's quite terrifying. […] I only have basically one year left and then I'll be 18. Just even leaving school is kind of scary to me because everything that I've known my whole life it's going to change. I've got to start looking for jobs […] What scares me is kind of just being stuck. That's something for me, which is really big, is that, yeah, I don't want to ever feel trapped. What's scary to me is that I'll go into a job and maybe I realise that it's not what I want to do and then I feel like that's it (#29, female, 17, grade 11)

This excerpt indicates that some participants' anxieties were fuelled by an understanding of their current life stage as one in which decisions now would be diffcult to change later on. There is a sense of being expected to take responsibility accompanied by doubts about being ready for this degree of autonomy. Autonomy here was not longed for as a means of self-actualisation but contrasted with concerns about affordances (e.g., for experimentation or making mistakes) extended to young people in school becoming unavailable once secondary education was completed. These dynamics coincided with pressures to perform well in high stakes' exams that would infuence options for future educational and occupational choices. Participants commented on how conficting pressures affected other activities they would usually engage in to cope with stress or just feel good:

Sports have been reduced a lot and because I've been more focused on my studies, I haven't been able to […] spend that quality time with them [family and peers] lately […] at school, I'm doing four pre-tertiary subjects and that's quite hard to manage because I'm trying to get … good grades in all of them and that's—towards this time when […] exam time is getting really tough and […] it's quite a large workload for me. (#19, male, 17, grade 11)

It is not surprising that adolescence presents an array of emergent pressures. Parents, family, peers, and school staff are the primary sources of support during this period. At the same time, the demands and emphasis on autonomy in adolescence change how support can be made available further challenging the way parents engage in their children's learning. High aspirations in early adolescence have been linked to improved mental health (Almroth et al., 2018), so it is possible that despite these emergent stressors, students' high aspirations will be a protective factor for their mental health, possibly through increasing positive attitudes towards their future life goals.

#### **Conclusion**

The convergence of developmental and contextual transitions in adolescence increases the risk for poor outcomes across the life course. It is therefore imperative to identify sources of support in home and school environments that can mitigate risks. Parents from disadvantaged backgrounds are at greater risk of experiencing barriers to both forming partnerships with schools and engaging in their child's learning more generally (Fox & Olsen, 2014; Kim, 2009; Turney & Kao, 2009), therefore interventions targeted at these families and the support systems they draw on are needed to break the cycles of deep and persistent disadvantage.

Our chapter investigated the role of parent engagement during early, middle, and late adolescence on student outcomes and experiences. Findings from Part 1 show that parental interest in education was related to more positive academic or ability related self-concept, after controlling for family and school factors, although the effect was small. It was also related to lower levels of depression symptoms, although not anxiety after controlling for school factors such as bullying and school belonging. Parental factors, such as academic socialisation through parents discussing adolescents' futures with them, and perceiving high parental expectations, were predictive of tertiary study aspirations. In contrast, general parental interest in their education was not related to adolescents aspiring for higher education. These results help to illustrate that parent engagement is important in early and middle adolescence, even when accounting for family and school context factors. However, it is also important to remember that the parent engagement or academic socialisation an adolescent receives may be contingent on socio-economic factors. While parent engagement can drive motivation towards higher education—an important facilitator of improved economic outcomes in adulthood—families must have resources to provide this form of engagement.

The fndings from Part 2 build on those from Part 1 by allowing us to analyse parent-student engagement for students from highly disadvantaged backgrounds. The fndings indicate that being attuned to adolescents' shifting priorities, experiences and support needs is key. Engagement and support can be provided by parents, but also by signifcant others such as teachers, service providers, or community members. This engagement in their learning and support is perceived as important by these students. This may be especially true for students who are experiencing social disadvantage, and for whom continued involvement in education may strongly infuence their outcomes in later life.

**Acknowledgements** This chapter is based in part on research funded by the Paul Ramsay Foundation. Any opinions, fndings, or conclusions expressed in the chapter are those of the author(s) and do not necessarily refect the views of the Foundation. We would also like to acknowledge the support of the research assistants in the Learning Through Covid-19 project, for assistance with recruitment, data collection and analysis.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 7 Differences in Higher Education Access, Participation and Outcomes by Socioeconomic Background: A Life Course Perspective**

#### **Wojtek Tomaszewski, Francisco Perales, Ning Xiang, and Matthias Kubler**

Educational inequalities are a critical area of policy concern in countries across the globe (OECD, 2017b), including Australia (Harvey et al., 2016b). In particular, the impacts of family background on educational outcomes—including access, participation, and post-graduation outcomes—have received considerable attention from both academics and policymakers (Tomaszewski et al., 2018). However, previous studies have often focused on a particular stage of the student life cycle, instead of approaching the issue through a more holistic lens. This chapter provides such a holistic perspective through leveraging the life course approach as an overarching framework.

Inequalities in higher education are particularly suited to be understood from a life course perspective, as they constitute the culmination of long-term processes of accumulation of advantage and disadvantage beginning at birth, or even prior to birth (Tomaszewski et al., 2018). Higher education is also a well-recognised mechanism for promoting intra-generational mobility, since university qualifcations open the path to the most lucrative positions within the occupation structure (Desjardins & Lee, 2016; Heckman et al., 2016). In this chapter, we argue that inequalities in educational outcomes manifest across the key stages of a university student 'life

N. Xiang · M. Kubler Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia e-mail: n.xiang@uq.edu.au; m.kubler@uq.edu.au

W. Tomaszewski (\*)

Australian Research Council Centre of Excellence for Children and Families over the Life Course. Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia e-mail: w.tomaszewski@uq.edu.au

F. Perales

Australian Research Council Centre of Excellence for Children and Families over the Life Course. School of Social Science, The University of Queensland, Brisbane, QLD, Australia e-mail: f.perales@uq.edu.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_7

cycle': the access, participation, and post-participation stages. We propose that, at each of these stages, individuals from socio-economically disadvantaged backgrounds will exhibit poorer outcomes than individuals from more socio-economically advantaged backgrounds. We link educational inequalities at each of these stages to indicators of family background that approximate different positions within the social hierarchy. This chapter focuses largely on one important status: the role of family socio-economic status (SES) commonly measured using parental occupational or educational status, or a composite measure of these.

Overall, it is well known that individuals born in low-SES families—however defned—are less likely to achieve positive outcomes, including participation and success in higher education and high labour-market returns to university qualifcations. We know less, however, about how these differences come about, the mechanisms producing and reproducing them, the extent of such disparities at different points of the life course, and how we should intervene to ameliorate them. In this chapter, we showcase the value of the life course approach to understanding and intervening to resolve the persistence of educational inequalities at different stages of individuals' educational careers. We also summarise recent empirical evidence taking a life course approach to enhance our appreciation of how socio-economic inequalities translate into differences in educational outcomes along different stages of university students' life cycle. We conclude by pointing to policy implications and avenues for further research.

#### **Inequalities in Higher Education from a Life Course Perspective**

The life course approach is an increasingly infuential lens through which researchers can study the intersections between socio-economic status and educational inequalities (Elder, 1995; Elder et al., 2003). As discussed in other chapters in this volume, this interdisciplinary research paradigm proposes principles to guide research on how relationships, life events and transitions, and social forces, infuence people's lives—including their educational outcomes—from birth to death (Elder, 1995). The role of time in producing and reproducing social inequalities is one of the central tenets of the life course approach. Broadly, this perspective highlights the importance of recognising 'longitudinal dependencies'; that is, the infuence of previous life events and experiences on later life outcomes. A toolset of theoretical principles follows from this general proposition, some of which are useful to understanding differences in higher-education access and participation by SES.

Unlike cross-sectional perspectives, where accumulation of disadvantage pertains to combinations of concurrent disadvantaged statuses, accumulation of disadvantage understood through the lens of life course theory refers to the recurrent experiences of disadvantage over time. Disadvantage is treated not as a static state, but rather as a cumulative process that unfolds over the life course (Elder et al., 2003). Compared to a one-off experience, repeated or chronic exposure to barriers and stressors can be more harmful to individuals' chances to succeed in various aspects of their lives (Ferraro & Kelley-Moore, 2003; Kuh & Ben-Shlomo, 2004; Laub & Sampson, 1993; Sampson & Laub, 1996). Hence, individuals' outcomes across life domains (including education) must be understood in the context of their earlier experiences in those and other domains. A related and important concept is that of individuals' life course trajectories, defned as the sequences of events, transitions and social roles across multiple and interconnected life domains that individuals encounter as they age. Observing people at a single location within such trajectories is but a poor proxy for where they come from, and where they are heading.

In the context of this chapter, one could think of individual-specifc trajectories capturing young people's cumulative experiences and success, or lack of it, in the higher education system from entry (access to higher education) through participation, and onto exit (including post-graduation outcomes). These individual educational trajectories will be intertwined with those for other life domains, such as family or health trajectories. Individuals from low SES backgrounds experience greater exposure to stressors and barriers to educational success at multiple stages of the life course, which can result in educational trajectories that differ from those of those from more advantaged backgrounds.

Life course trajectories (including those pertaining to education) are not deterministic, and can be altered by life course experiences, including events and transitions. When negative, trajectories can be improved by individuals' receipt of formal or informal support. In life course jargon, visible changes in the direction of a longterm developmental trajectory are referred to as turning points. Closely related to this notion are the concepts of sensitive and critical periods. Sensitive periods are periods within individuals' life courses in which their outcomes are comparatively more likely to be affected, or affected more strongly, by a certain factor (e.g., the experience of adversity). Similarly, critical periods are those in which such exposures have lasting, even lifelong, consequences on individuals' outcomes (Ben-Shlomo & Kuh, 2002). In addition, different factors may become important for individuals' outcomes at different turning points of the life course. In the context of higher education, belonging to certain socio-economically disadvantaged groups may have different effects on the onset and maintenance of educational disadvantage at different points of the educational life course.

#### **Opportunity Pluralism and 'Bottlenecks' to Success**

Fishkin's (2014) recent theory of opportunity pluralism offers additional insights into how educational inequality, including inequality in higher education, needs to be understood as a long-term process. Fishkin argues that the goal of Government policy should be to increase the range of opportunities available to individuals at all life stages, so that people can freely choose among different life paths leading to

'human fourishing'. The pursuit and attainment of a university degree is a clear example of such a path. This argument resonates with the notion of 'capability development', coined by economist Amartya Sen, and its associations with social disadvantage (Sen, 1992).

Opportunity pluralism involves revising the ways in which we think about equality of opportunity by focusing on how opportunities in a society are created, distributed and controlled. Fishkin uses the analogy of 'bottlenecks' to refer to "the narrow places through which people must pass if they hope to reach a wide range of opportunities that fan out on the other side" (Fishkin, 2014, p. 1). In the context of this chapter, higher education is in itself a 'bottleneck', one that regulates access to labour market outcomes, social recognition and prestige, and other desirable socioeconomic outcomes. However, earlier bottlenecks regulate individuals' progress through the education system leading to higher education participation. These include: (i) Developmental bottlenecks—whereby certain developmental capacities are required to develop higher-order abilities (e.g., cognitive or non-cognitive skills necessary to do well at school); (ii) qualifcation bottlenecks—when a formal qualifcation or test score is required to gain access to the next set of opportunities (e.g., a pass mark in a university admission test); and (iii) instrumental-good bottlenecks—when some material good is required to attain a goal (e.g., money to fund one's university studies).

Consistent with the life course principles outlined before, these educational bottlenecks are sequentially arranged and span from very early in life to the point of higher education entry (and beyond). As Chambers (2009) puts it "each outcome is another opportunity" (p. 374): access to later bottlenecks is contingent on successful passage through earlier ones, producing a concatenation of opportunities. For example, with some exceptions, young people typically face the opportunity to enter higher education if they have previously obtained a secondary school qualifcation, and only obtain such qualifcations if they have developed the necessary cognitive skills. Critically, many of these opportunities are shaped by individuals' socio-economic background, particularly at the earlier stages of the life course.

At the early stages of the life course, both genetic and socio-economic factors running though families affect children's capacity to go through the frst set of bottlenecks (Heckman, 2006; Kautz et al., 2014). The former include the child's genetic endowments, while the latter include parental investments in the child, e.g., money, parenting practices, role modelling, etc. At this point, Fishkin's thinking deviates from other perspectives emphasising the importance of early intervention—e.g., that of Nobel Prize awardee James Heckman. While Heckman advocates for intervening as early as possible to shift a disadvantaged life trajectory (within the frst 5 years of life and prior to school entry), Fishkin argues that early intervention is never 'early enough' and thus not suffcient in isolation. In his words, "there is no fair place to put the starting gate" as "any starting gate will have the effect of amplifying past inequalities of opportunity" (Fishkin, 2014, p. 71). An alternative to early intervention may consist of early and sustained intervention over the life course: "instead of building a starting gate at one specifc place, we have to do some of the work of addressing or mitigating inequalities at every stage" (Fishkin, 2014, p. 73). In addition, the necessity for certain bottlenecks should be questioned in the context of their consequences for structuring opportunities and life trajectories, such as the necessity of passing certain formal academic tests at specifed ages.

#### **Life Course Theory and Opportunity Pluralism in Practice: A Conceptual Model of the Student Life Cycle**

Applying the life course approach and opportunity pluralism theory to inform our understandings of educational disadvantage in Australia hinges on the ability to identify the nature of the bottlenecks encountered by young people navigating the process. Recent efforts have provided some groundwork in this area in the context of higher education participation (Australian Institute of Health and Welfare (AIHW), 2014). For example, research by the Australian Institute of Health and Welfare (AIHW, 2014) proposed a model in which student higher education disadvantage was evaluated at four phases in the student life course: (i) the pre-entry phase (i.e., the decision about whether to apply to university), (ii) the enrolment phase (i.e., experiences navigating the admission process), (iii) the university experience phase (i.e., how well young people cope with participation in higher education), and (iv) the graduate outcomes phase (i.e., the post-graduation returns to education).

A similar model was suggested as part of the Critical Interventions Framework for advancing equity in Australian Higher Education by Naylor et al. (2013). The categorisation of equity initiatives contained in that framework was based on existing equity programs, such as outreach programs, many funded by the Higher Education Participation and Partnerships Program (HEPPP), which is a testament to the fact that an understanding of the longitudinal nature of disadvantage in higher education has already been infuencing higher education policies in Australia for some years. Consistent with existing outreach programs, the aforementioned models and associated evaluation frameworks consider the stage of school (including primary school) as the earliest point for interventions and evaluations relevant to higher education outcomes.

Pitman and Koshy's (2014) framework for measuring equity performance in Australian Higher Education refects an even stronger acknowledgement of the longitudinal nature of disadvantage in higher education. Their framework includes the domain of early childhood and proposes to report indicators from the Early Child Development Census and attendance rates at pre-school as part of the higher education equity performance reporting. The framework also covers student achievement indicators at primary and secondary school level, which could provide the basis for a more thorough examination of bottlenecks in the Australian education system, particularly as they pertain to the experiences of students from families of different socio-economic status.

Recent reports by the Centre for International Research on Education Systems (Lamb et al., 2015, 2020) adopt a similar perspective. They consider a conceptual model in which educational opportunities in the Australian context are channelled through four sequential milestones: (i) Developmental readiness at school entry, (ii) school performance in Year 7, (iii) completion of school Year 12 by age 19, and (iv) full engagement in education, training or work at age 24. Progress through these milestones constitutes a 'leaking pipeline' of sorts, with some students meeting the milestone and others dropping out of the race. This research provides important background and directions for the identifcation of 'milestones' and 'bottlenecks' in Australian youth's pathways towards higher education (and beyond). The importance of considering longitudinal dependencies to study disadvantage in higher education is internationally recognised (OECD, 2017a, b) and apparent in the policy models in other countries, providing support that long-view solutions can work in practice.

#### **The Three Phases of the Student Life Cycle: Access, Participation, and Post-participation**

Informed by the body of work outlined above, in this chapter we conceptualise the pathways into and through higher education along three broad stages: access, participation and post-participation. Within each of these stages, we are interested in how socio-economic background shapes outcomes relevant to higher education.

In our framing, the 'access' stage comprises the 'before university' part of the student trajectory, including school experiences and performance. The key aspects to consider with respect to the access stage include the multiple barriers that individuals from socio-economically disadvantaged backgrounds may face in their journey to university. Previous literature identifes a range of such barriers, including material barriers (Edwards & McMillan, 2015; Lamb et al., 2004; Sellar & Gale, 2011), cognitive and non-cognitive barriers (Dickson et al., 2016; Harvey et al., 2016a; Kautz et al., 2014), aspirational barriers (Harvey et al., 2016a; James et al., 2008), cultural barriers (Bok, 2010; Burke et al., 2016; Devlin, 2013) and institutional barriers (Armstrong & Cairnduff, 2012). In this chapter, we present the results of empirical analyses in which we tested the relevance of selected barriers on the likelihood of university enrolment among young people from socio-economically disadvantaged backgrounds.

The 'participation' stage follows from the 'access' stage and refers to the time at university. The key areas of interest identifed in the literature include issues of continued participation (attrition/retention) and completion of university studies. For instance, previous studies suggest that low SES students not only have a lower likelihood to enrol into university compared with their more socio-economically advantaged peers, but they are also less likely to continue and complete their studies once enrolled (Harvey et al., 2016a, b; Productivity Commission, 2019). In this chapter,

we explore the infuence of individual and family-level indicators of socio-economic background on the likelihood of university participation within the Australian context.

The 'post-participation' stage covers the time after university graduation and has been increasingly recognised in the literature as an integral stage when considering outcomes pertaining to higher education participation and success (Naylor et al., 2013; Tomaszewski et al., 2021). Completing a university degree has been associated with a number of positive outcomes spanning various life domains, including labour market participation (Heckman et al., 2016; OECD, 2019) and improved health and wellbeing (Cutler & Lleras-Muney, 2008; Heckman et al., 2017; Oreopoulos & Petronijevic, 2013). However, robust evidence on inequalities in these post-graduation outcomes by socio-economic background remains sparse. The analyses presented in this chapter afford us unique understandings of these processes.

In summary, this chapter aims to shed the light on how socio-economic background shapes differences in higher education outcomes at various stages of the student journey, including the mechanisms producing and reproducing these differences, and the extent to which such disparities emerge at various points of the university student life cycle. This evidence can in turn improve our understanding on how and when we should intervene to ameliorate socio-economic-background inequalities in higher education.

#### **Socio-Economic Status Differences in Higher Education Access, Participation, and Outcomes: Empirical Evidence for Australia**

The theoretical considerations and empirical evidence outlined earlier lead to the expectation that socio-economic disparities in family background may shape individuals' outcomes at various stages of their journeys into and out of higher education. Validating and refning these theoretical propositions, however, requires a body of robust empirical evidence. In the remainder of this chapter, we identify and briefy summarise key fndings from selected recent Australian studies relevant to the infuence of socio-economic background on higher education outcomes at various points of the life course. These fndings come from our long-standing program of research aimed at understanding how individuals from disadvantaged backgrounds, including low socio-economic status, engage with the higher education system. The evidence presented here is based on robust, nationally representative longitudinal data for Australia, which improves on earlier cross-sectional studies based on a single point in time and longitudinal studies based on small and nonrepresentative samples.

Over the next sections, we present empirical evidence from three studies covering the three stages distinguished in our conceptual model of student life cycle: access to higher education, participation in higher education, and post-participation outcomes. First, we present evidence on the infuence of various barriers to accessing university faced by young people at the cusp of making the transition out of school (at approximately 15 years of age). We then move onto exploring how various individual and family-level markers of socio-economic disadvantage shape the chances of university participation several years later, at the ages of around 18–22 years. Finally, we present empirical evidence on how socio-economic family background shapes a range of post-graduation outcomes spanning up to 15 years after obtaining a university degree.

#### *Access to Higher Education: Uneven Barriers to University Enrolment by Socio-Economic Background*

We start our presentation of empirical evidence by showcasing a recent analysis pertaining to the access to university stage, focusing on the relevance of various barriers to university enrolment among low and high SES students. The analysis leverages data from the 2006 *Longitudinal Surveys of Australian Youth* (LSAY). LSAY was designed to study transitions from school to further education, work and other destinations, following cohorts of students aged 15 years and tracking them over a 10-year period. The survey allows longitudinal tracking of post-school outcomes, while capturing information on a range of individual and family characteristics, including markers of disadvantage and potential barriers to participating and succeeding in higher education. The analyses presented in this chapter were based on the 2006 LSAY cohort, which tracks a cohort of young people aged 15 in 2006 up until 2015. The key variables representing various barriers to university participation were captured at LSAY Wave 1, when the individuals were approximately 15 years old.

Before we start discussing the specifc barriers to accessing university among low SES youth, it is useful to present an overall picture of the socio-economic disparities in the likelihood of university enrolment over time. Figure 7.1 shows the likelihood of enrolling into university for young people from low and high SES backgrounds based on the LSAY 2006 data. It is evident that young people start enrolling into university between Waves 3 and 4 (at about 18–19 years of age), and already at that point a slightly larger proportion of young people from high SES backgrounds do so, compared with their peers from low SES backgrounds. However, the socio-economic background gap in enrolment becomes larger by Wave 5 (at approximately 20 years of age). At this point, about 50% of young people from high socio-economic status backgrounds have already enrolled at university, compared with just over 20% of young people from low SES backgrounds. This university enrolment gap remains fairly constant up to the end of the observation period (Wave 11; age approximately 25). By then, around 60% of young people from higher socio-economic backgrounds and about 30% of young people from low

**Fig. 7.1** Proportion of students enrolling into university, by socio-economic background Notes: Kaplan-Meier (A Kaplan Meier graph was chosen to visualise the gap between the two groups and as such is the preferred approach than a life table) hazard rates, replicating analyses included in Tomaszewski et al. (2017) updated with LSAY 2006 data. Source: LSAY 2006 sample (n = 14,170)

socio-economic status backgrounds were observed to have enrolled into university. Hence, we fnd a large socio-economic status gap in university enrolment, one which emerges largely due to differential enrolment rates at a "standard" enrolment time (i.e., ages 18–19).

The available literature identifes various mechanisms that help explain the socio-economic status differences in university enrolment. Material barriers to accessing higher education relate to a lack of fnancial resources, which can compromise the ability of families or individuals to invest in education, including higher education (Sellar & Gale, 2011; Edwards & McMillan, 2015). Cognitive barriers to higher education involve multiple dimensions but have been approximated in the empirical literature by measures of school performance. In Australia, poor school performance has been recognised as a major barrier to accessing university for people from low SES backgrounds (Harvey et al., 2016a; James et al., 2008). Noncognitive barriers refer to lack of non-cognitive skills, such as perseverance ("grit"), conscientiousness, self-control, trust, attentiveness, self-esteem and self-effcacy, resilience to adversity, openness to new experiences, empathy, humility, or tolerance of diverse opinions (Kautz et al., 2014). Barriers related to aspirations and expectations for university participation have been another major focus of research and intervention in Australia in relation to people from low SES backgrounds

(Bennett et al., 2015; Gale & Parker, 2018).1 Cultural factors are another kind of potential barrier to accessing university. Some authors see universities as reproducing, valuing and relying upon particular cultural views and practices concerning ways of speaking, thinking and behaving, which can manifest as barriers for those socialised into different social and cultural worlds (Devlin, 2013). Relatedly, institutional barriers to accessing university refer to the barriers embedded in the structures and characteristics of schools and universities, and their associated practices (Armstrong & Cairnduff, 2012).

Figure 7.2 illustrates examples of such barriers with empirical data based on our research. The fgure shows how various barriers to university participation that

#### **Fig. 7.2** Selected barriers to university enrolment

Notes: Hazard ratios from a Cox regression model (Cox regression models are semi-parametric regression techniques of the event-history family, which are useful to determine how different factors infuence the occurrence of an event over time (Box-Steffensmeier & Jones, 2004). In the context of higher education research, event history models are commonplace in studies of university student retention/dropout (Bahi et al., 2015; Gury, 2011; Vallejos & Steel, 2017) and time to degree completion (Lassibille & Gómez, 2011; Wao, 2010). They have also been used to examine routes to University amongst non-traditional students (Brändle, 2017) and returns to University after intermissions (Johnson, 2006)) adjusted for gender, Indigenous status, school location and non-English speaking background (NESB), based on analyses included in Tomaszewski et al. (2018). Variables capturing barriers to university participation recorded at Wave 1. University enrolment measured across Waves 2–10. Source: LSAY 2006 sample (n = 9326)

<sup>1</sup>See Sellar and Storan (2013) for an analysis of the policy context in Australia; and Burke (2012), Sellar (2013) and Sellar et al. (2011) for critiques of the defcit connotations in the policy rhetoric of 'raising aspirations'.

**Fig. 7.3** Propensity to experience barriers to higher education participation by socio-economic background

Notes: Odds ratios from logistic regression model adjusted for gender, Indigenous status, school location and non-English speaking background (NESB), based on analyses included in Tomaszewski et al. (2018). All ORs are statistically signifcant in relation to the reference category (higher SES students) at p < 0.005. Variables capturing barriers to university participation and SES recorded at Wave 1 and Wave 2. Source: LSAY 2006 sample (n = 9326)

could be identifed in LSAY are associated with a lower propensity to enrol into university (for both low and higher socio-economic status students).

Next, we demonstrate that these common barriers to accessing university are more likely to occur among young people from low socio-economic status backgrounds. Figure 7.3 shows that the likelihood of experiencing these different barriers to university participation is signifcantly higher among people from low socio-economic status backgrounds, relative to their more advantaged peers. The results confrm that low socio-economic status is consistently associated with increased chances of experiencing the barriers to higher education participation. For example, the odds of low socio-economic status students having low PISA math scores were almost 3 times greater than those of students from more advantaged backgrounds, while the odds of attending a low socio-economic status school or receiving fnancial support in the form of youth allowance were even higher around 4 times greater than those of students from more advantaged backgrounds.

#### *Socio-Economic Background and Higher Education Participation: Contributions of Individualand Area-Level Factors*

Previous research demonstrates that, compared with their more socio-economically advantaged peers, low socio-economic status students not only have a lower likelihood to enrol into university, but are also less likely to continue their studies once enrolled (Harvey et al., 2016b; Productivity Commission, 2019). While the lower likelihood of higher education participation among low socio-economic status students is a well-established fnding, a number of research gaps remain. These include the concrete factors that may be driving continuing participation among disadvantaged students. Specifcally, one of the questions that remains unanswered is the relative infuence of markers of socio-economic advantage or disadvantage operating at various levels (e.g., the individual, area, or school level).

This section of the chapter presents the results of our recent research aiming to address this question though leveraging data from the *Australian Census Longitudinal Dataset* (ACLD). Specifcally, data from the ACLD 2011–2016 (ABS, 2018) panel were used to investigate university participation for people growing up in disadvantaged areas. We capture university participation at the ages of 18–22 and relate that to the data on socio-economic disadvantage captured 5 years earlier, when the young people were aged 13–17 and most of them still lived with their parents.

The empirical analysis presented here focuses on demonstrating the relevance of different aspects of socio-economic advantage and disadvantage captured at the individual-, local-area and school-sector levels to university participation. The individual-level markers of socio-economic disadvantage include an indicator for a university degree in the family (to approximate the relative disadvantage of being the 'frst in the family' to attend university) and the information on the occupation of the male parent. Area-based indicators include an indicator of the Socio-Economic Indexes for Areas (SEIFA) Index of Education and Occupation (IEO) score of the postcode area of a student's permanent address, which has been a measure commonly applied to monitor socio-economic disadvantage in higher education in Australia.2 We capture area disadvantage using deciles, with the lowest, most disadvantaged SEIFA decile 1 as the reference category. We further include a measure of area remoteness, which constitutes a separate dimension of area-level disadvantage (Burnheim & Harvey, 2016). Furthermore, we also include information on the school sector (Government, Catholic and Other Non-Government) that the young person attended in 2011. The systemic role of the school system in maintaining or

<sup>2</sup>The SEIFA IEO combines census data on occupational and educational characteristics of individuals living within communities into a composite index to rank geographic areas. The resulting index is then assigned to individuals living in a given postcode area. Postcodes falling in the bottom 25% of the population aged 15–64 are categorised as low SES in offcial higher education equity reporting.

generating disadvantage in higher education participation has been a topic of considerable debate in Australia (Goss et al., 2016; James et al., 2004; McConney & Perry, 2010) and elsewhere (Hanselman, 2018; Hauser et al., 1976; OECD, 2010; Van de Werfhorst & Mijs, 2010). Figure 7.4 shows the likelihood of participating in higher education depending on these individual-level, area-level and school-level indicators.

The results confrm the relevance of individual-level, area-based and schoollevel indicators of socio-economic status for university participation. Specifcally,

**Fig. 7.4** Association between university participation and selected individual, area, and school characteristics

Notes: Odds ratios from multiple logistic regression adjusted based on analyses included in Tomaszewski et al. (2018). All ORs are statistically signifcant in relation to the reference category at p < 0.01. People not attending secondary school in 2011 were excluded from the model. Source: Australian Census Longitudinal Dataset, 2011–2016, data extracted using TableBuilder in June 2018

the chances of participating in higher education decrease with residence in each lower SEIFA decile, as indicated by the increasing magnitude of the odds ratios (which are all greater than 1) in relation to the reference category. Similarly, students from regional and remote areas are less likely to participate in higher education within the 5-year period, compared with students from major cities. Furthermore, the results demonstrate that the added individual-level markers exert an independent statistical effect on the probability of university participation. For example, secondary school students aged 13–17 years in 2011 had more than twice the odds of participating in higher education within 5 years when someone in their family had a degree in 2011 (ORs >2). Similarly, there are independent effects of the father's occupation, and the type of school they attended in 2011, with students attending Independent or Catholic schools more likely to subsequently participate in higher education.

Taken together, the results suggest the independent effects of multiple individuallevel, area-level and school-level markers of socio-economic advantage and disadvantage (parental occupation, 'frst in family', SEIFA IEO scores, type of school, and area remoteness) on the likelihood of participation in higher education over the next 5 years. Each of these markers has its own independent statistical effect on the probability of university participation.

#### *Beyond Graduation: Long-Term Impacts of Socio-Economic Background on Students' Post-Graduation Outcomes*

Research has consistently demonstrated that university education has positive impacts on individual outcomes. For instance, only 7% of university-educated adults aged 25–34 years are unemployed across OECD countries, compared to 9% of those with upper-secondary or post-secondary education, and 17% of those with lower qualifcations (OECD, 2017a). Data for Australia also shows that employment rates are substantially higher for individuals holding postgraduate (82%) and bachelor (80%) degrees, compared with individuals without post-school qualifcations (54%) (ABS, 2017). Furthermore, university graduates in Australia and internationally are also more likely to work in more prestigious occupations and earn higher wages (Card, 1999; Cassells et al., 2012; Daly et al., 2015; Desjardins & Lee, 2016; Heckman et al., 2016). The positive effects of higher education qualifcations are not limited to the labour market, with research documenting infuences across a range of other domains, including, general health (Cutler & Lleras-Muney, 2008), mental health (Heckman et al., 2017) and subjective wellbeing (Oreopoulos & Petronijevic, 2013; Oreopoulos & Salvanes, 2011). However, few studies have explicitly considered differences in these impacts by socio-economic background of the graduates. The scarce studies that explicitly consider socio-economic differences among graduates, tend to focus on short-term post-graduation outcomes, often due to limited data over long-run trajectories, and only consider labour market outcomes.

Internationally, Hansen (2001) reported higher economic returns to a university degree in Norway for high socio-economic status graduates than for low socioeconomic status graduates, net of qualifcation level and feld of study. Similarly, graduates in Norway, Italy and Spain whose parents had university qualifcations were shown to be more likely to attain a high-status occupation 5 years after graduation, compared with similar graduates whose parents did not hold university qualifcations (Triventi, 2013). Furthermore, a study analysing the effect of parental education on German and British university graduates' occupational outcomes at the point of entry into the labour market, and 5 years after graduation showed that high socio-economic status individuals had a relative advantage over low socioeconomic status graduates in securing high-status occupations (Jacob et al., 2015). However, the effect at the point of entry into the labour market was stronger than 5 years after graduation, suggesting a weakening of the socio-economic status gradient over time.

The limited Australian research in this area provides mixed evidence. For instance, Richardson et al. (2016) showed that low socio-economic status graduates were less likely to be employed 6 months after graduation, compared with high socio-economic status graduates. However, Li et al. (2017) reported no signifcant differences in employment rates between low and high socio-economic status graduates. Taking a longer perspective, Edwards and Coates (2011) demonstrated that low and high socio-economic status graduates had similar rates of employment and employment in a high-status occupation and median annual salaries 5 years after completing their university studies.

This part of the chapter presents recent empirical evidence drawing on our analysis of longitudinal data from the *Household, Income and Labour Dynamics in Australia* (HILDA) Survey. The analysis aims to examine post-university trajectories of low socio-economic status and high socio-economic status graduates over a long run (up to 10 years post-graduation), capturing outcomes across multiple domains, including those that extend beyond indicators of labour-market performance.

HILDA is an annual household panel survey following a sample of individuals aged 15 and older over time. The HILDA Survey sample is largely representative of the Australian population in 2001 (Watson & Wooden, 2012). Data are collected using a complex, multi-stage sampling strategy at the household level, and a mixture of self-complete questionnaires and computer-assisted face-to-face interviews. In the empirical analyses summarised below, we use the frst 16 Waves of the HILDA data, covering years 2001 to 2016. We use these data to construct an analytic sample of individuals who were observed at least twice and obtained a Bachelor's degree during the life of the panel. In order to examine trends in outcomes post-graduation, we only include observations subsequent to individuals obtaining their degrees. We apply growth curve models to estimate the trajectories of post-graduation outcomes for individuals from low socio-economic status and high socio-economic status backgrounds. In this analysis, high socio-economic status graduates are defned as those with at least one parent in a professional or managerial occupation, and low socio-economic status as those where none of the parents are in a professional or managerial occupation.

We focus the analyses on four outcome variables spanning labour-market circumstances, health and wellbeing (Fig. 7.5). Specifcally, we model:


The results show that hourly wages and fnancial prosperity increase with time since graduation for all graduates, while mental health and job-security satisfaction

**Fig. 7.5** Selected outcome trajectories for university graduates from low and high SES backgrounds

Notes: Marginal effects from growth-curve models. Covariates held at their means and random effects at zero. Whiskers denote 90% confdence intervals. Source: HILDA Survey (2001–2016). Originally published in Tomaszewski et al. (2021)

remain stable over time. The picture is somewhat mixed in relation to differences in outcomes by SES. The hourly wages and mental health of low socio-economic status graduates appear to be on par with those of high SES graduates throughout the observation period, that is, no differences between low socio-economic status and high socio-economic status graduates is detected. However, low socio-economic status graduates (red lines) have lower job-security satisfaction and fnancial prosperity in the frst 4 years post-graduation, compared with high socio-economic status graduates (blue lines). These initial differences are statistically signifcant, as can be inferred from non-overlapping 90% confdence intervals. Subsequently, the trajectories of low and high socio-economic status graduates on these outcomes converge over time. That is, there is evidence of a 'catch up' effect for low socioeconomic status graduates, which results in the outcomes comparable to those of high socio-economic status backgrounds towards the end of the observation period. Still, it takes a relatively long time—at least 4–5 years after graduation—for the average low socio-economic status graduate to achieve outcomes comparable to the average high socio-economic status graduate.

#### **Discussion and Conclusion**

In this chapter, we used a conceptual model of the university student life cycle as an organising framework to present a body of recent Australian evidence on differences in pathways through the higher education system among individuals from low and high socio-economic status backgrounds. The model introduced in the chapter splits individuals' higher education trajectories into three distinct stages: access, participation and post-participation, with empirical evidence presented for each of these stages on the factors infuencing the outcomes of people from low socioeconomic status backgrounds. To do so, we leveraged three fagship longitudinal data sources for Australia: the *Longitudinal Survey of Australian Youth* (LSAY), the *Australian Census Longitudinal Dataset* (ACLD), and the *Household, Income and Labour Dynamics in Australia* (HILDA) Survey.

At the *access* stage, we used the 2006 LSAY data to investigate the relevance of various barriers to university enrolment identifed in previous literature. This analysis demonstrated large gaps in university enrolment by socio-economic status, and confrmed the relevance of various barriers, including material, cognitive, noncognitive, aspirational, cultural and institutional barriers, to accessing higher education. We further show that these barriers disproportionately manifest among low socio-economic status students at the age of 15, compared to their high socioeconomic status peers. At the *participation* stage, we focused on the relative infuence of various markers of socio-economic advantage or disadvantage operating at various levels: the individual, area and school levels. We used ACLD data to demonstrate that socio-economic disadvantage at all these levels matters for university participation at the age of 18–22. At the *post-participation* stage, we leveraged the HILDA Survey data to assess the differences in post-university outcomes for graduates from low socio-economic status and high socio-economic status backgrounds. We evaluated outcomes spanning the domains of work and health, demonstrating that the effects of socio-economic background extend beyond graduation—at least for some of the outcomes. While we fnd some evidence of a 'catch up' effect for low socio-economic status graduates, it does take them several years to achieve outcomes that are on par with those of their high socio-economic status peers.

The fndings collated throughout this chapter bear important implications for educational and social policy. Overall, they suggest that, in contemporary Australia, socio-economic background continues to play a role in shaping outcomes at various points of individual's educational trajectories. This is manifested by the lower chances that low socio-economic status individuals have to access and participate in higher education, and to fnd satisfying and secure employment in the frst few years post-graduation. Specifc fndings presented in this chapter point to the presence of various barriers to accessing university, which is already evident among secondary school low socio-economic status students. This fnding underscores the need for early intervention to align educational opportunities and trajectories experienced by low socio-economic status students with those of their high socio-economic status peers. For such barriers not to manifest at the age of 15, much earlier interventions—most likely starting in primary school—are required. Further, our analyses demonstrated the presence of multiple barriers that span various domains and dimensions of disadvantage, consistent with the life course principle of interdependence of life domains. This piece of evidence indicates that any interventions must be multi-pronged—focusing on an array of needs, instead of prioritising a single dimension (such as material disadvantage). Further, our analyses of individual-, area- and school-level factors infuencing university participation highlight the importance of developing such interventions with attention to particular context geographical, school and community context. Moreover, because our analyses demonstrate that socio-economic disadvantage across various levels has independent effects on educational outcomes of young people, interventions should ideally be designed to be simultaneously delivered across multiple levels, for instance targeting individual, schools and communities at the same time. Finally, the evidence showing that socio-economic disadvantage extends to the post-graduation stage makes a stronger case for the importance of *sustained* interventions—even if we do intervene early, many low socio-economic status people will require additional support down the track.

Our research also points to avenues for further research. For instance, we note opportunities to further strengthen the evidence base in this space by leveraging administrative data covering whole populations, rather than large samples from these populations. Only with these data we will be able to estimate more precisely the extent of educational and socio-economic disadvantage, particularly for smaller and vulnerable groups (e.g., Indigenous students or students from refugee backgrounds) (Perales, 2021). Second, we are only able to present a 'piece-meal' picture of how student life cycles differ for low and high socio-economic status individuals. As ft-for-purpose data become available, future studies could attempt to track the same individuals over the full course of their student life cycle—from pre-school and school, through university, and beyond. Again, emerging administrative datasets offer distinct opportunities to accomplish this. Such a holistic analysis would also enable more comprehensive engagement with key concepts pertaining to the life course approach, such as consideration of longer-term trajectories, turning points, sensitive and critical periods, and sequences of transitions across multiple statuses.

**Acknowledgements** Parts of the chapter are based on materials included in Tomaszewski et al. (2018), which was funded by the Australian Department of Education and Training. Parts of the chapter also draw on materials published in Tomaszewski et al. (2021), which was supported from a grant by the National Centre for Student Equity in Higher Education (NCSEHE).

#### **References**


Burke, P. J. (2012). *The right to higher education: Beyond widening participation*. Routledge.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 8 Emerging Adulthood in Australia: How is this Stage Lived?**

**David C. Ribar and Clement Wong**

Young people's transitions to adult roles, including completing school, entering full-time work, leaving their parents' home, marrying, and becoming parents, are occurring at older ages and more heterogeneously in Australia and other developed countries (see, e.g., Cobb-Clark, 2008; Elzinga & Liefbroer, 2007; Furstenberg, 2010; Settersten & Ray, 2010). Whereas many of these transitions occurred at or near the time of reaching legal adulthood in the 1960s and 1970s, they often occur in people's mid- and late-20s—or later—today. This has created a new period in the life course of "emerging adulthood" in which young people are legal adults but also dependent on their families and not fully autonomous. As a life course period, emerging adulthood entails developmental and growth opportunities, especially for young people who engage in schooling, training, work experiences, developing relationship skills, and other activities. It is also shaped by a host of contexts, including family environment, economic opportunities, institutional supports, and social norms.

This chapter takes a life course perspective to review studies that have examined emerging adulthood and to examine empirically the patterns and conditions of emerging adulthood in Australia. As described in Chap. 2, the life course perspective is a multidisciplinary conceptualization of the sequence of socially meaningful transitions in a person's life that places those transitions in overlapping contexts of family, community, and society. The perspective extends theories of development to

D. C. Ribar (\*)

C. Wong

Andrew Young School of Policy Studies, Georgia State University, Atlanta, GA, USA e-mail: dribar@gsu.edu

Society and Lifecourse Research team, Australian Institute of Family Studies, Melbourne, Australia e-mail: clement.wong@aifs.gov.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_8

a person's entire lifetime and also incorporates ecological theories. In addition to these theoretical considerations, the chapter empirically analyses 2001–2018 data from the *Household, Income, and Labour Dynamics in Australia* (HILDA) Survey, which conducts annual interviews of people aged 15 years and older in Australian households and follows those people over time, including when they leave their original households. As with many previous analyses, the chapter documents the numbers of young people who can be categorized as emerging adults, how these numbers have changed over time, and how they differ across groups. More distinctively, however, the chapter examines the conditions of emerging adulthood—that is, how this stage is lived.

#### **Emerging Adulthood**

#### *Defning Emerging Adulthood*

The introduction of the term "emerging adulthood" is widely attributed to Arnett (2000), who used it to describe a distinct and historically new developmental period that extended from people's late teens to their early 20s. Arnett conceptualized emerging adulthood as being bracketed by and distinct from people's adolescence and early adulthood. He emphasized the amount of choice and volition during this period regarding eventual adult roles and characterized it developmentally in terms of identity formation, especially in the areas of romantic relationships, career, and "worldviews." Others (see, e.g., Côté, 2014, Côté & Bynner, 2008), while acknowledging the growing prevalence and in-betweenness of emerging adulthood, have sharply criticized Arnett's developmental framing. Far from a period of choice and volition, they emphasise how emerging adulthood likely refects constraints from reduced economic opportunities, high housing costs, and inadequate social supports. They further see the developmental framing as dangerous as it distracts society and policymakers from addressing young people's growing structural disadvantages. The life course perspective helps to reconcile both approaches by positing that developmental and structural explanations must be considered together, rather than exclusively. Of course, in addition to researchers who have used the term "emerging adulthood" formally, many others have used it more loosely to categorize young adults.

Just as there is debate on the defnition and meaning of emerging adulthood, there is debate on when this period ends and early adulthood begins. Consistent with the notion of volition, Arnett's (2000) initial formulation saw the end in terms of subjective elements of young people feeling responsibility over life domains and having the ability to make independent decisions. More recently, Arnett (2007) has described the end of emerging adulthood in terms of undertaking, though possibly not completing, the major adult transitions of fnishing school, starting work, living independently, partnering, and parenting. Benson and Furstenberg (2006) directly asked young Americans about whether they considered themselves to be adults. In support of Arnett's demarcation, they found that adult identity was strongly tied to living independently and becoming a parent but was only moderately tied to completing schooling or working. Interestingly, they also found that reversals in transitions, such as having to move back in with parents, reduced young people's identifcation as adults.

In this chapter, we will use the term emerging adulthood to describe people who have reached the legal age of adulthood (18 years in Australia) but who remain residentially and fnancially dependent on their families and thus have not taken on all the responsibilities or gained the full autonomy of adulthood. Our defnition differs from some others by demarking the end of the period in terms of the behavioural outcomes of attaining residential and fnancial independence, which occur at different ages for different people. Because some people attain both forms of independence by age 18, our defnition also implies that emerging adulthood is not universally experienced. Our analyses examine partnering and parenting, but we do not include them in our defnition because recent cohorts of young Australians tend to experience these outcomes many years after achieving residential and fnancial independence.

#### *Characteristics Associated with Receiving Support*

Logically, parents' economic resources will affect their ability to support their young adult children. Empirical research has found support for this conjecture across many different measures of resources. Specifcally, coresidential and fnancial support tend to be higher for parents with greater fnancial wealth (Schoeni, 1997) and higher socioeconomic status (Angelini et al., 2020) and lower for parents with long histories of welfare receipt (Cobb-Clark & Gørgens, 2014), histories of poverty (Kendig et al., 2014), and fnancial stress (Cobb-Clark & Ribar, 2012). Having more children also limits parents' effective resources and has been found to reduce parental support (Schoeni, 1997). In addition to these characteristics, Cobb-Clark and Ribar (2012) have found that confict between parents can prompt young adults to leave home sooner.

Characteristics of the child who might receive support are also important. To the extent that parents' motivation for providing support and children's willingness to accept support are guided by concerns for children's wellbeing, exchanges of support will rise with children's needs. Consistent with this, empirical evidence shows that material support increases with different measures of children's needs (Hartnett et al., 2013; Schoeni, 1997) and falls as children age (Angelini et al., 2020; Hartnett et al., 2013). Within Australia, norms for supporting adult children also increase with children's needs and fall with children's ages (Drake et al., 2018). Co-residential and fnancial support have also been found to be lower for children with emotional distress (Sandberg-Thoma et al., 2015). Support falls as young adult children identify more strongly as adults (Hartnett et al., 2013), and it is lower for daughters than sons (Cobb-Clark, 2008). More closely examining gender differences, Wong (2018) found that young men who experienced negative life shocks, such as the illness of a close friend, tended to remain at home longer while young women left home sooner. Wong also found that heavy drinking contributes to young men and women leaving home earlier and that an early initiation of smoking contributes to young women leaving earlier.

Support may also be affected by characteristics of the parent-child dyad, especially the relationship between the parents and their young adult children. However, the evidence regarding the effects of parent-child relationships on support is mixed. Cobb-Clark and Ribar (2012) have found that unsatisfactory parent-child relationships are associated with earlier nest-leaving. However, Gillespie (2020) fnds that close parent-child relationships are positively associated with leaving home, especially for the relationships between mothers and daughters. The conficting fndings may be explained by relationship quality and support mutually affecting each other. For example, Johnson (2013) has found that fnancial assistance at one age is associated with children being closer to parents at a later age.

#### *Effects of Support*

If emerging adulthood serves a developmental purpose, we would expect to see benefts associated with it. The largest advantages appear to be in terms of completing more schooling (Swartz et al., 2017; Wong, 2018, see also the review by Cobb-Clark, 2008). However, evidence on the associations of emerging adulthood with other outcomes is equivocal. Fingerman et al. (2015) found that intense support from parents increased young adult offsprings' sense of defned goals and life satisfaction, and as mentioned, Johnson (2013) uncovered a positive association between parent support and subsequent close parent-child relationships. However, Johnson also found that parental support increased young adults' depressive symptoms, and Mortimer et al. (2016) found that support reduced young adults' self-effcacy. In a more developmentally focused study, Roisman et al. (2004) found that three developmental tasks that were salient to adolescence—building friendship, academic, and conduct skills—were predictive of subsequent adult success. However, two tasks identifed by Arnett (2000) for emerging adulthood—building work and romantic skills—were not consequential for later adult success.

#### *What Emerging Adults Do*

Although many studies have been conducted of who emerging adults are, how long emerging adulthood lasts, and whether emerging adulthood is consequential for later outcomes, Furstenberg (2010) and others have noted that much less research has considered emerging adulthood as it is lived. The authors have contributed to several of the available studies. Wong (2018) used the HILDA Survey to examine smoking and heavy drinking among co-resident emerging adults. Botha et al. (2020) used the HILDA Survey to examine how fnancial autonomy grew and developed over the course of emerging adulthood for men but not women. Ribar (2015) also used these data and found that emerging adults suffered fewer fnancial hardships than newly independent adults. In addition to these studies, Hartmann and Swartz (2007) interviewed emerging adults to document their understandings of their roles. They found that emerging adults see themselves positively in a new and distinct role that is shaped by dynamism.

#### **Description of the Analysis Sample from the HILDA Survey**

Our empirical analyses draw data from the frst 18 waves of the *Household, Income, and Labour Dynamics in Australia* Survey. The HILDA Survey is a large national longitudinal survey that began with 19,914 people in 7682 Australian households in 2001 and has subsequently followed those people and their families in annual interviews. Each wave asks every household member aged 15 years and older about personal and household economic conditions, demographic circumstances, life changes, and other characteristics through multiple instruments, including face-toface interviews and self-completed questionnaires. Importantly, the HILDA Survey continues to follow members and children of original sample households even after they leave their initial homes, allowing us to examine young people while and after they co-reside with their families. Attrition has been modest; by the 18th wave, 62% of the original survey respondents completed interviews (Summerfeld et al., 2019). We extracted the HILDA data with the PanelWhiz add-on for Stata (Hahn & Haisken-DeNew, 2013).

For our analyses, we select annual observations for 11,660 HILDA Survey respondents who were ever observed when they were 15–25 years old. Because of the longitudinal structure of the survey, the sample includes multiple annual observations for most of the respondents. Some of our analyses include observations for adolescents, aged 15–17 years, and some are restricted to young adults, aged 18–25 years. Although the HILDA Survey provides weights that address sampling issues and attrition for the general sample, we only conduct unweighted analyses because we are using a selective sample.

As with the analysis by Cobb-Clark and Gørgens (2014), we consider two forms of support that parents might provide. The frst is co-residence with one or both of the young person's parents, which is measured at the time of the HILDA Survey interview. The second is an indicator for whether the young person received a direct fnancial transfer (of any amount) from his or her parents. Most of our analyses distinguish among young people who receive both co-residential and fnancial support, receive only co-residential support, receive only fnancial support, or receive neither type of support. Our analyses label young people who receive one or both types of support as "emerging adults" and those who do not receive either type of

support as being "independent." In several analyses, we further distinguish between young adults who have been independent for three or fewer consecutive years and those who have been independent for longer, as previous research by Ribar (2015) shows that young people's conditions change with the duration of independence. We also examine young people who "boomerang" back into co-residence with parents after living apart for at least one wave of the HILDA Survey.

Our analyses consider many other measures from the HILDA Survey of conditions and characteristics of the young people, which we explain within each analysis.

#### **Parental Support in Adolescence and Emerging Adulthood**

We begin our empirical analysis by examining the percentages of Australian adolescents and young adults who co-reside with their parents, receive fnancial transfers, or both. Figure 8.1 shows these percentages at each age from ages 15 to 25, using all the data from the HILDA Survey from 2001 through 2018. We consider young adults who co-reside with or receive transfers from their parents to be emerging adults, so the graph shows the extent of emerging adulthood at each age.

**Fig. 8.1** Receipt of parental support, by age. (Source: Household, income and labour dynamics in Australia survey)

The estimates indicate that at age 15 almost all (98% of) young Australians coreside with their parents. Co-residence falls with age but remains very prevalent through adolescence. At age 17, 90% of young people are co-residing with parents. The incidence of co-residence falls more noticeably starting at age 18. At age 21, the incidence of co-residence drops below 50%, and by age 25, only 21% live with their parents.

The incidence of fnancial transfers is much lower but follows the same age pattern. At age 15, just under half of young Australians are receiving transfers from their parents. By age 18, about a third are receiving transfers, and by age 25, only 9% are receiving transfers. In adolescence, transfers are mostly received in conjunction with co-residence. By age 22, transfers are slightly more likely to be received while living independently than while co-residing.

The estimates also reveal the heterogeneity in coresidential and fnancial support. For example, at ages 18 and 19, most young people are receiving some type of support from their families, but substantial fractions are not and thus effectively skip any period of emerging adulthood. By age 22, nearly equal proportions of young people are and are not receiving support, and by age 25, most young people are independent, but a substantial fraction (27%) are not. Consistent with the fndings for other countries by Elzinga and Liefbroer (2007) and others that transitions to adulthood are heterogeneous, the estimates indicate that there is no single, uniform path through emerging adulthood in Australia.

We next consider whether and how young Australians' receipt of coresidential and fnancial support changed from 2001 to 2018. Figure 8.2 shows the year-byyear percentages of people aged 15–17 years old (left panels), 18–21 years old (centre panels), and 22–25 years old (right panels) in the HILDA Survey who coreside with their parents (top panels) or receive fnancial support from their parents (bottom panels). The incidence of co-residence among adolescents (people 15–17 years old) has remained relatively constant within a range of 92–95% over the period. In contrast, co-residence among 18- to 21-year-olds fell modestly from 2001 to 2006 and then generally grew from 2006 to 2018. On net, co-residence for 18- to 21-year-olds increased over the period. Co-residence among 22- to 25-yearolds was relatively constant from 2001 to 2010 but grew after 2010 and increased on net over the entire period. Overall, co-residential support for young Australians has increased since 2010, especially among 22- to 25-year-olds.

The patterns for fnancial support are different, with the prevalence increasing for adolescents and both groups of young adults and with the increases being larger for adolescents and 18- to 21-year-olds. There was no discernible trend for any of the groups in the early 2000s but a strong upward trend since 2009 and an upward trend on net over the entire period. When we consider co-residential and fnancial support together, it is clear that the proportion of 18- to 25-year-old Australians who are emerging adults has increased.

**Fig. 8.2** Co-residential and fnancial support from 2001 to 2018 among different age groups. (Source: Household, income and labour dynamics in Australia survey)

#### **Characteristics, Activities, and Resources of Emerging Adults**

The preceding analyses indicate that young people differ considerably in whether and when they co-reside with parents and receive fnancial support. As we will show in the next analyses, the heterogeneity in co-residence and fnancial support extends beyond young people's ages and occurs across other characteristics. Additionally, we consider how young people's activities and resources differ conditional on the dependence or independence from parents. For each of these analyses, we consider cross-tabulated results, which reveal associations between young people's dependence and other outcomes but not causal relationships.

#### *Characteristics and Activities*

We consider the characteristics and activities that are associated with emerging adulthood in Table 8.1. Table 8.1 reports means of 18- to 25-year-old Australians' characteristics and activities conditional on whether they receive both co-residential and fnancial support from parents (column 1), receive only co-residential support (column 2), receive only fnancial support (column 3), have been independent for 3 years or less (column 4), or have been independent for more than 3 years (column 5). It also reports the means for the entire sample of young adults (column 6). The frst three columns indicate characteristics and activities of emerging adults, while the next two columns examine independent adults.

The top panel of Table 8.1 reports conditional averages of several characteristics. The frst row reports the young people's average ages and, consistent with Fig. 8.2, indicates that the receipt of support is more likely to occur among younger rather than older people. The next rows report averages for other demographic characteristics, which can be compared against the overall sample average in the rightmost column. In general, women are more likely than men to be independent of parents; however, women are also more likely to only receive fnancial support from their parents. Young Indigenous Australians are more likely than other young Australians to be independent of parents, while young people from migrant families are less likely to be independent. Young people from single-parent families and in rural areas are more likely to be living with their families but less likely to only be receiving fnancial support than those from other families. High school graduates are less likely to be independent of parents, while college graduates are more likely to be independent.

The bottom panel of Table 8.1 reports the average levels of several activities and conditions of young people. Two-thirds of young Australians who are co-residing and receiving fnancial support are studying, while very few of those who are independent are still studying. Employment is common among all young Australians and especially prevalent among those who are independent. Although most emerging adults work, the majority of those jobs are on casual contracts, and their average earnings tend to be low. Thus, many emerging adults have transitioned to work but not to permanent, well-paying jobs. Household-level rent is lower once living independently—independent young adults may live in smaller/lower-cost accommodation than the parental home. However, among those living independently, household rents are slightly higher for young people receiving fnancial support, which may assist with renting better housing. Rates of marriage and partnership and of parenthood are very low for young Australians aged between 18 and 25, but the rates are proportionally much higher among those who are independent and those who live apart from parents but receive fnancial support from them.


166


Source: Household, Income and Labour Dynamics in Australia survey

Notes: Authors' calculations of averages for each listed measure for 39,121 annual observations of 18- to 25-year-olds from the HILDA Survey

Emerging adulthood is conceptualised to involve autonomy and responsibility in some life domains but not others, and the estimates from Table 8.1 provide strong evidence for this. One potential area of autonomy is engaging in risky health behaviours. Rates of consuming fve or more standard alcoholic drinks in a day are higher among emerging adults than young independent adults; however, rates of smoking are lower among emerging adults. Co-resident emerging adults also report moderately high levels of autonomy over their social lives and work choices, but young adults who live independent of parents report even more autonomy. In contrast to these domains with moderately high autonomy, co-resident emerging adults report little responsibility or autonomy for day-to-day spending in their households, large purchases, or fnancial matters.

#### *Wealth, Debt, and Financial Wellbeing*

Table 8.1 indicates that most emerging adults work and receive independent resources from earnings. In addition to these fows of resources, we can examine other aspects of their fnancial lives, including their bank account holdings, debts, superannuation, and fnancial wellbeing. Table 8.2 reports statistics on these outcomes. It is organised the same way as Table 8.1 but reports statistics for fnancial characteristics. Because many fnancial outcomes have skewed distributions (many low values but a few very high values), the table lists averages (the frst fgure in each table cell) and medians (the second value in each cell, listed in parentheses) of the outcomes.

For the period that we study, the average individual bank account balance for Australians aged 18–25 years old was just under \$4800, but the median balance was \$900, indicating that the distribution is very skewed towards low balances. The balances also provide little in the way of a savings buffer—the median balance of \$900 amounts to less than a week and a half of young workers' average earnings. Young people who are receiving fnancial support and those who are independent have smaller average account balances, while those who are co-residing with parents but not receiving fnancial support have larger average balances. Young people who are independent of their parents also have lower median balances. The results suggest that co-residence helps emerging adults build savings, especially when we consider the younger average ages and lower earnings of co-resident young adults. The higher balances among co-residing young people who do not receive fnancial support further suggest that these transfers may be conditioned on need.

Few young Australians held joint bank accounts. The median balance for all 18 to 25-year-olds and for each of the subgroups was zero. However, average balances were higher for those who were independent of their parents, which is consistent with their higher rates of partnership and marriage. Similarly, few young Australians had credit card balances, and the average balances were low. In contrast, average Higher Education Contribution Scheme (HECS) and Higher Education Loan


**Table 8.2** Financial resources and wellbeing of 18- to 25-year-old Australians by receipt of residential and fnancial support

Source: Household, Income and Labour Dynamics in Australia survey

Notes: Authors' calculations of averages and medians (in parentheses) for each listed measure for 39,106 annual observations of 18- to 25-year-olds from the HILDA Survey

Programme (HELP) balances and other debt balances are higher, refecting the high rates of study particularly among emerging adults, though median balances are zero.

Young adults hold modest superannuation balances. Balances are very low for emerging adults who are both co-residing with and receiving fnancial support from parents, and balances are moderate for young adults who are living independently. These results, too, are consistent with independent adults being older, earning more, and having longer work histories.

When we consider other measures of fnancial wellbeing, we see that most young people would struggle to raise funds in an emergency, though emerging adults report slightly less diffculty than young independent adults. Nearly a quarter of young co-resident adults and 37–51% of non-coresident adults also report at least

one experience of material deprivation as a refection of fnancial hardship, such as not being able to pay a utility bill, in the previous year. Despite these negative indicators of fnancial wellbeing, young Australians report moderately high average satisfaction with their fnancial situations. The fndings for fnancial wellbeing mirror those reported by Ribar (2015).

#### **Differential Experiences in Becoming Independent**

#### *Timing of Independence*

The longitudinal data from the HILDA Survey allow us to investigate more closely the timing of young people's transition from emerging adulthood to independence from parents. For this analysis, we select observations from our general sample of 18- to 25-year-olds for young people who are continuously interviewed from age 15 on. We follow them until the frst interview at which they are neither co-residing nor receiving fnancial support from parents, age 25, or they attrit from the sample. For people who are observed to transition to independence, we can measure the duration to this event. For people who reach age 25 or attrit from the sample without transitioning, we know that the transition occurred after the last time we observed them. To address the partial loss of information from the people whose transitions are not observed, we estimate Kaplan-Meier, or life-table, hazard probabilities of transitioning to independence and use these hazard probabilities to estimate the median age of transitioning (formally, the point at which the cumulative hazard function reaches 50%). Table 8.3 reports these median ages for the entire sample and for subgroups defned in terms of household income, parents' household structure, indigenous background, and migrant background; Fig. 8.3 presents the corresponding probabilities of "surviving" (that is, continuing to receive parental support) across the age profle.

Consistent with the characteristics of emerging adults described previously, the median timing to independence also differs by economic resources and demographics. Overall, 50% of emerging adults become independent of co-resident and fnancial supports by age 22.2. When considering household income as ranked at age 15, emerging adults from more affuent households tended to establish independence at older ages, indicating that those from an advantaged or well-resourced household receive parental support for a longer period of time. A similar pattern emerges for both household structure and indigenous status, where emerging adults from singleparent households or from an indigenous background also tend to become independent sooner. This is especially pronounced for Indigenous Australians, where the median age of independence is 19.1. Conversely, emerging adults with a migrant background tend to receive parental support for longer, with a median independent age of 23.7.


**Table 8.3** Median age at independence by family background factors

Source: Household, Income and Labour Dynamics in Australia survey

Notes: Authors' calculations of median ages based on cumulative hazard functions of transitioning to neither co-residing nor receiving fnancial support among HILDA Survey respondents who are continuously followed from age 15

#### *Prevalence of Returning Home*

We use a similar event-history methodology to examine the probability that young people who transition to independence later "boomerang" and return home. For this analysis, we select young people in the HILDA Survey who are observed to transition to independence. We then follow them until they return home, reach age 30, or attrit from the HILDA Survey. We use the Kaplan-Meier approach to calculate the cumulative hazard of the years until they return home, presenting the information in Table 8.4. As with the previous analysis, the cumulative hazard accounts for the partial loss of information from people who reach age 30 or attrit from the sample without being observed to make a transition. Persons who return home within their frst year of independence cannot be distinguished in the HILDA survey from others who continued to receive parental support; there are consequently no "boomerang" cases in the frst year.

The cumulative hazard of returning home is initially quite steep. Over 9% of young adults return to their parents' home in the year after becoming independent. This indicates that for some young Australians—intentionally or otherwise—independence is a brief or temporary arrangement, and parental support is accessed soon afterwards. Although the cumulative hazard rate continues to increase in subsequent years it also fattens out substantially; young adults who have been independent for four or more years have likely left the parental home for the last time.

#### Kaplan Meier Survival Estimates

**Fig. 8.3** Kaplan meier survival estimates of receiving parental support among different demographic groups. (Source: Household, Income and Labour Dynamics in Australia survey)


**Table 8.4** Cumulative hazard probabilities of "Boomeranging"

Source: Household, Income and Labour Dynamics in Australia survey Notes: Authors' calculations of cumulative hazard functions of returning home after becoming independent among HILDA Survey respondents who are continuously followed after becoming

#### **Conclusion**

independent

Our analyses of longitudinal data from the HILDA Survey examine emerging adults—people who are 18–25 years old and who co-reside with parents and receive fnancial support from them. We fnd that a majority of young Australians who are 22 years old or younger are dependent in this way and that substantial minorities of 23- to 25-year-olds also meet the defnition of being an emerging adult. We also fnd that the proportion of young people who are emerging adults has grown over time, especially since 2009.

Emerging adulthood is conceptualised to be distinct from both adolescence and from independent adulthood. Consistent with this conceptualisation, we fnd that young people who are emerging adults have autonomy in some spheres of their lives but not others, while young people who are independent adults have much more autonomy in all the spheres that we can measure. We fnd that most emerging adults are enrolled in school and therefore investing in better adult opportunities. We also fnd that most emerging adults work. However, their work often occurs through casual jobs and with low earnings, so they have not completed the adult transition to permanent, economically sustaining work. Similarly, few have made the transition to partnering or to bearing children.

As with many other forms of human development, young people from privileged backgrounds appear to have more opportunities to invest through emerging adulthood. Young people with high-income parents receive co-residential and fnancial support longer than young people with low-income parents. Similarly, non-Indigenous young people and young people from two-parent families receive support for longer than Indigenous Australians or young people from single-parent backgrounds. However, there are some other patterns, with young people from migrant families and rural areas receiving assistance longer.

**Acknowledgements** The authors thank Janeen Baxter and Deborah Cobb-Clark for many helpful discussions.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 9 Labour Market Participation: Family and Work Challenges across the Life Course**

**Barbara Broadway and Guyonne Kalb**

Over the past few decades, the labour market in Australia has changed drastically, as it has in many other countries (Thévenon, 2013). Figure 9.1 shows this clearly. Female labour force participation in 2021 was 1.5 times as high as in 1978: steadily increasing over this period from just over 50% to just over 75% for women of working age. Male labour force participation has not changed as much, except that it slightly decreased over four decades: from just over 87% to around 83% for working-age men. Employment rates followed a similar pattern to participation rates, except that it is more affected by the shocks of increased unemployment during recessions, especially in the early 1980s and early 1990s recessions. Small dips in employment are also observed in the Global Financial Crisis of 2008 and the COVID crisis of 2020, with the latest dip being a particularly sharp one. Women were affected to a lesser extent than men in most of these crises, except in the 2020 COVID-related crisis, but they had recovered largely by March 2021.

Hérault and Kalb (2020) examine the labour market outcomes of prime workingage individuals (25–55 years) from 1994/95 to 2015/16, using the Survey of Income and Housing (aka Income Distribution Survey) collected by the Australian Bureau

B. Broadway

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Melbourne Institute: Applied Economic & Social Research, University of Melbourne, Melbourne, Australia

e-mail: b.broadway@unimelb.edu.au

G. Kalb (\*)

© The Author(s) 2022

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Melbourne Institute: Applied Economic & Social Research, University of Melbourne, Melbourne, Australia

Institute of Labor Economics (IZA), Bonn, Germany e-mail: g.kalb@unimelb.edu.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_9

**Fig. 9.1** Monthly labour force participation and employment of 15–64 year old men and women in Australia – February 1978 – March 2021 (in percentage). (Source: Trend numbers from the Labour Force, Australia, Australian Bureau of Statistics catalogue number 6202.0)

of Statistics. Their data show that employment has increased dramatically among partnered women and single parents (by 13 and 17 percentage points, respectively). Meanwhile, single men were the only group whose employment dropped, by 4 percentage points. These changes happened against a backdrop of slowly, but steadily increasing employment rates overall, including for partnered men and single women. Average hours worked changed in similar ways to these employment rate changes.

One aspect of employment that has changed substantially for men and women is the proportion of workers who are in a full-time job. A decrease of around 14 percentage points for men and 11 percentage points for women is observed in the over four decades that have passed since the late 1970s (see Fig. 9.1).1 The larger decrease for men is from a 95% high compared to 66% for women; and it shows the change in the role of employment in people's lives over the past few decades. Theoretically, this shift in the role of full-time employment should have allowed both men and women to obtain a better balance between family and work over the life course.

Increased employment among partnered women and single parents implies – more or less directly – increased employment of mothers. This brings with it the

<sup>1</sup>The larger decrease for men seems mostly due to a reduction for single men (Hérault & Kalb, 2020).

potential for increased confict between one's working and home life, as competing responsibilities need to be fulflled. The observed shift in the role of full-time employment may represent one tool used by families to obtain a better balance between family and working lives. Changes in workplace conditions could fulfl a similar function – including fexible start and fnishing times, telecommuting or employer-provided parental leave. The *Household, Income and Labour Dynamics in Australia* (HILDA) survey is the perfect tool to assess this, as it asks all individuals in selected households a range of questions on their circumstances, including employment and labour force participation, in addition to a wide array of questions on their wellbeing and subjective views. The same households have now been interviewed for more than 20 years, starting in 2001.

The majority of changes in the proportion of part-time employment have occurred in the 1980s and 1990s, and thus changes observed in HILDA are modest, especially for women. However, the HILDA data allow us to observe the reason for working part time: Fig. 9.2 (panel a) shows that men are now more likely to say that they work part time for family reasons than in 2002. It has increased by 1 percentage point (from 6.1% to 7.1%) and for women it decreased by 2.8 percentage point (from 39.9% to 37.1%). This suggests there may be a (slow) change in the division of paid work and caring roles at home.

There was a substantial increase in the proportion of employees (from 45% in 2002 to 71% in 2018), who indicate that their employer provides paid maternity leave (Fig. 9.2, panel b). This should further assist parents in achieving a better balance between family and work. However, there has been little change in employers providing childcare facilities or subsidies to assist with childcare expenditures, which has remained around (a low) 9% of respondents' employers.

Although there has not been that much change in the proportion of workers who do any work from home (increased by just over 2 percentage points) or in the hours worked from home on average (increased by just over 1 h), there are more likely to be formal arrangements in place for working from home (for nearly 45% of workers by the end of 2018 versus 33% in 2002) and the proportion of workplaces that allow their employees to do home-based work has increased from 19% to 32% between 2002 and 2018. It should now be easier for employees (who want this) to work from home as the infrastructure for this is in place. With the COVID-19 pandemic necessitating work-from-home for many employees in 2020, working from home may experience a larger increase in the next few years than appeared possible in 2019.

Putting all these (modest) changes together, we expect some improvements in parents' ability to balance work and family. This is confrmed in practice when examining the proportion of men and women who are not satisfed with the fexibility they have at the workplace to balance work and non-work commitments. This decreased from 32% to 26.4% for men and from 29.4% to 24.7% for women between 2002 and 2018. Nevertheless, combining family responsibilities and employment remains challenging.

In the next section, this chapter examines these challenges for different types of families and at different points in the life course, considering families with children frst and then families without children, focussing on partnering and fertility

**Fig. 9.2** Family-friendly policies used by employees and offered at workplaces, 2002 versus 2018. Notes and defnitions: The sample is restricted to male and female employees up to age 65, who are not in full-time education. Results are weighted using cross-sectional weights. Employees are considered to be working part-time, if they report less than 35 working hours per week in a usual week across all their jobs. Reason for working part-time is self-reported main reason for working part-time; respondents can choose between 13 response categories. Family reasons combines the response categories "caring for children", "caring for disabled or elderly relatives (not children)" and "other family responsibilities". Whether respondent is working from home and

decisions. This is followed by a consideration of the family and work interactions at different stages of the life course: early life, mid-life and later in life. The chapter concludes with a section discussing policy implications.

#### **Family-Work Challenges in Early Adulthood**

Balancing family life and work life can be challenging as both academic and anecdotal evidence show. Particularly the impact of having children on employment is well documented in a large literature, but there may also be a reverse impact from employment (or lack thereof) on family formation decisions, and especially on fertility. This section frst discusses how parenthood and resulting work-familyconficts affect labour market outcomes. We focus on the role that skills, disadvantage, and policy and institutions can play in alleviating or exacerbating such conficts, and hence in shaping labour market outcomes. In the second part of this section, we explore the reverse relation: how do job situation, skills and earnings potential affect the decision to have children?

#### *The Effects of Parenthood on Labour Market Outcomes*

#### **Direct Effects of Parenthood on Labour Market Outcomes**

Although relatively recent developments in the labour market have substantially narrowed the gap between male and female labour force participation over the past decades, the difference between men and women is still large, and in Australia, the gap is still considerably wider than in, for example, the Nordic countries. Even in the Nordic countries women have slightly lower labour force participation than men once children arrive, but the drop in labour force participation is much smaller than in Australia (Kalb & Thoresen, 2010). OECD data show that in 2019, Australia's maternal employment rate was lower than the average across OECD

**Fig. 9.2** (continued) whether a formal arrangement is in place, are self-reported variables. Satisfaction with workplace fexibility is self-reported on a scale from 0 [Totally dissatisfed] to 10 [Totally satisfed]; respondents are coded as not satisfed if they report a satisfaction level of 6 or lower. For workplace entitlements, respondents are asked if they, or other employees working at a similar level to them at their workplace, would be able to use these if needed. "Family leave" combines employer-funded paid maternity leave, employer-funded paid paternity leave and Special leave for caring for family members. Childcare facilities/subsidies refers to childcare services provided by the employer, or childcare subsidies provided by the employer to purchase childcare services in the market. (\*) This question was not asked in wave 2002, the reported numbers for this category are calculated using HILDA wave 2006 and wave 2018. (Source: Household, Income and Labour Dynamics in Australia (HILDA) Survey, waves 2002, 2006 and 2018)

countries, and it ranked about one third from the back compared with around 40 other countries.2 This ranking was similar to that in 2014 (Kalb, 2018), although the gap between Australia and the countries with the highest maternal employment rates is getting smaller, as Australia's maternal employment rates continue to increase, and increased by more on average than those in the highest performing countries. In 2019, Australia's maternal employment rate was 69% compared to 87% in Iceland and Slovenia, and 86% in Sweden (vs 62% for Australia and 82% for Sweden in 2014).

There is a large literature on how children affect their parents' labour supply, and how it differs for mothers and fathers. Disentangling the true effect of parenthood from pre-existing differences between parents and non-parents is no trivial task. The ground-breaking study by Angrist and Evans (1998) was the frst to confrm that the pattern we observe around the world is indeed a causal relationship: the presence of preschool children substantially reduces labour force participation and hours worked for women, while in most cases, men are hardly (if at all) affected by the presence of preschool children. Although the impacts are smaller than for lowereducated mothers, even when the mother is highly educated, the presence of young children affects her labour supply much more than for similar fathers. An analysis of general practitioners' and medical specialists' labour supply shows that female medical doctors (general practitioners and specialists) reduce their labour supply substantially when preschool children are present, while male specialists are not affected at all and male general practitioners are affected to a much smaller extent and only for preschool children (Kalb et al., 2018). Female general practitioners, on the other hand, reduce labour supply at least to some extent when any child under age 15 is present. This shows that even very high education – as in the case of medical doctors – does not completely eliminate the difference between fathers' and mothers' labour supply.

Although many researchers estimate the impact of children on labour supply, it is diffcult to separate pre-existing differences between parents and non-parents from the causal impact of having children: that is individuals who have children may already be different from people who do not have children before the children were born. Moschion (2013) quantifes the impact of an additional child for women who already have one child or two children, using an approach aimed at estimating the causal effect. She fnds that impacts are relatively large in Australia: going from one child to two children or going from two children to three children decreases labour force participation by 12 percentage points, and it decreases labour supply by 4 h and 3 h respectively. A wage penalty of 5% for the frst child and 9% for two children is estimated for Australia by Livermore et al. (2011). This impact of children grows over time through slower wage growth for mothers than for other workers rather than immediately lower wages after childbirth. A frst child reduces annual wage growth by 1.2 percentage points on average, while no further impact is

<sup>2</sup>See https://www.oecd.org/els/family/database.htm: the labour market position of families (LMF1.2 Maternal employment).

observed for a second child. Not surprisingly these impacts are most apparent when the child is still an infant. No differences are observed by occupational, marital or part-time work status.

Most studies investigate the immediate impact of having children, but Kahn et al. (2014) extend such research to investigate the impact of career penalties over the life course in the US. In addition to the impact of having children on mothers' wages, they examine the impact on labour force participation and occupation of mothers up to their early 50s. The impacts on all three outcomes lessen as women enter their 40s and 50s, with only mothers of three or more children facing persisting wage penalties. The Livermore et al. (2011) study only covers 8 years of data so that such a long-term view could not be taken in their analysis. A recent metaanalysis (including the Livermore et al. study) fnds an average motherhood wage penalty of 3.8% for having one child and 3.6% per child for the total number of children (Cukrowska-Torzewska & Matysiak, 2020). The authors of this study distinguish impacts for different groups of countries (including Australia in the Anglo-Saxon group), fnding that mothers in Anglo-Saxon countries experience relatively high wage penalties of around 4.6%. The study also explores the causes of the wage gaps and fnds that in the studies focussing on wage gaps for the total number of children, the gap is largely explained by human capital loss. However, in the studies on the wage gap arising from having one child (i.e., the frst child), the choice of job and occupation appears to be the main driver of the gap, with new mothers tending to choose lower paying jobs and occupations, presumably to gain better (more fexible) non-pecuniary work conditions.

Evidence from the US shows that these motherhood wage penalties have not improved much in recent decades despite women investing considerably more in their human capital now than a few decades ago. Jee et al. (2019) use panel data to estimate separate wage penalties for three periods of time: 1986–1995, 1996–2005 and 2006–2014. They fnd that the motherhood penalty did not change much over time, except for a slight decline for women with two or more children. This latter decline appears completely due to women's increased human capital investments. When controlling for these investments, mothers experience a larger wage gap now than they did two to three decades ago.

In addition to the immediate or direct "shock" of having children and the impact this has on labour supply, the presence of children also means that other, additional shocks may affect parents differently from non-parents. For example, when experiencing job loss due to frm bankruptcy, Meekes and Hassink (2020) show that pregnant women spend more time unemployed and if they are re-employed, they are more likely to reduce their hours worked and commuting distance. Men who are expecting a baby have higher re-employment rates. Similarly, women who already worked part-time or had a short commute before their dismissal (possibly because of childcaring responsibilities) hold on to these job characteristics after their dismissal at the cost of taking longer to fnd another job.

These different impacts arising from other shocks have also become quite clear when observing the impact that the COVID-19 pandemic has had on different people. COVID-19 has for example affected parents (and especially mothers) quite differently from non-parents (or fathers) in many countries. For the UK, Andrew et al. (2020) show that mothers are more likely than fathers to lose their job, spending less hours on paid work and more hours on unpaid work such as on childcare activities. Mothers who are no longer in paid employment do more additional domestic work than fathers who are no longer in paid employment. For the US and Spain, a similar conclusion is reached, leading to an increase in gender inequality (Collins et al., 2021; Farré et al., 2020). The pattern of impacts for parents was not the same in all countries: for example, in the Netherlands the impact for partnered parents (mothers and fathers) was not worse than for others, but the impact for single parents (both mothers and fathers) was more severe than for singles or for partnered parents, indicating relative gender equality of the impact and the importance of a supporting co-parent to absorb shocks (Meekes et al., 2020).

#### **The Role of Skills and Disadvantage**

When young (pre-school) children are present in the household, care needs to be provided by one of the parents or through formal or informal care by a third person. Informal care requires the presence of trusted family or friends in the neighbourhood while formal care usually is expensive and thus requires suffcient income. Families thus need either time, a social network or money to fulfl the care needs of their children. Choices can be limited if a reliable social network is not available locally (which is the case for many migrant families), leaving parents the choice between working part-time or full-time and outsourcing part of their childcare, or become a full-time stay-at-home parent. However, due to high childcare costs, even this choice is not always available (or feasible) as parents with low earning capacity may not earn enough to make labour force participation worthwhile. Although the presence of children is likely to affect the labour supply of all parents in some way, fathers and mothers who have attained higher education levels are more likely to remain in the workforce even when preschool children are present as they are better able to afford the cost of childcare (Doiron & Kalb, 2005). Single parents are even more at a disadvantage than partnered parents, as they usually need more childcare to facilitate their employment than partnered parents. For example, partnered parents may be able to organise their work hours so that one parent starts work early and fnishes early while the other parent starts work late and fnishes late thus reducing the total amount of childcare needed. As a result, employment may be less fnancially viable for single parents, especially if they have more than one preschool child.

Employers are also more likely to be supportive of high-skilled (usually highpaid) employees, and are more likely to accommodate their needs in trying to balance family and work. This was clearly evident, for example, before the introduction of universal paid parental leave in Australia in 2011. Broadway et al. (2020a) used data from the HILDA survey to show that an estimated 56.8% of 20–45 year old women in employment in Australia had access to employer-provided paid parental leave. This was concentrated among women on above-median wages (71.3% compared to 37.8% for those on below-median wages), women with higher education (77.5% for those with a university degree compared to 39.8% for those with less than a high school certifcate), and women in professional occupations (76.5% compared to 32.9% for labourers). Overall, more-advantaged women were more likely to have access to employer-funded paid parental leave than less-advantaged women. Examining current family-friendly work entitlements in 2018 HILDA data shows that more-advantaged women still have access to better work conditions than lessadvantaged women.3 As a result, it is likely to be easier for these women to have higher work hours while balancing family and work. Comparing 2002 with 2018 HILDA data, the gap between advantaged and disadvantaged women has been growing for most entitlements (except for paid maternity leave), and policy reforms (like the introduction of universal paid parental leave) may be needed to achieve more equitable conditions.

These advantages for higher-educated more-advantaged parents are compounded by the fact that highly educated men tend to partner highly educated women, and similarly men and women with low levels of education are more likely to partner each other. As a result, the increased female labour force participation has created a widening divide between two-earner families where both parents earn a high income and one-earner families where one parent earns a low to medium income, but high childcare costs may preclude the other parent from re-entering the labour market. Although this may be the best choice for the latter family in the short-term (due to a secondary earner not being fnancially worthwhile if the additional income would mostly go towards paying for childcare), in the long run, this absence from the labour market can have major implications for the primary carer. Re-entry in the labour market at a later point in time may be diffcult, leading to continued low income for the primary carer and lower household income for her/his family.

#### **Effects of Policy**

Well-designed family policies can mitigate some of the impacts of having children and the resulting diffculties in balancing family and work. Such policies are especially important for low-skill and/or low-wage parents to avoid disadvantage being reinforced in the child-rearing years and being transmitted from one generation to the next.

The previous sections already mentioned the importance of childcare in enabling (both) parents to return to employment, and the often high costs associated with formal childcare. Policies that subsidise these costs, particularly for low-income

<sup>3</sup>For example, using HILDA wave 18 data and distinguishing between women with and without a tertiary education, we fnd that women with tertiary-level qualifcations are more likely to have access to carer's leave (85.2% vs 75.4%), paid (employer-provided) maternity leave (80.8% vs 68.2%), parental and/or paternity leave (76.2% vs 58.5%), or any of these leave types (85.9% vs 72.0%). Smaller differences are observed for access to fexible start/fnish time (55.9% vs 53.4%), home-based work (36.2% vs 25.3%) and childcare facilities or employer-provided childcare subsidies (12.7% vs 8.7%).

families, are an important tool to encourage a full return to employment. However, the targeted nature of the Australian social security system means that subsidies are withdrawn once household income reaches a certain threshold. As a result, secondary earners who are partnered to a high-income primary earner may not be eligible for much childcare subsidy. If the secondary earner is not a high-wage worker, the family may not gain much by their labour force participation and/or increase in hours. This could discourage their return to the labour force, especially when they still have pre-school children at home.

In addition, in Australia, withdrawal of income-tested family payments occurs at low- to medium-level household income as it does in many other countries. When the withdrawal occurs within the household income range where the secondary earner has to decide whether or not to enter the labour force again or whether to increase hours of work, this can provide a disincentive to participate or to increase hours. This is due to the cumulative effect of the withdrawal of family payments, which may combine with income tax rates, with childcare costs and other costs of working (such as commuting costs) to create a very high effective marginal tax rate for the secondary earner in specifc household income ranges. The primary earner's income is likely to be taken as a given, so that the decision by the secondary earner is made conditional on the primary earner's income, and this may mean that the fnancial return to an additional hour of work is very low, while reducing the time that can be spent at home in productive activities like childcare.

To ensure secondary earners have a real choice regarding whether they want to return to employment or not after having children, childcare subsidies, family payments and other family and income support policies need to be carefully designed to avoid such disincentives as much as is possible. In addition, workplace fexibility and support can play a major role to assist employees (men and women) in achieving work-family balance. There is a large literature investigating the impacts of different policy settings, and this chapter does not aim to provide a full (or even extensive) review. Instead, we briefy discuss a few larger cross-country studies and some Australian studies which are representative of the broader literature.

Uunk et al. (2005) use micro data on households for 13 European countries combined with macro-level information on institutional structures to assess the impact of institutions on female labour supply. They focus on childcare provision. In addition they take into account two other macro-level factors: the economic need to work and gender-role values in society. They fnd that more generous provision of childcare (measured as the number of public childcare places per child under age three) and lower economic welfare (measured as gross domestic product per capita) both decrease the negative impact of childbirth on women's labour supply and thus increase female labour supply. When economic welfare increases, the institutional impact decreases. Female labour supply is also higher in countries with more egalitarian gender-role values (measured by attitudes to employed mothers), but these values do not change the impact of institutional factors, and once institutional factors are included the gender-role-value variable becomes insignifcant. The authors interpret this as the importance of institutional support for female employment, but they conjecture that the institutional support may be an intermediate factor and may be the result of the gender-role values in society.

Such fndings are consistent with the results reported in Morrissey (2017). Although high childcare costs are expected to decrease childcare usage and (mostly female) labour force participation, the estimated impact of a 10% decrease in childcare costs varies considerably across countries and studies. Morrissey conducts an extensive review of the recent literature (from 2001 onwards) and fnds estimates ranging from 0.25% to 11% increase in labour force participation, with a concentration of estimates indicating a likely increase between 0.5% and 2.5%. The estimated impacts tend to be higher for single mothers, low-income families and families with high childcare costs.

For Australia, Breunig et al. (2012) estimate the elasticity of working hours with respect to the gross childcare price for an "average" partnered mother with one child under age 13 of −0.65 (implying a 6.5% decrease in labour supply for a 10% increase in childcare costs). Employment rates are expected to decrease by 2.9% for a 10% increase in childcare costs. Again, focussing on partnered mothers, Breunig et al. (2014) estimated lower hours of work elasticities with regard to net and gross childcare prices of −0.10 and −0.14 (on average across the sample of analysis) respectively; that is a decrease of 1% and 1.4% for a 10% increase in net and gross childcare prices respectively. Similar elasticities for the employment rate were estimated at −0.06 and −0.09 respectively. Earlier Australian estimates of a 10% increase in gross childcare prices by Doiron and Kalb (2005) showed a relatively modest average of 0.5% expected decrease in hours for single parents and an average of 0.2% decrease for partnered women. However, specifc subpopulations are shown to have much higher elasticities. For instance, single parents with a preschool child are expected to have a 1.8% decrease in hours worked for a 10% increase in gross costs, while partnered mothers with a preschool child are expected to have a 0.5% decrease in hours worked. The impact of a 10% increase in gross costs is highest for single parents who have a pre-school child and earn below the median wage.

An effective childcare policy would also entail that adequate childcare is available to cover those who want to enter/re-enter the labour market. Therefore, besides the cost of childcare, the availability of childcare close to home or work is extremely important in making labour supply for families with young children feasible. Relatively little data is available on this in Australia, but Breunig et al. (2011) use information by location as reported by parents in a survey and Yamauchi (2010) complements this with information on estimated available childcare places per 100 pre-school children at the Statistical Local Area level.4 Yamauchi establishes a clear relationship between the number of childcare places in the community where a family lives, and the diffculty in fnding good-quality care and mothers' satisfaction with the amount of free time she has available. Having at least 15 places per 100 children decreases the diffculty in fnding good-quality care by a third and increases

<sup>4</sup>There were 1353 SLAs in Australia at the time of the study.

satisfaction levels by 16%. The benefts are larger for lower-educated women (who fnished high school or less). Breunig et al. (2011) use survey data on families reporting a lack of childcare, low-quality childcare or childcare that is too costly and aggregate this information at the local level (excluding the family's own response). They fnd that this aggregate information explains hours worked and employment rates, but due to the correlation of the three types of diffculty, the impacts cannot be apportioned more to one diffculty than another. Women living in areas with fewer reported childcare diffculties are more likely to be employed and work more hours. Breunig et al. use Yamauchi's estimates for additional computations to derive some policy implications: they link the number of childcare places per 100 pre-school children to the estimated decrease in the diffculty of fnding good-quality care which can then be used to predict the impact on labour supply. They compute that a local change in the number of childcare places from zero to 15–25 per 100 children could potentially increase the employment rate by 1.8 percentage points and the hours of work by 1.5 h per week on average which would be substantial increases.

Paid parental leave is another important family policy. Thévenon and Solaz (2013) examine the impact of length of leave on female and male employment rates, hours of work and earnings using data on 30 OECD countries over 40 years (1970–2010). They fnd that up to about 2 years, increased leave duration has a small positive impact on female employment rates, but beyond 2 years the impact turns negative. Similar impacts are found for hours worked, while any increase in leave widens the earnings gender gap.

To illustrate the importance of family policies in Australia, we report on the impact of the Australian universal Paid Parental Leave scheme introduced in 2011 providing 18 weeks of paid leave at the minimum wage. Broadway et al. (2020a, b) evaluate the impact on mothers' labour market outcomes in the frst year after birth using data which were collected for the purpose of this evaluation. Consistent with other literature, they fnd a positive impact on leave taking in the frst half year (thus reducing labour force participation in that period). This is complemented by a higher overall probability of returning to work in the frst year. They also provide evidence of a positive impact on continuing in the same job and under the same conditions. Perhaps most importantly, they show that the impact of the PPL scheme in terms of taking leave and returning to work was largest for disadvantaged mothers – low income, less educated, without access to employer-funded leave. This new policy had a substantial impact on this group's fnancial and mental wellbeing just after childbirth (as indicated in qualitative interviews), and it also ensured a substantial increase in the probability of returning to paid employment within a year of childbirth.

This is perhaps one of the few successful examples of social policy in Australia encouraging mothers' labour supply. Hérault and Kalb (2020) examine data for Australia from 1994 to 2016 to determine the factors infuencing female labour force participation. They fnd that the large rise in female labour force participation since the mid-1990s is nearly completely explained by (i) changes in real wages, (ii) population composition changes, and (iii) changes in labour supply preference parameters, with only a relatively small role remaining for tax and transfer policy reforms. Despite the ongoing emphasis of public policy on improved work incentives for women in Australia and elsewhere– changes in fnancial incentives due to tax and transfer policy reforms have contributed relatively little to achieve these large increases in participation. Despite a 20 and 14 percentage point increase in employment rate over the two-decade period, for single parents (85% single mothers) and partnered mothers respectively, the role of tax and transfer reforms is even smaller for these two groups than for partnered women without children.

#### *The Effect of Labour Market Circumstances on Family Formation*

In the beginning of this chapter, we outlined various pathways by which family situation, and in particular parenthood, infuence a person's labour market outcomes. However, the reverse is also true: a person's labour market situation has a strong infuence on partnering decisions and fertility decisions, including total number of births and timing of frst and subsequent births.

Standard economics would predict that, all else equal, an increase in *family income* should drive up fertility, as it improves the family's ability to cover the cost associated with having children. At the same time, an increase in *wages* increases the opportunity cost of having children – an effect especially relevant for women – which should lower total fertility. To the extent that a family's income is generated from wages, these two mechanisms work in opposite directions, and it is unclear whether they add up to a positive or negative relationship between (women's) labour market activity and the family's likelihood of having a child. The competing forces could work out differently for families: with different preferences or at different points in life (with ambiguous effects on the total number of births over a woman's lifetime), in different institutional contexts, or under different macroeconomic conditions.

Turning to the role of institutions and macro-economic conditions, Adsera (2004) frst examined a puzzling fnding in much of the developed world. The relationship between fertility and female labour force participation across countries had changed over time. We used to fnd that countries with the lowest female labour market activity, had the highest fertility rate, and vice versa. However, since the 1980s, we fnd a positive relationship. Adsera (2004) shows this is driven by labour market institutions. There is always an underlying trade-off between women's labour market activity and fertility in the short run. But this trade-off is amplifed where high female unemployment or rigid labour market institutions geared towards sole breadwinners, make re-entry after temporary joblessness diffcult and turn the short-run trade-off into a quasi-permanent life choice. This trade-off is, however, eased where low unemployment or fexible markets keep the future cost of temporary career breaks, to a minimum. Some countries – for example, much of Southern Europe – have maintained institutions that forced women into a rigid choice of either raising a family or engaging in paid work – resulting in overall low levels of engagement in

both spheres of life, as women were forced to 'pick one'. Others, such as the Nordic countries, have adopted institutions that eased the long-term trade-off faced by women, such as parental leave provisions and high public sector employment. This resulted in higher fertility and higher labour market activity at the same time, in places where women could have both.

Women's labour market activity and earnings prospects can thus be a barrier to or an enabler of fertility, depending on available institutions and context. This heterogeneity in the relationship between work and fertility, can also be observed within countries at the micro-level. For example, Kreyenfeld (2010) analyses longitudinal survey data for Germany fnding that women's unemployment and perceived job insecurity accelerates the decision to have a child in the overall population. In the context of a relatively generous welfare state and strong social norms in favour of mothers' primary role as caregivers, insecure female employment presented as an accelerator of family formation. However, Kreyenfeld (2010) found the opposite for the subset of university-educated women, for whom social norms in favour of exclusive caregiving are arguably weaker and welfare payments are a less generous replacement of potential lifetime income: they postponed frst births in response to economic uncertainty, suggesting that stable employment is not a barrier but a prerequisite for family formation for this subgroup. Pailhe and Solaz (2012) analyse a similar question in France – a country with very different norms, where strong public investment in affordable childcare encourages maternal employment. They indeed fnd that periods of economic uncertainty for French women have the opposite effect of that found for German women. French women do not accelerate childbirth in response to own unemployment, and signifcantly delay it in response to overall unemployment or own precarious employment. These results lend some weight to the theory that in countries where combining motherhood with employment is institutionally encouraged, positive labour market outcomes go hand in hand with high fertility. But in countries with norms and institutions that encourage mothers to focus on caregiving, the two outcomes are in competition.

The highly context-dependent nature of the effect of labour market outcomes on fertility, leaves the question: how does this play out in Australia? Kingsley (2018) uses data from the HILDA survey to estimate the effects of current weekly working hours and hourly wages on Australian women's likelihood of giving birth to their frst child in the following year. She fnds that Australian women delay childbirth if they have higher earnings potential. Laß (2020) extends this analysis by examining the uncertainty of labour market income, in the form of non-standard work contracts. She shows that securing a permanent contract is an important prerequisite to having a child for Australian women, with part-time hours further increasing the likelihood of a frst birth. Full-time employment is especially important for Australian men. These fndings are consistent with Australia as a liberal welfare state, where the male-breadwinner/female-caregiver model dominates social norms, and work-family-confict is seen as the individual's responsibility to solve (Pocock et al., 2013). Women are assumed to be the primary caregiver and secondary earner, who will temporarily exit the labour market after childbirth, and thus need a permanent contract to facilitate re-entry into the labour market later on; men are assumed to be the primary breadwinners whose full-time employment is crucial for family income during the period when the mother does not earn a wage, and governmentprovided parental leave is paid only at a low rate and for a short time. Nevertheless, a small impact of the government-provided paid parental leave on intended fertility is estimated by Bassford and Fisher (2020), increasing the number of children by 0.34 conditional on at least wanting one child. This impact is seen mostly for highly educated women. Consistent with this, McDonald and Moyle (2019) show that between 2006 and 2016, fertility in Australia has been falling – but only for women without post-school qualifcations. They argue this is because this group is disproportionally affected by a considerable reduction in income support payments in the course of the Welfare to Work reforms beginning in 2008. Furthermore, they are more likely to lose employment, in which case they lose access to the new paid parental leave and a large portion of their subsidised childcare. These differences across countries, within countries and over time, illustrate the large role governments and institutions can play in mitigating or exacerbating the conficts between fertility and labour market participation.

#### **Family and Work Interactions at Other Life Stages**

The previous section of this chapter has focused on the relationship between family life and work life in a particular phase of life – early adulthood, when many people make crucial decisions about future career paths and family structure, and when the tension between family and work life tends to be strong and immediate. However, we fnd cross-effects from one sphere of life to the other across the whole life course. In what follows, we discuss the interaction of family and work life over the life course, beyond the mechanisms discussed in the previous section. Each subsection introduces a different feld of research, all of which have a vast and rapidly developing literature. We do not aspire to give a complete overview of the state of research in any of these felds, but rather to briefy introduce some of the important questions studied, and to provide examples of studies in these felds.

#### *Early Life – The Impact of Parents*

In the early stages, lives are shaped by parental labour market situations and decisions. A parent's labour market situation – their employment status, working hours, wage and the stability of their job – determines how much money they can spend on their children, where the family lives, where the child goes to school, whether a child attends early education and care and of what quality, their access to health care, what food or entertainment they consume, and what educational resources are available in the home. A parent's job situation also affects how much time they can spend with their children, and how they spend that time together – their parenting style, how often they are actively engaged in education, play, or supervising multiple children while simultaneously performing other work. Moreover, parental employment and occupation can shape a child's values and aspirations. It is through this multitude of channels, that we see intergenerational transmission of labour market outcomes (such as earnings and joblessness) and other life outcomes (such as health, education, welfare dependency, or parenting style), which in turn can affect the next generation through the same channels.

This transmission begins very early in life. Currie (2009) shows in her seminal paper how parents' socioeconomic status (including their labour market situation) affects children's education and labour market outcomes later in life, through health in infanthood and even before they are born. For Australia, using data from the *Longitudinal Study of Australian Children (LSAC)*, Broadway et al. (2017) show that provision of paid parental leave – one aspect of job quality – improves the health of disadvantaged children. Paid leave alleviates the trade-off faced by mothers between securing their pre-birth job and spending time with their infant, which appears to be benefcial for their children's health.

As children grow up, the ways in which parents' job quality can affect their children's life trajectory, only grow more diverse, and many aspects of a parent's job are found to be connected to their children's development. Using LSAC data, Strazdins et al. (2010) show that poor job quality – measured by low control over how the work gets done, infexible working times, low job security and no access to paid family leave – leads to increased parental distress among parents in Australia, which in turn puts children at risk of emotional and behavioural diffculties. Li et al. (2014) review empirical studies from around the world spanning 30 years, that examine the effects of parental non-standard working hours on children. They fnd that nonstandard work hours negatively affect pre-school children's cognitive and behavioural development. Johnson et al. (2012) show that job instability among parents, as well as long working hours, negatively affects their children's behaviour and academic outcomes, but stable, high-quality employment does not.

While low-quality jobs are frequently linked to undesirable child outcomes, so are joblessness and unemployment. Mooi-Reci et al. (2019) use rich survey data linked to detailed administrative data on Dutch parents, who experienced unemployment during the economic crisis in the 1980s, and their children. They show that fathers' unemployment has a lasting negative effect on their children's education, partly refecting a loss of family income and deterioration of children's learning environments, and partly refecting that parents tended to experience a reduction in the intrinsic value placed on work.

There is a vast literature on the effect of the parent's labour market situation on children's development *and* education – which we know to partially determine children's labour market outcomes in adulthood. Other studies directly assess the adult offspring's labour market outcomes: for example, Cobb-Clark et al. (2020) show that reduced unemployment among fathers of adolescents causes a drop in unemployment among their offspring in early adulthood. They exploit the introduction of Mutual Obligation for recipients of unemployment benefts in Australia in 1999. The policy affected recipients of unemployment benefts up to age 34, and the study compares the children of unemployed fathers just below and just above the age threshold in a regression discontinuity design.

Starting from birth, parents' employment status, wages, job quality, or joblessness affect a young person's development, values, and educational choices. As a result, the choices available to different young adults are already vastly different, by the time they embark on their working lives and begin to shape their own families.

#### *Mid-Life – Past Decisions Compound*

The previous section discussed various ways in which decisions regarding family formation – especially when to have children and how many – interact with labour market decisions: for example, what occupations men and women sort into, what skills they acquire, how many hours they work, and what wages they earn. Due to the constraints that young children impose on parents' time use, it is during the early child-rearing years that gaps between parents and childless individuals, and between men and women, frst appear and then solidify. As children age, their caring needs decrease and once children reach school age and beyond, one might expect those gaps to close again.

But past decisions can cast a long shadow. Occupational choices made with fertility intentions in mind will affect earnings across the entire life course. Skills lost or not accumulated, due to reduced employment or lower working hours during early childhood years, could take a long time to recover, and women whose career progression was halted, may only fully catch up many years after they have fully re-entered the workforce, or never at all. There could be a parenthood penalty in the labour market, and especially a motherhood penalty, long after the immediate care needs have disappeared or been drastically reduced.

Some early studies directly compare the labour market outcomes of women who never had children with those of mothers, to examine how such a motherhood penalty evolves as the women age. Davies et al. (2000) construct synthetic employment and earnings biographies, using data from the British Household Panel Study, and fnd that mothers earn less than childless women well into their 40s and 50s and even beyond, especially if they have more than one child, or are low- to mediumskilled. Sigle-Rushton and Waldfogel (2007) follow a similar approach using data from a range of industrialised countries. They fnd great variation by country: while mothers in the United States lost between 11% and 19% of their earnings by age 45, mothers in Germany and the Netherlands lost between 42% and 63%.

As mentioned in the section on "Direct effects of parenthood on labour market outcomes", Kahn et al. (2014) fnd that while mothers have mostly caught up to childless women in terms of employment, hourly wages and occupational prestige, some gap remains even by age 50, and it increases with the number of children. We cannot be certain that this difference is completely due to motherhood changing a woman's employment prospects or earnings potential – instead, some women might decide to become mothers or remain childless, in part because of their labour

market outlook (see the section on "The effect of labour market circumstances on family formation"). To deal with this issue, Lundborg et al. (2017) study a group of women who all received fertility treatment, but only some of them were successful. Then they compare both groups' labour market biographies following treatment. Their results confrm what other researchers found before, which they summarise as follows *"the labor market consequences of having children are large for women. When children are young, women earn less because they work less. When children are older, women earn less because they get lower wages."* (Lundborg et al., 2017, p. 1627).

#### *Later in Life – New Caring Responsibilities Emerge*

We have caring responsibilities not only towards children, but also towards family members who need help due to illness, disability or old-age. These caring responsibilities tend to manifest somewhat later in life than those towards one's children, but when they do, they pose similar challenges for the carer's labour market activity, because this type of caregiving is time-consuming, mentally stressful and physically exhausting. Caregiving for elderly or disabled family members is predominantly supplied by women, and the gender patterns observed while a couple raises small children, are largely mirrored in this later life stage. There were more than 2.6 million carers in Australia in 2018, and around 860,000 primary carers – those who provide the most assistance to another person. Among primary carers, seven in ten were women, and the majority was 45–64 years old (Australian Bureau of Statistics, 2019).

Estimating the effect of caregiving on labour market outcomes is no trivial task. Caregiving should reduce working hours in the paid market because hours spent providing care, cannot be used for formal employment. At the same time, if several adults could potentially provide care to a family member in need, individuals who are unemployed or working part-time are more likely to become these caregivers. It is diffcult to disentangle the effect of working hours on care provision, from the effect of care provision on working hours. Longitudinal data can help with this, for example by accounting for individual fxed effects, and by analysing the timing of caregiving uptake and changes in employment status and working hours.

Bauer and Sousa-Poza (2015) provide an extensive review of the international literature on this issue. Overall, they conclude that most studies fnd informal caregiving to have a small or no causal effect on participation, but a somewhat larger effect on working hours at the intensive margin and wages – with large variation by country and institutional environment, as well as by estimation method and data set used.

This large variation in results is also found in studies specifc for Australia. Berecki-Gisolf et al. (2008) use two waves of the Australian Longitudinal Study on Women's Health, to examine to what extent care provision is determined by previous labour market outcomes. They compare employment and wages of *future*

caregivers (women aged 45–50) before care provision starts, with those of noncarers. They fnd no reason to suspect that the caregiving decision is determined by previous outcomes, but that it is followed by a strong reduction in labour force participation. In contrast, Leigh (2010) used seven waves of the HILDA survey to estimate the effect of informal caregiving on employment, working hours and wages, among men and women of working age. He fnds that individual heterogeneity explains almost all the difference in labour market outcomes between carers and non-carers, and that informal caregiving reduces labour force participation by 4–6 percentage points, while having no effect at all on wages, working hours and life satisfaction. Meanwhile, Nguyen and Connelly (2014) also use a HILDA sample of working-age men and women, but fnd that informal caregiving can reduce labour force participation by as much as 12 percentage points. Nguyen and Connelly (2014) analyse caregiving by *main* carers, while Leigh (2010) uses several broader defnitions of being a caregiver. This difference in caregiving intensity – together with different estimation approaches – likely explains the different estimated effects.

The variation in estimated effects means that the true effect of caregiving in later life on labour market outcomes is ambiguous and more research is needed. However, a clearer picture emerges, if we restrict our view to carers near retirement age. It appears that employment outcomes and in particular the retirement decision of informal carers near retirement age, is quite strongly affected by caring responsibilities, as shown by Ciccarelli and Van Soest (2018) and Heger and Korfhage (2020) for 15 European countries, Jacobs et al. (2014) for Canada, and Jacobs et al. (2017) for the US. Similar studies for Australia, focussing on mature-aged workers only, are not available.

#### **Policy Implications**

The evidence presented in this chapter shows the large impact of parenthood on labour market outcomes across all life stages. It also highlights the large differences by gender as it is still mostly women who adopt the responsibility of primary carer. This has substantial impacts on their ability to participate in the labour force in the short and medium term, and on their ability to continue in high-level, demanding occupations, and hence impact on their pay. The consequences go beyond those experienced by the individual; conficts between parenthood and labour force participation have far-reaching implications for society more broadly.

First, there are fscal implications: where parents can continue to work while caring for young children, they continue to pay taxes. Moreover, having two incomes in a family works as a safeguard against fnancial shocks, including those caused by disability or unemployment; families with two earners are less likely to have to rely on income support if one income is lost or reduced. This 'insurance effect' appears to also be keenly felt on a personal level. In the immediate aftermath of job losses caused by the COVID-19 pandemic in early 2020, many men experienced mental distress if their partner was no longer employed (Broadway et al., 2020b).

Second, low labour force participation among women puts their economic security at risk, and the consequences are felt long after the child-rearing years, as shown earlier in this chapter. This becomes particularly important in old-age, when lack of retirement savings poses a signifcant poverty risk. In Australia in 2016, just under a quarter of all women approaching retirement age (55–64 years) had no funds in the primary, government-mandated retirement savings scheme ('superannuation'), and women's median savings in this age group were \$96,000, while the equivalent fgure for men was 73% higher (Australian Bureau of Statistics, 2018). Not only does this put women at risk of poverty, especially if they do not own a home, it also leads to higher public expenditure on the means-tested Age Pension. Jefferson (2009) studies the literature on this problematic link between women's caring roles across their working lives and the funding from public pension schemes they have available to them in retirement for much of the English-speaking world (the United States, the UK, Canada, Australia and New Zealand). She also fnds that policies that facilitate and encourage women's workforce participation are crucial for poverty prevention in old age among women in all studied countries.

Third, conficts between workforce participation and parenthood are also relevant because they pose a threat to an equitable society. The severity of family-work conficts differs not only along gender lines. Women living in disadvantaged households and those who have low skill levels, and thus low earning power, are often experiencing the largest impacts – which poses the immediate risk of exacerbating inequality in society, and most importantly, intergenerational transmission of inequality and disadvantage. As discussed before, if the policy environment makes it hard for low-skilled mothers with low earnings potential to make a decent living, this can lead to poverty as well as a high-stress family environment, which in turn jeopardises children's health, education, career aspirations and future earnings prospects. A change in direction is needed if a lack of family-friendly policies deepens a cycle of entrenched disadvantage and intergenerational transmission of poverty.

And fourth, this chapter discussed the negative impact of family-work confict on fertility. Demographic ageing threatens the sustainability of retirement funding in many countries around the world, while simultaneously posing challenges for the training and recruitment of an adequately-sized health care and aged care workforce. Easing work-family conficts can help bringing fertility rates closer to replacement levels, and is thus of immediate societal interest, above and beyond its impact on the individual.

What can be done? In a nutshell: policy focus should be on (i) adequate provision of paid parental leave after birth or adoption, (ii) access to high-quality, affordable childcare for preschool children, and (iii) a tax and transfer system that rewards men and women for sharing paid work and caring responsibilities.

When deciding on the optimal length of paid parental leave, two counteracting effects need to be considered: if leave is too short, work-family confict remains high and forces many women into pursuing one at the expense of the other, resulting in overall low levels of both fertility and labour force participation. If leave is too long, women are effectively being provided with a strong incentive to stay out of the workforce by being paid for *not* returning to their job. A review of the international literature by Kalb (2018) showed the 'sweet spot' between these conficting mechanisms to be somewhere between 6 and 12 months of paid absence from work after birth, with a return-to-job guarantee.

The most important feature of a childcare system – when looked at it purely as a means to facilitate mother's workforce participation – is that it needs to effectively reduce families' out-of-pocket expenses, especially for single parent families and secondary earners with low earnings potential. This can be achieved through many different systems: from public provision of childcare to public subsidies of services provided by private markets under various levels of public regulation, or a mix of both.5

Finally, the tax and transfer system should be set up in a way so that it does not discourage secondary earners' workforce participation. The exact policy settings necessary to achieve this goal, will vary as widely from jurisdiction to jurisdiction as their tax and transfer systems do. However, one policy parameter that is important in any progressive tax system, is that taxing the individual rather than the household, is more likely to leave a secondary earner with a reasonable take-home pay that makes market work worth their while. Likewise, slow withdrawal of meanstested income support payments makes it less likely that government transfer income effectively locks recipients out of employment.

While such policies come at a price – and in many institutional contexts, that price could be hefty – this should be seen as an investment, not just an expenditure. It is important that policymakers are aware of the multitude of societal benefts that arise from easing work-family confict. These include higher fertility; lower poverty risks among women and the elderly; and a fairer, more equitable world for children to grow up in and develop their full potential.

#### **References**


<sup>5</sup>Different funding models have different implications for the quality of care and equitable access to care that are beyond the scope of this chapter. Lloyd and Penn (2012) provide an overview and discussion of various models.


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

## **Chapter 10 Marriage Matters. Or Does It?**

**Rennie Lee and Janeen Baxter**

There have been numerous changes in marriage and cohabiting patterns in Australia over the last few decades signaling shifts in the meaning and place of marriage in the life course. Marriage remains a relevant milestone for Australians with most still marrying, but how, why and when marriage takes place has changed markedly (Qu, 2020). As is the case for other western countries, the crude marriage rate in Australia (marriages per 1000 head of population) has declined steadily since the 1970s from over 9% in 1970 to 4.5% in 2019 (Baxter et al., 2015; ABS, 2019a). In 1981, over 80% of individuals aged 40–44 were married while in 2016 the comparable fgure was only 60% (Qu, 2020). Single-parent families currently comprise about 12% of family households, with single-male-parent households projected to be the fastest growing family type over coming decades (ABS, 2019b).

How people marry has also changed with only about 20% of marriages in Australia in 2020 performed by a minister of religion, down from 70% in the 1970s. This signals not only a decline in religiosity but also changing views about the role of religion in endorsing personal relationships (Qu, 2020). At the same time, unmarried cohabitation has become much more common, both as a precursor to marriage and as an alternative. In Australia, the percentage of people cohabiting was almost 20% in 2016 compared with only 5% in 1986 (Qu, 2020). While cohabitation is more common among younger individuals, it has increased across all age groups in Australia. Overall, aggregate trends suggest that cohabitation is becoming a more

© The Author(s) 2022

R. Lee (\*) · J. Baxter

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia

e-mail: rennie.lee@uq.edu.au; j.baxter@uq.edu.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_10

important and permanent means of partnering in Australia. Views on marriage have also changed with about 29% of people who participated in the Australia Talks National Survey (Zhou, 2021) reporting marriage is an outdated institution, with a higher proportion of women (33%) feeling this way compared to men (24%) and a much higher proportion of younger women aged 18–39 years feeling this way (43%) compared to older women over age 75 (13%) (Zhou, 2021).

These fgures portray important changes in the ways Australians think about and organise their personal and family lives. They also have important implications for understanding patterns of disadvantage in families. In general, marriage has long been associated with social advantage, conferred status, and economic security (Sassler & Schoen, 1999; Smock et al., 1999). Single-parent households are consistently found to be amongst the most vulnerable in Australia in terms of poverty, housing insecurity and reliance on income support (Wilkins et al. 2020). There is evidence to suggest that those with higher socioeconomic positions are more likely to marry while those with lower socioeconomic positions are more likely to cohabit (Carlson et al., 2004; Baxter et al., 2015; Perelli-Harris et al., 2019). The relationship between marriage and social advantage likely refects a combination of causal and selection effects with marriage both leading to economic advantage and those with economic advantage more likely to marry.

In this chapter we examine marriage, cohabitation and singlehood in the Australian context with a focus on how partnering is related to variations in social disadvantage and wellbeing for men and women. First, we discuss recent theories that explain the changing place of marriage in life course trajectories. We also review evidence on who gets married, who cohabits or remains single and their associated outcomes. Second, we utilize longitudinal data that enables us to examine, within individuals, how marriage and cohabitation affect outcomes relating to social disadvantage and wellbeing. Third, we discuss implications for life course theories, current understandings of partnering and family relationships and policy implications.

#### **Marriage as a Foundation or Capstone?**

Cherlin (2004, 2009) has argued that trends over the last few decades suggest that marriage is undergoing a process of deinstitutionalization where there is a weakening of the norms and behaviour patterns that characterise marriage. The emergence of new forms of family, such as cohabitation, the increase in age at frst marriage, increasing numbers remaining single, the rise of childbearing outside marriage, and increase in divorce all suggest that marriage is optional, and that those who marry and stay married are unusual. As laws have changed to enable alternatives to marriage and new forms of partnering have become more socially acceptable, marriage

may have lost its practical and social signifcance. Legislation also increasingly recognises and protects shared fnancial, family and estate contributions for both partners, and particularly women, if the relationship should end, enabling fnancially viable pathways out of unhappy relationships. In Australia, alternative partnering options include registered partnerships which provide cohabiting couples with many of the same legal arrangements as traditional marriage and same-sex marriage which became legal in 2017.

Marriage was traditionally a springboard to adulthood, a life course event that indicated the beginning of adulthood entailing leaving the family of origin home, setting up a new household, economic independence, employment and start of parenthood (Marini, 1978). However, as discussed in Chap. 8 in relation to emerging adulthood, young people are increasingly delaying marriage, perhaps due to the diffculty in attaining social and fnancial independence. Increasingly, marriage is associated with a certain level of fnancial stability and those with fewer fnancial resources and less stable employment are more likely to delay marriage (Edin, 2000; Aldo, 2014). As a result, marriage has shifted from a springboard into adulthood to a capstone that must now wait until several milestones have already been achieved (Cherlin, 2004; Edin & Kefalas, 2005; Holland, 2013). If true, this has implications for both the meaning and returns to marriage. In particular, those who are the most disadvantaged may not be able to reach this capstone making it a desirable but unachievable life course event.

One implication of the deinstitutionalization of marriage and its shift to a capstone event is that some of the social and psychological gains of belonging to a well-recognized and established institution are also diminishing. As cohabitation becomes more widespread and institutionalised, it may, in turn, produce similar social and psychological benefts as marriage. Research has identifed several advantages associated with marriage. Married individuals have reported greater subjective wellbeing, mental and physical health (Waite & Gallagher, 2000), and economic resources (Sassler & Lichter, 2020). Given that cohabitation is increasingly similar in form and function to marriage, it is possible that these benefts may now extend to those who cohabit (Musick & Bumpass, 2012). There are several similarities between cohabitation and marriage. Cohabitating unions mark the start of coresidential partnerships and also increasingly represent settings for having and raising children (Perelli-Harris & Sanchez Gassen, 2012). As cohabitation becomes more widely accepted, it is possible that we are witnessing a convergence of the benefts of marriage and cohabitation (Sassler & Lichter, 2020).

Nonetheless, there are some important differences between marriage and cohabitation that may lead marriage to retain distinct benefts. Cohabitation may represent a less stable family type, even among those who have children (Andersson et al., 2017; Musick & Michelmore, 2018). This may negatively affect wellbeing with cohabitors more likely to have lower life satisfaction (Soons & Kalmijn, 2009) lower relationship quality (Wiik et al., 2012), higher levels of depression (Brown, 2000; Lamb et al., 2003), and worse health (Musick & Bumpass, 2012). In general, it may be that cohabitation is associated with a pattern of disadvantage that continues across the life course, including less stable relationships, less fnancial security, higher unemployment, and lower wellbeing.

#### **Who Marries and Who Cohabits?**

One reason why the benefts to marriage may differ is because cohabiting couples and married couples differ in their composition. Though this has changed over time, cohabitation has been previously more common amongst those with lower economic resources or educational attainment (Heard, 2011; Evans, 2015). In Australia, individuals with higher education levels are more likely to marry than those with lower education (Hewitt & Baxter, 2011; Evans, 2015; Heard, 2011).

Much of our understanding of the link between marital patterns and social disadvantage is based on the United States which varies from Australia in several key ways that are related to marital patterns and social disadvantage. Although the United States and Australia are often considered liberal welfare states with a limited safety net and reliance on means-tested benefts, Australia offers more universal benefts overall (Arts & Gelissen, 2002; Esping-Andersen, 1990). In turn, women in Australia may be less dependent on marriage for fnancial security. Furthermore, married and cohabiting couples in Australia share many of the same rights and this is much less true in the United States (Perrelli-Harris et al., 2018). For instance, in Australia, cohabitors share similar access to family courts in the event of union dissolution and similar rights to inheritance (Evans, 2015). Overall, Australia provides greater state support for cohabiting couples than the United States, which may refect the general acceptance of cohabitation and in turn, lower incentives for marriage. Moreover, there is a strong link between cohabitation and disadvantage observed in the United States in a way that is not the case in Australia (Perelli-Harris & Lyons-Amos, 2016). Therefore, it is possible that the strong links between marital patterns and social disadvantage are more characteristic of US society than Australian society. To examine these issues and to provide an overview of the associations between different forms of partnering and social disadvantage, we address the following questions in this chapter:


#### **Results**

#### *Who Marries and Who Cohabits in Australia? Evidence from HILDA*

To answer these questions, we use data from The *Household, Income, and Labour Dynamics in Australia* survey, waves 1–18. To address our frst two research questions, we examine three groups of individuals: individuals who are single1 (never married and not living with a partner but could be partnered) throughout the 18 waves of the survey (n = 38,330); individuals who move from single to cohabiting (never married and living with partner) (n = 19,966); and those who transition from single to married (n = 189,39).2 We restrict our sample to men and women who were single in their frst wave of the survey so that we can observe the effects of the transitions to partnership on various outcome measures. We use multinomial logistic regression to predict whether individuals transition from single to cohabitation, single to married, or stay single (reference category) and to examine the characteristics of those who make these different transitions.

This approach allows us to extend the existing literature by examining the outcomes of those who stay single along with those who transition to cohabitation and marriage. While most previous research has compared outcomes for the latter two groups of partnered individuals, we widen our lens to include single individuals as a way of understanding the broader effects of marital status and coresident partnerships. By including singles as a reference category, we can compare outcomes for those who cohabit and marry as well as make comparisons between coresident partnerships and other individuals. Additionally, given our interest in the relationship between marital status and social disadvantage, including single individuals expands the depth of our knowledge.

We include several independent variables in our model: gender, racial/ethnic background (Indigenous Australian, English Speaking Background Immigrant, Non-English Speaking Background Immigrant, and non-Indigenous Australian educational attainment (high school degree, diploma or certifcate, Bachelor's degree, graduate/postgraduate degree, and less than high school), individual income, age, labour force status (unemployed, not in the labour force, and employed), remoteness (inner regional Australia, outer regional Australia, and major city), presence of children (age 14 or less) in the home, and survey wave.

Figures 10.1 and 10.2 present the predicted probabilities of being in each marital status by racial and ethnic background for women and men, respectively.

<sup>1</sup> It is not possible in HILDA to distinguish between individuals who are unpartnered from individuals who are partnered and not living together. It is likely that some of the respondents are in relationships as 22% report having children under 14.

<sup>2</sup> Individuals who transitioned from single to married do not include those who were previously cohabiting. In our analyses, we drop individuals who moved from single to cohabiting to marriage to isolate the effects of moving from single to cohabiting only and single to married only.

**Fig. 10.1** Marital status by race/ethnicity for women.

Note: Predicted probabilities of marital status are based on regression analysis controlling for gender, racial/ethnic background, educational attainment, personal income, age, labour force status, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

**Fig. 10.2** Marital status by race/ethnicity for men.

Note: Predicted probabilities of marital status are based on regression analysis controlling for gender, racial/ethnic background, educational attainment, personal income, age, labour force status, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

The predicted probabilities are based on the regression model with independent variables held constant at their means. Taken together, both Figs. 10.1 and 10.2 show that across all racial/ethnic groups, about half stay single over their duration in the survey. Among those who partner, women and men who are Indigenous Australians and immigrants from English-Speaking Backgrounds show a greater probability of cohabitating than marrying. In contrast, immigrants from non-English speaking backgrounds show a greater probability of marriage. Thus, we see that despite being foreign-born, immigrant background infuences partnership pathways.

Additionally, when we compare how the relationship between marital status and racial and ethnic background differs by gender, we fnd that relative to their female counterparts in each racial/ethnic group, men show higher probabilities of staying single, but lower probabilities of cohabiting across each group. Men and women show similar probabilities of marriage by racial/ethnic group. For instance, among non-Indigenous native-born Australians, we fnd that relative to men, women are less likely to be single (0.47 versus 0.52), more likely to cohabit (0.28 versus 0.24), and less likely to marry (0.25 versus 0.24). This suggests that among men, staying single is the most common but for those who do partner, they are slightly more likely to marry. In contrast, women are more likely to transition into cohabitation.

Both men and women with Indigenous backgrounds show the highest probabilities of being single or cohabiting and immigrants from non-English speaking backgrounds show the highest probabilities of marriage. Overall, Figs. 10.1 and 10.2 show large racial and ethnic differences in the probability of staying single, transitioning to cohabitation, and transitioning to marriage. We fnd that the racial and ethnic variation across marital status is consistent for men and women despite men showing higher probabilities of staying single and smaller probabilities of cohabitation.

We also examine how marital status differs by education level and gender. Figure 10.3 presents the predicted probabilities of marital status for those with a high school degree and bachelor's degree or higher by gender. For both men and women, we fnd that those with a high school degree show the highest probabilities of staying single though this is slightly higher for men (0.47 versus 0.51). Moreover, the likelihood of cohabitation is also higher among women and men with a high school degree (0.28 versus 0.24). Men and women with at least a Bachelor's degree show the highest probabilities of marriage (0.36 versus 0.37). Overall, Fig. 10.3 shows that education is positively associated with marital status where those with higher levels of education are more likely to marry whereas those with lower education levels are more likely to stay single or cohabit.

Figure 10.4 also presents the predicted probabilities of marital status by labour force status and gender. In general, we see some variation in employment across marital status particularly when we focus on the role of unemployment and not in labour force in predicting marital status. Figure 10.4 shows that single men and

**Fig. 10.3** Marital status by education level and gender.

Note: Predicted probabilities of marital status are based on regression analysis controlling for gender, racial/ethnic background, personal income, age, labour force status, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

**Fig. 10.4** Marital status by labour force status and gender.

Note: Predicted probabilities of marital status are based on regression analyses controlling for gender, racial/ethnic background, educational attainment, personal income, age, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

women have the highest probabilities of unemployment (53.3 and 58.3) and not being in the labour force (54.8 and 59.6) respectively. In contrast, women and men who transition to marriage (15.4 and 14.8) or transition to cohabitation (31.3 and 27) have lower probabilities of unemployment respectively. Thus, this indicates that those who are unemployed or not in the labour force are more likely to be single or cohabitating relative to married individuals.

Taken together, the fndings thus far illustrate the role of racial and ethnic background, educational attainment, and labour force status on the probability of staying single, transitioning to cohabitation, or transitioning to marriage. Overall, those who marry are the least disadvantaged while those who remain single appear to be the most disadvantaged with cohabitors somewhere in between. Understanding why Indigenous Australians and English Speaking Background immigrants have higher probabilities of cohabitation whereas Non-English Speaking Background immigrants show higher probabilities of marriage has important implications for social disadvantage. These differences may be linked to cultural variations in the importance of marriage and family or may stem from compositional differences in the socio-economic characteristics of Indigenous and NESB groups.

#### *Is Partnership Status Associated with Variations in Social Disadvantage and Wellbeing?*

A second aim of this chapter is to examine the relationship between marital status and social disadvantage by observing how single, cohabiting, and married individuals differ along several health, wellbeing, and socioeconomic indicators. Our analyses will show descriptively how individuals across the three marital statuses compare in their health and wellbeing and socioeconomic status. We exploit the longitudinal feature of HILDA data to examine, within individuals, how marriage and cohabitation affects outcomes relating to wellbeing and disadvantage.

#### **Subjective Wellbeing**

Several studies have found that married individuals have higher levels of subjective wellbeing than non-married individuals (Mikucka, 2016; Stutzer & Frey, 2004; Waite & Gallagher, 2000), including cohabiting couples (Waite & Gallagher, 2000). Perelli-Harris et al. (2019) and Wilkins et al. (2020) fnd that married men and women in Australia have higher subjective wellbeing scores than those who are cohabiting, separated, divorced, widowed, or never married. Similarly, married individuals report greater satisfaction with all aspects of employment, fnances, housing, safety, and leisure (Wilkins et al., 2020).

We add to this body of work by showing descriptively the long-term mental health outcomes of those who stay single, those who transition from single to cohabiting, and those who transition from single to married in Australia. Figures 10.5

**Fig. 10.5** Predicted mental health scores for women by marital status over survey wave. Note: Predicted probabilities of mental health scores are based on random-effect linear regression analysis controlling for gender, racial/ethnic background, educational attainment, personal income, age, labour force status, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

and 10.6 show the predicted mental health scores by marital status over time for women and men, respectively, with independent variables held constant at their means. The predicted probabilities are based on random-effect linear regression models predicting mental health scores controlling for gender, racial/ethnic background, educational attainment, individual income, age, labour force status, presence of young children, and survey wave. We measure mental health using the Mental Component Summary (MCS) score, which is drawn from a patient-reported Short Form 36 (SF-36) (Butterworth & Crosier, 2004). It is created from several subscales measuring mental health, emotional problems, and social functioning. The scale ranges between 0 and 100 with a higher score representing better mental health. In Figs. 10.4 and 10.5, higher predicted scores on the vertical axis indicate better self-assessed mental health.

Figures 10.5 and 10.6 show that across all marital status groups, mental health scores decline over time for both men and women. Nonetheless, we fnd that those who transitioned from single to married consistently show the highest mental health scores, followed by those who transitioned from single to cohabiting and those who remain single. We fnd that these patterns hold for both men and women, though men in each marital status show better mental health scores than their

**Fig. 10.6** Predicted mental health scores for men by marital status over survey wave. Note: Predicted probabilities of mental health scores are based on random-effect linear regression analysis controlling for gender, racial/ethnic background, educational attainment, personal income, age, labour force status, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

female counterparts. This indicates a selection effect with men and women with better mental health more likely to select into marriage than either cohabitation or remaining single.

Additionally, Figs. 10.5 and 10.6 show that men have better mental health scores than women. Specifcally, married men report the highest mental health scores as indicated on the top line in Fig. 10.6 and single women report the lowest scores as indicated by the bottom line in Fig. 10.5. In fact, the mental health of married women is roughly equivalent to that of single men. This supports arguments showing that men experience a "marriage premium" though this has typically focused on the effects of marriage on men's labour market outcomes (Cohen, 2009). To sum up, Figs. 10.5 and 10.6 show that net of controls, individuals who transition to marriage consistently show the greatest predicted mental health scores, especially men, whereas those who remain unpartnered show the lowest predicted mental health scores, with unpartnered women being particularly vulnerable to poor mental health. Likewise, we fnd that people transitioning to marriage also start with greater mental health as indicated by their higher mental health scores, which suggests a likely selection effect in which those with greater health are more likely to select into marriage.

#### **Economic Wellbeing**

Another possible beneft of marriage is greater fnancial security and economic wellbeing (Hardy & Lucas, 2010; Sweeney, 2002). Married individuals are more likely to pool and jointly manage their resources than cohabiting partners (Brines & Joyner, 1999; Treas & De Ruijter, 2008). Cohabitation is not associated with the same norms of sharing fnances suggesting divergent patterns in how married and cohabiting individuals organise their relationships (Hardy & Lucas, 2010). In Australia, married individuals have higher household incomes than cohabiting individuals. Perrelli-Harris report that around 26% of married men were in the highest household income quintile compared with only 11% of cohabiting men in 2013 (Perrelli-Harris et al., 2019). Additionally, in the same year, nearly 18% of cohabiting men were in the lowest income quintile compared to 8% of married men (Perrelli-Harris et al., 2019).

We measure economic wellbeing using two indicators: access to emergency funds and employment. We measure access to emergency funds as a dichotomous variable with no access to emergency funds as the reference category. These are derived from several questions in the HILDA survey asking respondents how diffcult it would be to raise \$2000 for an emergency ("easily"; "with some sacrifces"; "would have to do something drastic"; or "couldn't raise emergency funds"). We measure employment as a dichotomous variable capturing employed versus unemployed/not in the labour force. This is derived from annual survey questions about current labour force status (employed, unemployed, or not in the labour force).

Figures 10.7 and 10.8 show the predicted probability of access to emergency funds for women and men respectively. A higher value on the vertical axis indicates a higher probability of access to emergency funds. The predicted probabilities are based on random-effect logistic regression models predicting access to emergency funds or not, controlling for gender, racial/ethnic background, educational attainment, individual income, age, labour force status, presence of young children, and survey wave. Overall, Figs. 10.7 and 10.8 show that men and women of all partnership statuses develop greater fnancial security over time. Likewise, both men and women who transitioned from single to marriage show the highest probability of fnancial wellbeing, followed by individuals who transitioned from single to cohabiting and single individuals. However, men show larger disparities in access to fnancial wellbeing by marital status than their female counterparts, suggesting a larger association between whether men transition to cohabitation or marriage and their access to emergency funds over time. This gender disparity may be shaped by our subjective measure of fnancial wellbeing, rather than an objective measure. Related, it is possible that men may feel a greater need to reach a certain level of fnancial wellbeing before transitioning to partnership.

We also consider the relationship between marital status and employment for men and women (fgures not shown). Not surprisingly, we fnd that men's employment probability is higher than their female counterparts. Even among those who stay single, men consistently show higher employment probabilities than their female counterparts. For both men and women, the probability of employment

**Fig. 10.7** Predicted probability of access to emergency funds for women by marital status over survey wave.

Note: Predicted probabilities of access to emergency funds are based on random-effect linear regression analysis controlling for gender, educational attainment, personal income, age, labour force status, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

slightly increases over time but it remains relatively fat. Again, we see that employment probabilities differ with those who transitioned to marriage showing the highest probabilities, followed by those who transitioned to cohabitation, and single individuals showing the lowest probabilities. We see the same pattern for women and men though again, the gaps in employment by marital status are larger for men.

In sum, our analyses show that, those who transition from single to married show the strongest mental health and fnancial wellbeing, net of controls. Additionally, we fnd that those who do not partner show the worst outcomes across these indicators. We fnd that these differences persist over time and are consistent for men and women, though men have higher wellbeing, employment, and fnancial security than women.

#### *Is Marriage Selective or Protective?*

Central to the discussion about the positive association between marriage and social advantage is whether the effect is causal or selective (Osborne et al., 2007; Perelli-Harris et al., 2019). In other words, is the greater subjective and economic wellbeing

**Fig. 10.8** Predicted probability of access to emergency funds for men by marital status over survey wave.

Note: Predicted probabilities of access to emergency funds are based on random-effect linear regression analysis controlling for gender, educational attainment, personal income, age, labour force status, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

observed among married individuals an effect of marriage or is it explained by the advantages that select individuals into marriage (Stutzer & Frey, 2004)? Relatedly, are the poorer outcomes observed among cohabiting individuals an effect of cohabitation or a result of cohabitation being selective of disadvantage (Kennedy & Bumpass, 2008; McLanahan & Percheski, 2008; Perelli Harris et al., 2018)?

While it is beyond the scope of this chapter to establish causality, we contribute by observing whether there are differences in subjective wellbeing and self-rated health, fnancial stability, home ownership, and employment by marital status. We include two additional measures—self-rated health and homeownership—to broaden our understanding of the relationship between marital status and social disadvantage. We include self-rated health which is known to be a valid measure of physical health (Noymer & Lee, 2012). In addition, we include homeownership as it is viewed as an indicator of material wellbeing and is an important source of wealth accumulation in Australia, but also central for promoting equality of opportunity (Kuebler & Rugh, 2013; Lewin-Epstein & Semyonov, 2000). Home ownership is measured as a dichotomous variable capturing whether individuals own their home or not (reference category). This is derived from an annual question in HILDA

where respondents are asked about their housing status (own their home, rent it, or live rent free).

Using the longitudinal feature of the HILDA data, we examine how these various wellbeing, health, and socioeconomic outcomes vary among individuals who are single, cohabiting, and married. We used fxed effects models to account for withinindividual transitions in marital status and the subsequent effects on each of the outcome variables. Fixed effects models allow us to address possible selection bias in which more advantaged individuals are more likely to select into marriage. Fixed effects models eliminate between-person variation that may confound the relationship between marital status and wellbeing and socioeconomic outcomes (Allison, 2009; Wooldridge, 2010). In turn, these models account for within-person changes in outcomes so respondents are compared to their own averages over the event of a move into cohabitation or marriage and thus, controls for some observed and unobserved factors.

In Figs. 10.9 and 10.10, we present the predicted health and socioeconomic outcomes across three marital statuses, for women and men respectively. We focus frst on self-rated health, which ranges from 0 to 100 with a higher score indicating better health. This is derived from a question asking survey respondents to rank their health as "poor", "fair", "good", "very good", and "excellent". A higher value on the vertical axis indicates greater self-rated health.

Note: Predicted probabilities based on fxed effect regression analysis controlling for marital status, educational attainment, personal income, age, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

**Fig. 10.10** Predicted wellbeing and socioeconomic outcomes by marital status for men. Notes: Predicted probabilities based on fxed effect regression analysis controlling for marital status, educational attainment, personal income, age, remoteness, presence of young children in the home, and survey wave. (Source: Household, Income and Labour Dynamics in Australia survey, Waves 1–18)

Figure 10.9 shows virtually no differences across the groups of women in relation to health or employment. On the other hand, when we examine homeownership, we fnd that about 74% of married women own a home compared with only 66.4% of single women and 57.4% of cohabiting women. Figure 10.9 also shows married women have much greater access to emergency funds (60.5%) than cohabiting (53.6%) and single (47.1%) women. We also show the wellbeing and socioeconomic outcomes by marital status among men in Fig. 10.10. Here we see similar patterns. Amongst men, there are no substantive differences in self-rated health or mental health across the groups. But like women, married men are much more likely than cohabiting and single men to own their own home and to have access to emergency funds.

Interestingly, we fnd that cohabitors have a lower probability of home ownership than their married or single counterparts. This suggests that home ownership may represent an important milestone or step for cohabiting individuals that infuences whether they transition to marriage. Finally, the results also show a clear gap between married and non-married individuals in access to emergency funds. Single (56.6%) and cohabiting (60.7%) individuals show much lower fnancial security in this regard relative to married individuals (64%). This indicates that fnancial security may be an important milestone that individuals or couples achieve before transitioning to marriage.

Taken together, Figs. 10.9 and 10.10 illustrate disparities across marital status depend on the specifc outcome. We fnd that married men and women have greater fnancial stability relative to cohabiting and single individuals, a result that is consistent with previous studies (Lee & Ono, 2012). We fnd no differences in relation to health. We also fnd some gender differences in how marital status shapes employment. In particular, we fnd that partnership increases the likelihood of employment among men whereas we fnd the opposite for women, although the differences for women are small. Specifcally, partnered women show slightly lower employment than single women whereas partnered men show higher employment than single men. For women these patterns are consistent with persisting gender divisions of labour in households where women do more unpaid care and domestic work than men. Australia still largely conforms to traditional gender roles with men serving as breadwinners and women taking up greater unpaid care and domestic work (Baxter et al., 2008; Baxter & Tai, 2016). Overall, our results show that even after controlling for selection, we fnd that married individuals show greater socioeconomic outcomes indicating that selection alone is unlikely to explain these outcomes observed among married individuals.

#### **Discussion**

Partnership transitions are important life course events that infuence wellbeing and disadvantage outcomes. On average, individuals who do not transition to marriage fare worse than those who do on a number of the indicators considered here. Overall, we fnd that across several measures of socioeconomic wellbeing, married individuals show better outcomes than their cohabiting counterparts and single individuals. Married individuals are more likely to be employed, own a home, and have access to emergency funds, net of various socioeconomic and demographic controls. Additionally, these advantages remain even when we consider their outcomes after they have transitioned to marriage controlling for unobserved and observed bias. While it is beyond the scope of this study to determine whether such differences are due to selection or the returns to partnership, it is clear that cohabiting individuals are more socioeconomically disadvantaged than their married counterparts and single individuals are the most disadvantaged. In sum, our fndings suggest that the socioeconomic advantages incurred by married individuals is unlikely to refect selection alone. We fnd no substantive differences in health and wellbeing across individuals of different marital statuses and no evidence that marriage is associated with greater subjective wellbeing or health. An avenue for future research may be to consider why marriage provides greater socioeconomic benefts relative to cohabiting and single individuals but not the same benefts for subjective wellbeing and health.

Across gender, nativity, and Indigenous background, we fnd a strong association between marriage and social advantage and similarly, strong associations between being single and social disadvantage. Across all groups, we fnd that married individuals continue to show greater socioeconomic outcomes, health, and wellbeing. However, we fnd that the gap between married and cohabiting individuals varies by gender, nativity, and Indigenous backgrounds. This lends support to the idea that selection into marriage, cohabitation, or staying single varies by subgroup. Similarly, this may refect variation in the meanings attached to marriage and cohabitation and differences in whether groups view cohabitation as a stepping stone or a fnal partnership destination. Future work may consider how group membership alters our understanding of marriage, cohabitation, and being single. It is possible that the meaning of marriage or the acceptance of cohabitation may vary by subgroup in complex ways. Additionally, it is clear that the effects of cohabitation may lead to greater benefts for some outcomes but not others. Therefore, greater research in understanding where cohabitation acts as a proxy for marriage and where it provides similar outcomes to those who are unpartnered can extend our knowledge on the unique benefts of marriage.

#### **Conclusion**

Marriage patterns have changed over time and there is considerable evidence that Cherlin (2004) is correct in arguing that marriage is now more effectively viewed as a capstone life course achievement rather than a springboard into adulthood. Policies aimed at supporting individuals to achieve fulflling lives must recognise increased diversity in partnership arrangements and provide strong supports to those who choose not to pursue traditional marital arrangements. In this chapter, we show that individuals who do not partner have the greatest disadvantage and this is exacerbated for single women. Whether single individuals have disadvantages that preclude their opportunities for partnership or whether their disadvantages accrue by not partnering, our fndings show that their longstanding disadvantages in the labour market and limited fnancial stability suggest greater needs for social and economic support. This suggests a clear need for policy supports that focus on single individuals, especially single women and parents.

We also fnd those who cohabit show greater disadvantages across several outcomes relative to their married counterparts. In some cases, cohabiting individuals show only slightly better outcomes than single individuals. This highlights the need for policies that focus on unmarried individuals. Tailored policy approaches are required as policies for all unmarried individuals do not focus specifcally on the challenges facing single individuals. Likewise, among single individuals, women are particularly disadvantaged and need additional resources to address their specifc challenges.

#### **References**


Woolridge, J. M. (2010). *Econometric analysis of cross section and panel data*. MIT Press.

Zhou, C. (2021, May 26). *Australia Talks National Survey reveals what Australians think about marriage and children*. Australian Broadcasting Corporation. Retrieved from https://www.abc. net.au/news/2021-05-26/australia-talks-national-survey-children-marriage/10014639

**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 11 Parenthood: Disrupting the Intergenerational Transmission of Social Disadvantage**

#### **Kylie Burke and Cassandra K. Dittman**

Parenthood represents one of the most signifcant transitions in an individual's life and once commenced, crosses all stages of the life course. Just as a child's needs and activities change as they develop and grow, so too do the activities, demands and pressures of parenting. Parents' capacity to create and maintain nurturing, responsive, and stimulating home environments across their child's development and across their own parenting experience is strongly infuenced by the social and economic resources they have available to them. To better understand the mechanisms by which parenting infuences child and adolescent outcomes, it is critical to investigate the impact that social and economic inequality has on the day-to-day decisions, choices, and tasks of raising children.

It is well understood that parenting is critical for supporting the development and wellbeing of children. Parenting characterised by responsivity, warmth, acceptance, encouragement, clear boundaries, effective confict management, problem-solving and supervision (Komro et al., 2011) is associated with positive outcomes for children and adolescents, including greater academic engagement and achievement (Kelly et al., 2012), and better social and community connectedness (Ben-Zur, 2003). Furthermore, effective parenting also protects against negative outcomes, including lowering the risk of emotional, social and behavioural problems (e.g.,

C. K. Dittman School of Psychology, The University of Queensland, Brisbane, QLD, Australia

Central Queensland University, Rockhampton, Australia e-mail: c.dittman@cqu.edu.au

K. Burke (\*)

Metro North Health Service, Mental Health, Queensland Health, Herston, Australia

School of Psychology, The University of Queensland, Brisbane, QLD, Australia

Australian Research Council Centre of Excellence for Children and Families over the Life Course, The University of Queensland, Brisbane, QLD, Australia e-mail: Kylie.Burke3@health.qld.gov.au

aggression, truancy, antisocial behaviour; Haskett et al., 2008) and preventing early sexual experience, alcohol and other drug use (Kelly et al., 2011).

Many parents are raising children in very diffcult circumstances, affected by factors such as extreme poverty, homelessness, intergenerational violence, mental illness and substance addictions (Davidson et al., 2020). While these factors clearly impact on parents' resources (both internal and external) and are risk factors for poor child outcomes, they do not in themselves prevent parents from being effective in supporting the development and wellbeing of their children (Azar & Cote, 2002). Rather, it is the complex interplay between the contextual and individual factors that shape parenting experiences and behaviours and whether and what aspects of parenting and disadvantage may be passed from one generation to the next. There are many factors that infuence both an individual's experience of parenting (e.g., satisfaction, wellbeing, quality of life) and the way in which they raise their child (i.e., parenting practices). These include the individual characteristics and history of the parent (e.g., mental health; cognitive ability; own experiences of being parented; adversity; age; gender; education), the context of the family unit (e.g., family structure; number, ages and developmental stage of children; social emotional, behavioural and physical wellbeing of children); community and neighbourhood factors (e.g., social cohesion; collective effcacy; relative neighbourhood advantage/disadvantage; neighbourhood disorganisation) and government policies that are designed to support the individual, family and society.

In this chapter we describe the individual, neighbourhood, and community factors that infuence the experience of parenthood and subsequent outcomes for children. We argue that this infuence is complex and multi-directional, acknowledging the reciprocal associations among children, parents, family systems and the broader social and economic ecology in optimising child developmental outcomes. We also discuss the importance of focusing on parenting as a mechanism by which the intergenerational transmission of disadvantage may, at least in part, be interrupted. The chapter concludes with a discussion of how interventions and policy can better support parents and families, and what further research is needed to enhance our understanding of the determinants of parenting.

#### **Understanding Parenting**

Parenting is a signifcant aspect of life in which parents experience multiple transitions in line with the growth and changes in the lives of their children. Thus, while a unifying theory of parenthood does not exist, with parenthood often described in light of the child's developmental stages and milestones (e.g., infancy, toddlerhood, preschooler and so on), theoretical attention has been given to considering the tasks and activities that the parent must undertake and master. Two theories have predominated in the literature: Parenting Styles (Baumrind, 1971; Maccoby & Martin, 1983) and Social Learning Theory (Bandura, 1961). Parenting styles comes from a social-emotional perspective, with parenting styles defned as the larger context or emotional climate in which parenting behaviours are expressed (Maccoby & Martin, 1983). Three specifc styles of parenting were initially defned based on levels of demandingness-control and acceptance-responsiveness in parent-child interactions: authoritarian, authoritative and permissive (later divided into indulgent and neglectful; Maccoby & Martin, 1983). Research in this area has suggested that authoritative parenting, involving high levels of behavioural control in the context of high responsivity and acceptance, is the optimal approach, with positive child outcomes such as educational commitment and achievement (Purdie et al., 2004) and lower levels of antisocial behaviours such as alcohol and other drug use (Adalbjarnardottir & Hafsteinsson, 2001). The parenting styles model has been criticised as lacking suffcient evidence within culturally diverse or low socio-economic circumstances (Kim & Rohner, 2002). Further, the model is descriptive rather than explanatory, and does not account for the infuence of family context (e.g., being a single parent in a high crime neighbourhood versus being a single mother living in a middle-class suburb) on the appropriateness or effectiveness of specifc parenting practices for child wellbeing and development.

Social Learning Theory offers an alternative approach to understanding effective parenting. Social Learning Theory is functionally based, focusing on the unique contingencies that impact the individual parent, child and family unit. It also takes account of parental and child learning histories, skills, personal goals, and the context in which parenting occurs. Thus, parenting styles can be considered to set the tone within which parenting practices are delivered, and so are differentiated from parenting practices, which are defned as the context-specifc, goal-directed behaviours that parents use to achieve a particular end. Social Learning Theory approaches to parenting interventions have the strongest evidence base for producing positive outcomes for children and adolescents (van Aar et al., 2017).

#### *Parenting in Context*

Parenting is contextually driven. There is no single right way to parent, nor is there a single set of actions and responses that will result in "a well-developed child". Rather, the optimal approach to parenting varies according to characteristics of the parent, their child (e.g., temperament, age, cognitive ability), family circumstances (structure, confict), access to resources (fnances, employment, support) and community and policy factors. Thus, different approaches are needed to effectively parent in different contexts and with different children. Azar and Cote (2002), in their seminal work on child maltreatment, described effective parents as having a broad repertoire of parenting strategies and the capacity to fexibly apply these strategies according to the specifc demands of the varied developmental and parenting situations they face with each of their children. The Australian Federal Government's "Parenting Information Project" (Centre for Community Child Health, 2004) defned effective parenting practices as "actions that best achieve the goals of parenting a particular child in a particular context" (Centre for Community Child Health, 2004, p. 56).

It is important, then, to understand that the social context for parenthood is changing across the world. Families are now more likely to have fewer children and parents are older at the time of birth of their frst child than three decades ago. More women are participating in the workforce, often with both parents engaged in employment, and the educational attainment of women has increased signifcantly (OECD, 2016). There is now much variation in family structure, with an increase in non-traditional relationships between parents (e.g., cohabitation rather than marriage, same-gender parents) and the number of sole-parent households. Greater social mobility and globalisation means that many families may live in communities isolated from their extended families (Weldon-Johns, 2013) and the advice and practical support that they can offer.

Thus, the identifcation of parenting styles and practices is not enough to fully understand what makes an "effective" parent. Achieving the "right" mix in terms of when and how parents negotiate, set boundaries or get involved in their child's life needs to be contextually driven, because behaviours that are effective in one environment may be potentially ineffective in another. The next section of this chapter will explore how parenting is shaped and changed in the context of disadvantage.

#### **How Parenting Differs Across Levels of Social Disadvantage**

Theory and research on social disadvantage and parenting are inextricably linked with the impact that parenting quality, parent-child relationships, and parental wellbeing have on child developmental outcomes. These models of social disadvantage and parenting tend to view parenting and parental wellbeing as intervening or mediating variables between social disadvantage and child outcomes; that is, that social disadvantage indirectly infuences child development via its effects on parents (Grant et al., 2003). Even in more recent conceptualisations of social disadvantage and parenting that have broadened from developmental psychology to disciplines including economics (Heckman, 2006) and cognitive neuroscience (Johnson et al., 2016; Noble et al., 2015), the focus has been on the role that parenting and the home environment plays in the long-term attainment, achievement, mental health and productivity of individuals across childhood, adolescence and adulthood. While this is clearly critical, it is also important to view parenting and parental wellbeing as outcomes, by evaluating the impact of social disadvantage on the day-to-day experience of being a parent. Thus, this section of the chapter will examine differences in parenting and parental stress and wellbeing based on levels of social advantage and disadvantage.

The construct of parenting encompasses the behaviours and responses parents engage in during interactions with their children, the home environments they create for their children, and the opportunities they provide for stimulation and community connection outside of the home (Hoff & Laursen, 2019). Each of these domains of

parenting has been investigated in terms of how they are affected by differing aspects of social disadvantage, but there has been a particular focus on interactional behaviours and home environments. Much of this research uses individual and family indicators of disadvantage (e.g., education, income, fnancial stress), but neighbourhood indicators of social disadvantage are noted where research is available.

#### *Parental Interactions with Children*

#### **Time Spent with Children**

One of the primary ways that patterns of interaction between parents and children has been examined is through assessment of the quantity and quality of time parents spend with their children. Using time use diary methodology, international research suggests that better educated mothers tend to spend more time with their children overall (Sayer et al., 2004) and that more economically and educationally advantaged parents spend more time in stimulating and enriching activities both in the home (e.g., reading daily to children, actively teaching literacy and numeracy, telling children stories) and outside the home (e.g., visiting the library, zoo, or museum; Kalil et al., 2016). It seems also that educated parents are able to use the time they spend with their children more effectively, by adapting what they do with their children based on the child's stage of development (Kalil et al., 2012). This research from the United States shows, for instance, that educated mothers spend more time caring, comforting and playing with their children when they are infants and toddlers, more time in teaching activities (e.g., reading, helping with homework) when children are at school entry, and more time managing, organising and attending outof-home activities during middle childhood and adolescence (Kalil et al., 2012). Evidence for a developmental gradient in educated mothers' time use patterns have been found in Spain (Gracia, 2014) and Italy (Rebane, 2015).

Recent Australian research, drawing on data from the Birth and Kindergarten cohorts from the Longitudinal Study of Australian Children (LSAC; *N* = 7007 children), extended this work by examining whether the developmental gradient in maternal time use extended to fathers and other caregivers (O'Flaherty & Baxter, 2020). The fndings supported an overall education gradient, in that, compared to mothers without a University degree, children of mothers with tertiary education degrees spent more time with mothers, fathers and other caregivers in teaching (e.g., reading, telling stories, talking) and enrichment activities (i.e., organised lessons and activities), and more time with fathers in play activities. In comparison, only partial support was found for the notion that better educated mothers tailor their activities according to the development of their child. The children of educated mothers spent more time in teaching activities when they were preschool-aged with their mothers, fathers and other caregivers, which was consistent with the work of Kalil et al. (2012). However, there were also large educational gaps in teaching activities during infancy, favouring the children of more educated mothers, and no

developmental gradient was observed in play or enrichment activities. Given that these fndings held even when family economic and social resources were controlled, the authors concluded that variation in time spent with children was likely due to better understanding among educated mothers of the value of teaching and enrichment activities for children than less educated mothers (O'Flaherty & Baxter, 2020), perhaps refecting sociocultural infuences on parental values among more socially advantaged Australian families.

#### **Discipline Practices**

Parents' use of ineffective or harsh discipline practices have been reliably linked to detrimental child and adolescent psychological and developmental outcomes (Bayer et al., 2011; Hoeve et al., 2009; McKee et al., 2007). Ineffective discipline practices include inadequate or permissive limit-setting, inconsistent or inappropriate application of consequences for problem behaviour, and poor monitoring of children's whereabouts and activities. Harsh discipline practices include those that are negative and coercive, are characterised by hostility and criticism, or involve physical punishment. These latter type of discipline practices are associated with child maltreatment (Theodore et al., 2005).

With increasing focus on evaluating more complex ecological models of social and family infuences on child development, research examining the direct effects of social disadvantage on discipline practices has dwindled in recent years (Roubinov & Boyce, 2017). International research using large population-level samples mostly comprising mothers, suggest that there is strong evidence that living in social disadvantage is a major risk factor for engaging in harsh discipline practices (Jansen et al., 2012; Weis & Toolis, 2010) and physical punishment (Barkin et al., 2007; Berger, 2005). Further, research within at-risk samples indicates that cumulative social disadvantage may place parents at particular risk of harsh discipline (Arditti et al., 2010; Pereira et al., 2015), and characteristics of the family neighbourhood (e.g., level of danger, access to services) also increase the likelihood of harsh parenting (Pinderhughes et al., 2001).

When it comes to child maltreatment, recent population-based data from outside the United States, including from Australia (Doidge et al., 2017) and Canada (Lefebvre et al., 2017), indicate that economic hardship and social instability (e.g., household and school mobility) are important risk factors. Offcial Australian child maltreatment statistics indicate that child maltreatment disproportionately affects children from more socially disadvantage backgrounds. Over a third of children (36%) aged 0–12 years who had at least 1 substantiated notifcation of child maltreatment during 2017–2018 came from the lowest socioeconomic group of families, compared with 6% of children from the highest socioeconomic group of families (Australian Institute of Health and Welfare, 2020).

One major gap in the literature that is slowly being addressed is the limited information about fathering generally, and paternal discipline practices more specifcally (Guterman & Lee, 2005), which is important to study given the increasing involvement of fathers across the community in children's lives (Cano et al., 2019). The available research on paternal discipline is mixed, with some research showing an association between economic and employment markers of social disadvantage (Lunkenheimer et al., 2006), and other studies fnding no relationship (Jansen et al., 2012; Lee et al., 2008). Clearly, further research is needed to assess differences and similarities in determinants of mothers' and fathers' discipline practices.

#### **Warmth and Support**

Another key domain of parenting practices are the behaviours parents engage in to build strong and positive relationships with their children. A large body of evidence, including large-scale experimental research (Kim et al., 2018) and behaviourgenetic studies (Roisman & Fraley, 2012) indicate that children will have better academic, behavioural, social, and psychological outcomes across childhood and into adolescence when parents are warm, sensitive, encouraging, and responsive to their physical, emotional and psychological needs (Biglan et al., 2012). In comparison to the research on discipline practices, much less research has explored the relationship between social disadvantage and warm and supportive parenting. There is evidence, however, that both individual and neighbourhood markers of social disadvantage are associated with lower levels of maternal warmth and responsivity (Evans et al., 2008; Odgers et al., 2012; Weis & Toolis, 2010), and autonomy supportive behaviours (Van Holland De Graaf et al., 2018) across childhood and adolescence. Further, an Australian 8-year longitudinal study of 166 adolescents found that warm and supporting parenting buffered the effects of neighbourhood disadvantage on adolescent frontal lobe development, with carryover effects on academic functioning (Whittle et al., 2017). Given the potential moderating role of positive parenting in minimising the effects of social disadvantage on children, further research is needed to examine the impact that living in social disadvantage has on both mothers' and fathers' ability to be responsive, supportive and encouraging of their children and adolescents.

#### *Home Environments*

#### **Home Learning Environment**

Parents contribute to creating a home environment for their children in many ways. This includes the physical aspects of the home environment, from the provision of stable, safe, and clean housing, and appropriate bedding and furnishings; to the availability of toys, books and other activities and resources that make the environment stimulating and enjoyable for children.

Because of its association with child cognitive, language, and academic outcomes (Christensen et al., 2014; Dubow & Ippolito, 1994), a large body of research has focused on variability in home stimulation and enrichment across levels of social disadvantage. Both cross-sectional (Hurt & Betancourt, 2019) and longitudinal (Cooper et al., 2010; Rodriguez & Tamis-LeMonda, 2011; Toth et al., 2020) research suggests that parents experiencing higher levels of social disadvantage provide fewer developmentally-appropriate play resources (e.g., puzzles, craft, blocks and toys requiring fne motor coordination, pretend play activities, board games), are less likely to engage in informal (e.g., shared book reading) and formal (e.g., helping children to write their name) literacy activities, and are less likely to provide variability in cognitive stimulation (e.g., through conversations and storytelling, visits and outings). These fndings were replicated in a recent study looking at neighbourhood markers of disadvantage, with living in a poorer quality neighbourhood associated with lower levels of maternal stimulation (Rhoad-Drogalis et al., 2020). Later in childhood and during adolescence, social disadvantage also predicts lower levels of parental involvement in children's school learning, including in overt ways (e.g., supervising homework, encouraging reading, communicating with teachers, being involved in school events; Camacho-Thompson et al., 2016; Cooper et al., 2010; Toth et al., 2020) and more covert or subtle ways (e.g., attitudes towards school completion, expectations about their child's achievement; Stull, 2013; Wang et al., 2016).

There is minimal Australian research on the relationship between social disadvantage and the home learning environment. In one study, using data from the Birth cohort of LSAC (*N* = 3836), Hayes et al. (2018) found that mothers with a postsecondary school education were more likely to engage in shared book reading and developmentally appropriate activities with their toddler, and had a slower rate of decline in shared book reading from aged 2 to 6 years, compared to mothers with a secondary school education. However, there were no differences based on maternal education in declines in other home learning activities, and there were only weak and inconsistent associations between family income and shared book reading and home learning activities from aged 2 to 6 years.

#### **Structure and Routines**

Parental behaviour also determines the level of routine and structure in the home. Predictability and consistency in a child's environment are achieved when there are clear and appropriate expectations for behaviour, children have consistent routines for mealtimes and bedtimes, and there are appropriate limits set and enforced for children's daily lives and habits, such as sleep, diet, physical activity, completing homework, and screen time. Conversely, what has been termed a 'chaotic' home environment, is characterised by a lack of routines, high levels of background noise and ambient stimulation (e.g., TV always on, loud conversations), overcrowding and high levels of 'foot traffc', clutter and disorganisation (Bobbitt & Gershoff, 2016; Evans et al., 2005), and more recently, frequent electronic intrusions (e.g., phone calls, messages or notifcations on electronic devices) (Whitesell et al., 2015). A recent scoping review of 112 studies found that household chaos is reliably linked to a range of adverse child, parent, and family-level outcomes (Marsh et al., 2020).

Many studies, mostly from the United States and the United Kingdom, have examined the relationships between various indicators of social disadvantage and household chaos. For instance, children and adolescents from low-income backgrounds are more likely to be living in household chaos than those from middle- or higher-income families (Bradley et al., 2001; Evans et al., 2005). When it comes to specifc aspects of children routines and daily activities, social disadvantage has been found to be associated with inconsistent bedtime routines, lower quality sleep environments and lower parental bedtime availability among infants and young children (Hale et al., 2009; Hoyniak et al., 2021). Among adolescents, those living in socially disadvantaged circumstances (Marco et al., 2012) and with higher levels of household disorganisation (Billows et al., 2009) report greater diffculty obtaining suffcient and consistent sleep throughout the week. Similarly, children and adolescents from disadvantaged backgrounds are less likely to consume breakfast as part of their morning routine (Vereecken et al., 2009), have regular family meals (Neumark-Sztainer et al., 2013), and have greater exposure to screen time (including TV viewing and technology use; Carson et al., 2010; Gorely et al., 2009) than their counterparts living in more affuent circumstances.

#### **Interparental Confict**

Parents play a pivotal role in determining the emotional climate of the home. This is highly infuenced by the way that people living in the home, particularly adults, get along with one another. Confict between parents is a known risk factor for poor child outcomes (Giallo et al., 2021; Teubert & Pinquart, 2010) and has been found to affect the wider family system. Meta-analytic evidence suggests that interparental confict, particularly when it is characterised by high levels of hostility, tends to 'spillover' into parents' interaction with their children, making them more likely to engage in harsh discipline, and less likely to be warm and supportive toward their children (Krishnakumar & Buehler, 2000). Further, confict in the parental relationship is associated with poorer quality sibling relationships (Zemp et al., 2021).

There is strong evidence that families living in social disadvantage are at increased risk of interparental confict, both for married couples and for unmarried cohabiting couples (Conger et al., 2010). It is likely that the distress associated with social disadvantage, such as being under fnancial pressure or experiencing job or housing insecurity, makes it diffcult for partners to relate to one another in a calm and reasonable manner, and more likely for them to engage in either verbal (e.g., heated arguments, shouting and swearing, critical comments and insults, threats, displays of hostility and anger) or physical confict (e.g., pushing, shoving, hitting; Conger et al., 2010; Westrupp et al., 2015). Recent Australian research using mother reports indicated that verbal and/or physical interparental confict is present in around 35% of Australian families with children aged up to 8 years, with up to 6% of families experiencing persistent interparental confict during early and middle

childhood (Westrupp et al., 2015). This study, which drew on data from both cohorts of LSAC (*N* = 9080) found that social disadvantage was reliably associated with both single time point and persistent interparental confict over several years (Westrupp et al., 2015). This fnding has been corroborated in research drawing on father-report data from the birth cohort of LSAC (*N* = 4136), which found that fnancial hardship during their child's frst year of life was associated with high and increasing levels of interparental confict from infancy to aged 10–11 years (Giallo et al., 2021).

This impact of living with social disadvantage on the couple relationship also holds for more extreme forms of interparental confict, namely intimate partner violence (Abramsky et al., 2011; Goodman et al., 2009). Australian research using 30-year follow up data from the Mater-University of Queensland Study of Pregnancy (*N* = 2401) found that both male and female partners were more likely to report being a victim of intimate partner violence when living in economically disadvantaged circumstances (Ahmadabadi et al., 2020). Women living with social disadvantage may be at particular risk when they are pregnant or parenting young children. Examining hospital admissions for the mothers of children born in Western Australia from 1990 to 2009, Orr et al. (2021) found that neighbourhood-level social disadvantage was an important predictor of a mother being hospitalised for intimate partner violence, both during the 12 months preceding the birth of their child, and up until the child was 3 years old. Thus, overall, there is strong evidence internationally and within Australia that social disadvantage creates a context that places parents under pressure and can detrimentally affect the physical and emotional environment that parents create for their children.

#### *Parental Stress and Wellbeing*

Parenting is often described as one of the most challenging, yet rewarding, roles of an individual's life. Being a parent provides meaning and purpose, which can have benefts for an individual's wellbeing. However, the opposite is also true. Balancing the responsibilities, demands and conficts of parenting, particularly when this occurs in diffcult circumstances, can affect parental wellbeing and life satisfaction.

When a parent is under signifcant stress, the capacity to be available, responsive, and patient with a child, and respond to problem behaviour in a calm, consistent and appropriate manner is signifcantly reduced. Meta-analyses suggest that mental health problems, particularly depression, are associated with more harsh and negative parenting, and decreased warmth and supportive parenting in both mothers (Lovejoy et al., 2000) and fathers (Wilson & Durbin, 2010). Dealing with the many characteristics of social disadvantage, including economic hardship, unstable or poor accommodation, restricted access to resources, and dangerous or chaotic neighbourhoods, alongside the everyday tasks and stressors of raising a family, is likely to be highly stressful for parents. However, research on social disadvantage

and parental mental health is limited, with a tendency to statistically control for social disadvantage, rather explicitly testing the relationship between stress associated with social disadvantage among parents (Gotlib et al., 2020). There is, however, some evidence from Australia (Baxter et al., 2012) and internationally (Borre & Kliewer, 2014; Lyons-Ruth et al., 2002) that social disadvantage detrimentally affects parents' mental health.

#### *Intergenerational Transmission of Parenting Practices*

In Australia, research indicates that family circumstances are transmitted across generations. For instance, Australian children of parents who receive welfare payments are almost twice as likely to be on welfare payments as adults when compared to those who are not (Cobb-Clark et al., 2017), while other work has demonstrated the transmission of outcomes related to wealth (Lersch & Baxter, 2021), health (Huang, 2020) and education (Hancock et al., 2018). One likely mechanism for this intergenerational transmission of individual social and economic outcomes is the infuence of parents and parenting.

While literature from Australia is lacking, international research indicates that parenting practices are at high likelihood for intergenerational transfer, such that a child's own experiences of being parented infuence the way in which they will parent their own children. For instance, Capaldi et al. (2008) found that childhood reports of harsh discipline among at-risk men were related to their own use of harsh discipline with their 2- to 3-year-old children. Similarly, Chen and colleagues (2008) in their three-wave longitudinal study of 1560 of students, found that perceived positive experiences with parents during early adolescence were positively related to marital satisfaction and educational attainment in early adulthood, which was subsequently positively related to the student's later use of constructive parenting with their own children. Parenting has also been shown to impact the effects of adversity across generations. Bailey et al. (2009) found that high levels of positive parenting in one generation lessened the impact of adversity for that generation and also increased the likelihood that the next generation used high levels of positive parenting with their children. This type of intergenerational research provides important information regarding the protective role of parents, and their capacity to buffer their own children from the harm associated with adversity, resulting in better outcomes for the next generation.

#### **Parenting and the Community**

For many families the already challenging task of parenthood is complicated by the context in which they live. The parenting role is likely to be more stressful, demanding and challenging when families live in neighbourhoods where there are high levels of poverty (i.e., low household incomes, high unemployment, high incidence of single-parent families), greater household crowding and high density living, and high levels of neighbourhood disorder (i.e., vandalism, abandoned or deteriorating housing, unsupervised teenagers, high residential mobility, poor access to health care, leisure and educational facilities).

Research has documented the impact of adverse neighbourhood contexts on parenting practices and the subsequent impact on children (Ceballo & McLoyd, 2002; Odgers et al., 2012). Jocson and McLoyd (2015) in their study of low-income, multi-ethnic families of children aged 6–16 years, found that neighbourhood disorder and housing instability were related to higher levels of parental distress and the subsequent use of higher levels of harsh and inconsistent discipline practices, with accompanying lower levels of warmth displayed towards their children. This suboptimal parenting was in turn associated with higher levels of child internalizing and externalizing behaviours 3 years later. Poorer neighbourhood quality has also been implicated in both higher and lower usage of parental monitoring strategies (Cobb-Clark et al., 2018), although recent research indicated that neighbourhood collective effcacy increased both parents' knowledge and limit-setting regarding their adolescent's whereabouts and activities (Zuberi, 2016).

Access to support, in the form of emotional, instrumental or practical support, is another factor that has been shown to infuence the quality of parenting provided to children (Green et al., 2007) and outcomes for children (Oravecz et al., 2008). Lower levels of social support have been associated with higher levels of maternal parenting stress (Hong & Lee, 2019) and parenting behaviours (Byrnes & Miller, 2012). Emotional and instrumental (e.g., fnancial, physical assistance) support have been associated with more effective parenting practices such as increased parenting consistency, better parent-child communication and parental monitoring (Byrnes & Miller, 2012; Marra et al., 2009) and have been associated with reduced anxiety in the attachment relationship (Green et al., 2007).

The protective impact of emotional and instrumental social support for parents has been shown to decline in more disadvantaged communities. The resulting increased parental social isolation has been found to be associated with reduced nurturing parental behaviours and effective use of monitoring and discipline practices (Ceballo & McLoyd, 2002), placing families further at risk. For example, in their study of African American single mothers, Ceballo and Mcloyd (2002) found that social support was less benefcial for positive parenting behaviours in more disadvantaged communities, with a stronger relationship between higher social support and lower punitive parenting in higher quality neighbourhoods. In comparison, levels of social cohesion in a neighbourhood have been found to be related to greater parental social support, with parents who reported higher social support also reporting more effective parenting (Byrnes & Miller, 2012; Maguire-Jack & Wang, 2016). Community processes such as collective effcacy, social capital, trust and social ties have also been shown to have an infuence on parenting. Low levels of these processes have been associated with parenting practices such as corporal punishment, ineffective monitoring and reduced warmth or nurturance and with child outcomes

such as child maltreatment, truancy and antisocial behaviours (Ma, 2016; Ma et al., 2018).

#### **Parenting Support Programs as a Solution for Addressing Intergenerational Transmission of Social Disadvantage**

Overall, the evidence reviewed here shows that the context in which parenting occurs is a critical infuence on the parenting that a child receives. While many parents adjust to the ever-shifting context, most do so with support and advice from others, including from both formal (e.g., GPs, psychologists, counsellors, spiritual and community leaders) and informal (e.g., partners, family, friends, neighbours) sources. The type of advice or support needed by a parent is likely to be determined by the context in which the parent is raising their children and the phase of the life cycle they and their children are negotiating. Additionally, the intensity of support needed by a family will vary according to factors such as the child's temperament, the level of adversity experienced, the parent's own wellbeing and the social and economic stressors facing the family. Some parents may need longer-term, intensive, and one-to-one support from a qualifed health professional to address the multiple and complex issues facing their child and family. For others, access to brief, evidence-based information will be suffcient to address any parenting concerns. Across the lifespan the same parent may fnd themselves requiring information or support at different times and for a range of issues relevant to their current context.

Thus, programs, interventions and policies designed to support parents must consider context and seek to balance fexible tailoring to this context while adhering to evidence-based models of parenting support. Strategies should be designed to acknowledge the complex and interactive effects of community, family, and individual aspects of social disadvantage on parenting, along with the intergenerational legacy of social disadvantage on families.

There are several powerful examples from within Australia that parenting interventions delivered in the context of social disadvantage can improve outcomes for children aged under 12 years and their families, such as the Triple P–Positive Parenting Program (Sanders et al., 2004), Smalltalk (Hackworth et al., 2017) and Tuning Into Kids (Duncombe et al., 2016). In comparison, very little research has been conducted on programs targeted specifcally at providing parenting support to parents of adolescents, particularly among socially disadvantaged families. One study from the United Kingdom, however, indicated that a program designed specifcally to support socially disadvantaged parents of adolescents was feasible and well-accepted by parents (Michelson et al., 2014).

Available research with parents of children indicates that parents across all social contexts beneft from parenting interventions (Leijten et al., 2013). Further, because their positive effects persist over time, parenting interventions may even facilitate reduction in social disparities brought about by ongoing conduct problems in children (Gardner et al., 2019). However, a key challenge across the entire parenting population, but particularly with socially disadvantaged parents, is improving acceptability, engagement, and ongoing participation in evidence-based parenting programs (Piotrowska et al., 2017). Thus, we need to investigate alternative models to improving acceptability and engagement in parenting programs that take account of the many barriers, both internal and external, on parents living with social disadvantage. Several promising approaches have received empirical support, including peer co-facilitator frameworks (Day et al., 2012), drawing on multiple delivery modalities (e.g., home visiting, group workshops; Morrison et al., 2014), incorporating technology-assisted approaches (Harris et al., 2020), or by incorporating sustained, long-term and developmentally targeted support (Doyle, 2020).

Another alternative to improve engagement of families experiencing disadvantage is to take a population-based approach to implementation of parenting interventions. In this approach, evidence-based parenting support is available to all families, from those experiencing signifcant and complex needs to those seeking to enhance and learn strategies to support the positive development of their child/ren. The best example of such a population approach involves the availability of interventions at different levels of intensity and across different delivery modalities. This helps maximise fexibility and accessibility for parents living in complex circumstances and helps cater for the varying support needs of parents living with social disadvantage. A population health strategy for parenting is consistent with advocacy for the adoption of population health approaches in the child and adolescent mental health (Patel et al., 2007) and family services sectors, including child protection (Prinz & Sanders, 2007).

Several population-level trials of the multi-level system of parenting support, Triple P, have shown success in achieving improvements in child and family outcomes (Doyle et al., 2018; Sanders et al., 2008; Zubrick et al., 2005), as well as in community indicators of child maltreatment (Prinz et al., 2009). More recently, a population-based approach to the implementation of Triple P has been conducted in Australia with the aim of exploring the impact of a whole of community approach to parenting support on community level indicators of child wellbeing across communities experiencing relatively high levels of social disadvantage. While data collection for this study is still being fnalised, this implementation took a unique approach in that it aimed to support parents by building both individual skills and knowledge, and by activating collective effcacy at a community-level to view parenting as a means for producing better outcomes for children. Overall, however, much more work needs to be done to design and evaluate parenting programs that meet the diverse and complex needs of families living with social disadvantage, and that acknowledge that parenting support is likely to be needed across childhood, adolescence and beyond.

#### **Implications for Policy**

Reddel and colleagues in Chap. 14 of the current volume, noted that "living in a community where there are high rates of poverty or other indicators of disadvantage is a strong predictor of experiencing persistent disadvantage, and addressing disadvantage at the community or place-based level is seen as an important pathway in moving people out of entrenched disadvantage". Parenting is one mechanism by which governments seek to intervene. Government policies have signifcant bearing on parenthood. Much of this effort is focused on structural and fnancial support for the task of child care and increasing capacity for mothers to remain in the labour force (e.g., income support payments, childcare subsidies, paid parental leave). Other efforts apply statutory mechanisms that seek to promote child safety and engagement with education (Reddel, 2002).

As the disparity between those who have and those who do not has grown, governments have developed policies and funded services designed to better support the most vulnerable members of society. Policies designed to reduce child maltreatment, provide health care and safe neighbourhoods via crime reduction are all examples of strategies that have implications for parenting. Policies designed to build safe neighbourhoods also have the potential to impact parenting. Research shows that parenting is adversely effected when living in dangerous neighbourhoods and that effective parenting can act as a protective factor against the adversities associated with living in poverty and/or dangerous areas (Ceballo & McLoyd, 2002). Further, living in neighbourhoods characterized by common goals such as ensuring the health and safety of children and where services such as mental health and substance abuse support are accessible, are linked with lower levels of child maltreatment (Maguire-Jack & Klein, 2015).

In recent decades, policy makers have turned to initiatives and policy directives directly targeting parenting practices associated with child health issues. A primary example of this is the way that sleep safe policies for infants, and their associated public education campaigns have been shown to enhance parenting practices associated with these issues and to reduce the incidence of Sudden Infant Death Syndrome (SIDS; Pollack & Frohna, 2002; Moon & Task Force on Sudden Infant Death Syndrome, 2011). Given the detrimental effects of alcohol consumption during pregnancy, particularly the risk of Foetal Alcohol Syndrome (FASD), government initiatives have also begun to target alcohol use during pregnancy. For example, the Australian Government developed an action plan aiming to take a whole-ofpopulation approach to reduce the impact of FASD across Australia (Foundation for Alcohol Research and Education, 2012). However, as is often the case, rigorous evaluation of such initiatives are limited.

In recent decades governments have also turned to policy and funding initiatives that directly target parenting practices via the implementation of parenting support programs. Internationally, some governments take an evidence-based approach to their focus on parenting, while others have tended to take a more localized community approach with locally developed programs and less focus on manualized evidence-based programs. Generally speaking, the approach to parenting support has tended to be localized to focus on specifc contextual issues, but with many regions moving towards a greater emphasis on agencies selecting programs that are identifed as evidence-based on an approved registry, such as the California Evidence-based Clearing House for Child Welfare (CEBC; www.cebc4cw.org) or Blueprints for Healthy Youth Development (www.blueprintsprogram.com).

Reddel (2002) noted that place-based approaches targeting complex social issues in the Australian context have been characterised by trials, pilots, and time-limited programs, and a narrow focus on human service delivery rather than broader policy design. This is also true in relation to parenting initiatives, with numerous examples of place-based programs undertaken over the last two decades that were designed to specifcally target parenting practices and the parent-child relationship. These have tended to be state-based and time-limited with minimal evaluation of outcomes. The Triple P– Positive Parenting Program (Sanders, 2012) is arguably the most successful parenting program both nationally and internationally, and offers an excellent example of the short-term and fragmented approach to government policy relating to parenting. Since the mid 1990's Triple P has been subject to multiple large-scale government funded implementations across multiple Australian states, including Western Australia (Zubrick et al., 2005), Victoria (Cann et al., 2003), and Queensland (Sanders et al., 2008). Each of these initiatives has been funded or implemented at a state level. Some have targeted whole-of-population while others have targeted specifc regions. With the exception of Western Australia, where the program is embedded within their family-related policy and services, implementation in other states has been time-limited or small in scale. Queensland undertook a signifcant roll-out of the Triple P program during the 1990's and early 2000's before again funding a substantial state-wide implementation of the Triple P system of parenting support in 2015. Positively, the current implementation has been a sustained government priority since that time.

Triple P is by no means the only parenting intervention to receive support from government, with programs such as 1-2-3 Magic (Phelan, 2003) and Circle of Security (Powell et al., 2014) also subject to government funding. However, typically, this support is funded at the service level with agencies receiving funding to deliver parenting support to their consumers. Increasingly, the funding is tied to a specifc program, however, service-level agreements also often allow local agencies to determine the type and intensity of support for parenting that is provided.

That government policies have begun to invest in evidence-based parenting interventions represents an important shift in recognition of the role of government in supporting parenting to reduce social and health issues that have persistent and intergenerational effects on individuals and community indicators of social disadvantage and adversity. However, to be effective in producing long-term shifts in factors that promote effective parenting within communities and across generations, policies need to be developed systematically with sustainability in mind. Further, embedded high quality evaluations are needed to ensure that government resources are being effectively applied and to enable more comprehensive analysis of the impacts of programs on parenting, children and community contexts.

#### **Directions for Future Research**

Very few studies have examined the mechanisms for the associations among social disadvantage, parenting and parental wellbeing. Family stress models emphasise how social disadvantage places the family system under pressure, and thus the mechanisms for this association need to acknowledge the complex and bidirectional relationships between family economic pressure and hardship, parental personal stress and mental health, interparental confict, and harsh and ineffective parenting (Conger et al., 2010). In comparison, investment models argue that greater fnancial and economic prosperity, along with higher educational and occupational status within families, increase childrearing activities, values and expectations that foster the social and academic success of children (Bradley & Corwyn, 2002). Thus, within an investment model framework, more advantaged parents have the knowledge and means to invest time and resources into their child's long-term development, whereas more disadvantaged families must invest in more immediate family and child needs (e.g., safety, stable housing, food, basic school resources). It is likely that social disadvantage acts on parenting and parental wellbeing along both family stress and investment pathways. Thus, a major direction for future research is to test comprehensive, interactional models of the infuence of social disadvantage on families and children.

An important message from this chapter is that parenting continues well beyond infancy and early childhood and is a dynamic and evolving role that is strongly infuenced by the past and current social context and structures a parent experiences. However, when it comes to research on parenting support, much of the focus has been on the transition to parenthood, and the infancy and early childhood periods of parenting. There are signifcant opportunities at other stages of parenthood, including adolescence, emerging adulthood, and grandparenthood, where supporting parents is likely to make important differences for individuals and communities living with social disadvantage. For example, grandparents who provide informal care of their grandchildren make up a signifcant amount of the childcare burden, with rates estimated as greater than 20% in Australia ([ABS], 2012). Grandparents in these circumstances face the challenge of balancing provision of support and respect for their own child's parenting with the need to make moment-by-moment parenting decisions for the grandchild under their care. Thus, further research and development and dissemination of parent support is required to ensure that effective and developmentally responsive support is available across the life course.

#### **Conclusion**

Parenting is fundamentally linked to the wellbeing and development of children. The quality of parenting that a child experiences has clear impacts on their life outcomes during childhood and into later life with high likelihood of transfer across

generations. The demands and stressors of parenting shift over time according to the child's developmental needs and the context in which parenting occurs. Individual, community and neighbourhood factors each infuence and shape parenting and the parenting experience, and in turn outcomes for children. However, this infuence is complex and multi-directional. These factors impact parenting and parental wellbeing, as well as directly and indirectly increase risk of adverse child outcomes (mental health, alcohol and other drugs, maltreatment, family violence) via an association with parenting. In addition, effective parenting has a protective function, promoting child wellbeing and reducing the impact of contextual factors associated with social and economic disadvantage. Research is needed to comprehensively assess the mechanisms for the associations between parenting, parental wellbeing and social disadvantage. Parenting interventions, particularly those that target parenting practices and beliefs known to be effective, have shown positive effects on a range of child outcomes and across developmental stages of childhood and adolescence and have shown positive effects across levels of relative advantage and disadvantage. The most rigorous evidence is for interventions delivered during early childhood however, some evidence is also available for adolescence. As the family context changes, research is needed to explore whether these positive effects extend to other life stages (e.g., parenting an emerging adulthood, grandparenting). Despite decades of evidence for parenting interventions government policies remain fragmented with funding generally embedded within services that provide care for the most vulnerable or offered for time limited pilot programs or population-based implementations with only minimal evaluation. Systematic and sustained populationbased approaches that incorporate universal and targeted approaches are needed to ensure that parents across the community have access to evidence-based support tailored to their specifc needs. Such approaches will maximise opportunities to reduce the harms associated with suboptimal parenting and the intergenerational transmission of parenting practices.

#### **References**


P-Positive Parenting Program with parents at-risk of child maltreatment? *Behavior Therapy, 35*, 513–535. https://doi.org/10.1016/S0005-7894(04)80030-3


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 12 Intergenerational Processes of Disadvantage in the Lives of Lesbian, Gay and Bisexual Australians: From Relationships with Parents to Parenting Expectations**

#### **Alice Campbell and Francisco Perales**

While structural stigma has decreased substantially over the past two decades (Perales & Campbell, 2018), it remains a pernicious force constraining the opportunities and wellbeing of lesbian, gay, and bisexual (LGB) Australians. Structural stigma comes in several forms, including discriminatory policies and legislation such as the ban on same-sex marriage that existed in Australia until the end of 2017—and negative attitudes and stereotypes about LGB people. It also includes heteronormativity, an ideology that positions heterosexuality as the default norm and same-sex sexuality as inferior and "other". Heteronormativity further incorporates beliefs about the roles that men and women are expected to play in society, how relationships should be conducted, and what "family" means. It can provide an inhospitable backdrop against which the family lives of LGB Australians unfold.

Structural stigma can produce disadvantage among LGB people across several life domains, including employment and earnings (Mize, 2016) and educational attainment (Mollborn & Everett, 2015). However, the largest and most consistently observed disparities between LGB and heterosexual people tend to be for mental health, emotional wellbeing, self-harm and suicidal behaviours (e.g., Perales, 2016, 2019; Perales & Campbell, 2019). In locations where levels of structural stigma are higher, these disparities are even greater (Perales & Todd, 2018). Of concern, the disadvantages LGB people experience with regards to their mental health and

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia e-mail: alice.campbell@uq.edu.au

F. Perales

A. Campbell (\*)

Australian Research Council Centre of Excellence for Children and Families over the Life Course, School of Social Science, The University of Queensland, Brisbane, QLD, Australia e-mail: f.perales@uq.edu.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_12

emotional wellbeing have persisted in recent times, even though levels of structural stigma have started to decrease (Perales, 2019).

According to the life course principle of "linked lives", social structures can impact even our most personal and intimate relationships, with potentially harmful consequences for individual wellbeing. Chapter 2 provides a more extensive discussion of life course approaches. Consistent with this, one of the principal channels through which structural stigma produces mental health disadvantage among LGB people is by shaping their relationships within key contexts—including families, schools, workplaces, and communities. Recent evidence demonstrates that lower levels of social support and more frequent experiences of bullying, victimisation, and rejection in these contexts contribute to poorer mental health and emotional wellbeing among LGB Australians compared to their heterosexual peers (e.g., Perales & Todd, 2018; Perales & Campbell, 2020). However, signifcant gaps in our knowledge remain, and these must be addressed if we are to eradicate the mental health disadvantage experienced by LGB people. To that end, in this chapter we apply the life course perspective to explore the role of family dynamics in producing mental health disadvantage in the lives of LGB Australians.

In part one we focus on one of the most fundamental and enduring relationships or linked lives: that between a person and their parent/s. It is well-established that parents play a key role in shaping the emotional wellbeing of LGB youth (Perales & Campbell, 2020). However, most prior examinations of the parent-LGB child relationship have focused on a single point in time—overwhelmingly, adolescence. We build on this evidence by taking a longitudinal approach and drawing on the life course principles of "trajectories" and "turning points". Following individuals across the life course, we investigate levels of closeness, support, and overall satisfaction experienced by LGB and heterosexual people in their relationships with their parents. This allows us to identify when disparities by sexual orientation in the parent-child relationship emerge, the trajectories they follow over time, and the existence of turning points when relationships signifcantly worsen or improve. We also examine associations between parent-child relationship quality and emotional wellbeing from childhood into middle age, thereby shedding light on how parents contribute to mental health disadvantage among LGB people at different stages of the life course.

In part two we turn to parenthood from the focal point of view of the LGB person. We explore the parenting desires and expectations of childless Australians according to their gender and sexual orientation, as well as the associations between these and mental health. Studies from outside Australia have found that lesbian women and gay men are less likely to report aspirations or intentions for parenthood than their heterosexual peers (e.g., Gato et al., 2020; Riskind & Tornello, 2017). These differences appear to be attributable to the practical and psychosocial barriers to parenthood that lesbian and gay people face, which can lead them to believe that their identity is incompatible with being a parent and that they will be unable to attain parenthood even if they desire it. Given how fundamental family is to the human experience, having parenting desires suppressed or thwarted can result in negative repercussions for one's mental health (Payne et al., 2019). Therefore, this constitutes another example of how family dynamics can produce disadvantage in the lives of LGB people. Here, we investigate this in the Australian context for the frst time.

We also consider how disadvantageous family dynamics can accumulate within the lives of LGB people and across generations by posing a novel question: do poorer-quality relationships with parents constrain the parenting desires and expectations of LGB people? Positive relationships with parents may buffer LGB people against some of the negative impacts of heteronormative social structures, reducing the psychosocial barriers to parenthood they would otherwise face. In the other direction, being rejected by one's parents may increase internalised homophobia and harmful self-concepts among LGB people and decrease their expectations of receiving support as a parent. We therefore investigate whether LGB people who are less satisfed with their relationships with their parents are not only more likely to suffer from lower emotional wellbeing, but also less likely to desire or expect to have children of their own.

#### **Part 1: LGB People and Their Parents**

#### *Parent-Child Relationships During Adolescence*

Adolescence is characterised by rapid physical, cognitive, social and emotional changes. A critical stage of the life course, it is the time when future trajectories in education, employment, health, wealth, and family begin to take shape. While adolescents increasingly turn their attention to the outside world and spend more time with their peers, as Chaps. 6 and 8 in this volume show, parents continue to play a central role in shaping their development and wellbeing. By providing comfort, support, reassurance and resources, parents act as the secure base their adolescent child needs as they step into the world as independent beings. By taking an interest in their child's life, spending time together in joint activities, engaging in their education and refecting a positive appraisal of them, parents facilitate the development of skills that are fundamental to achieving one's potential and living a happy and healthy life. These include emotional self-regulation, confdence, resilience, and a positive self-concept.

One group who can especially beneft from these parental investments are LGB youth. The development of sexual and romantic identities is a core task of adolescence, as young people become increasingly aware of their attractions to others (or lack thereof). This undertaking comes with additional complexities for individuals who experience same-sex attractions in the context of a heteronormative society. The social construction of heterosexuality as the default norm places a unique burden on individuals who are not exclusively heterosexual to "come out". As LGB youth become aware of and reveal their "difference" from their heterosexual peers, they are at an increased risk of being bullied and rejected in school and the broader community (Perales & Campbell, 2020; Robinson et al., 2014). Parents can play a pivotal role in buffering their adolescent child from the negative impacts of these hostile environments (Hillier et al., 2010). By communicating unconditional love and acceptance, parents can challenge heterosexist and homophobic beliefs and help prevent adolescents from internalising them (Carastathis et al., 2017).

#### **Parental Rejection of LGB Youth**

The homophobia that permeates outside environments can also dominate within the family home. Thus, not only are LGB youth more likely than their heterosexual peers to need their parents' acceptance and support, they are also less likely to receive it. Disclosing their sexual orientation to parents can be a traumatic event for LGB youth, given their emotional and material dependence on their parents coupled with the possibility of rejection (Hillier et al., 2010). Some LGB youth will distance themselves from their parents prior to coming out as they grapple with these anxieties. Regardless of whether their fears are eventually realised, these young people miss out on the benefts of a close and supportive relationship with their parents at a critical stage of their development. On a positive note, Australian evidence indicates that the proportion of LGB youth disclosing their sexual orientation to and receiving support from their parents has been increasing over the past two decades (Hillier et al., 2010). However, this same evidence shows that a substantial minority still do not feel able to come out to their parents, or experience rejection when they do.

#### **What Does Parental Rejection Look Like?**

Carastathis et al. (2017) interviewed 21 gay and lesbian Australians about their lived experience of being rejected by their parents because of their sexuality. They found that rejection was sometimes subtle. For example, parents expressed disappointment and sadness that their child would not achieve the normative ideal of heterosexual marriage and parenthood that they had envisaged. While not explicitly rejecting their child, they instead rejected their child's partner or avoided any discussion of that aspect of their child's life. In other instances, parental rejection was more blatant. Some parents withdrew warmth, nurturing, interest and concern from their LGB child. Others subjected their child to verbal abuse or condemnation. In the most extreme cases, parents became a risk to their child's physical safety.

In the only study of its kind, we analysed time-use diaries completed by more than 3000 young people aged 14–15 years (Perales et al., 2020a). Over the course of a day, young people recorded where they were, what they were doing and who they were with. On average, we found that sexual minority youth (i.e., those who reported being attracted to the same-sex, both sexes, or who were unsure) spent the same amount of time with their mothers, but signifcantly less time with their fathers, compared to their heterosexual peers (i.e., those attracted to the other sex only). This is consistent with evidence that men hold more homophobic attitudes than women (Perales & Campbell, 2018), and that LGB youth are less likely to disclose to their fathers than their mothers (Hillier et al., 2010). Given the documented benefts for children and adolescents of time spent with an involved father (e.g., Cano et al., 2019), we concluded that reduced time with fathers may represent one mechanism of disadvantage affecting sexual minority youth.

#### **The Impacts of Rejection on LGB Youth**

The early teen years—the average age for coming out—are a particularly vulnerable time for homelessness risk among LGB youth (Dempsey et al., 2020). Physical and verbal abuse can leave LGB youth feeling unsafe at home, prompting some to fee (Hillier et al., 2010; Robinson et al., 2014). Other times, parents force their child to leave. For these youth, "home" can become a revolving series of refuges, foster homes, and hostels (Robinson et al., 2014). Changes of school may be frequent, resulting in interruptions to the young person's education. The negative implications of this disruptive adolescence for the remainder of the life course are not hard to imagine.

Even when LGB youth are "tolerated" by their parents and rejection is subtler, it is still likely to come at a high cost to the young person's social and emotional wellbeing. Parents are powerful agents of socialisation for their children. When they react negatively to their adolescent's sexual orientation and express homophobic attitudes, these can be internalised by the young person (Carastathis et al., 2017). Feelings of confusion, alienation and self-hatred can ensue, leading to mental health problems and self-destructive behaviours such as self-harm and substance abuse (Carastathis et al., 2017; Robinson et al., 2014). The ultimate consequences can be dire. A lack of family acceptance has been identifed as the defning feature of suicides by lesbian and gay Australians under the age of 30 (Skerrett et al., 2017).

In our own analyses of self-reported survey data from a nationally representative Australian sample, we found evidence of wide-ranging health and wellbeing disparities in LGB youth relative to their heterosexual peers (Perales & Campbell, 2019). These disparities were especially large for self-harm, suicidal thoughts and behaviours, socio-emotional functioning, and general health. LGB youth also rated their relationship with their parents as signifcantly less close and supportive than heterosexual youth on average. In a subsequent mediation analysis, we demonstrated that the health disadvantage experienced by LGB youth could be largely explained by these defcits in the parent-child relationship (Perales & Campbell, 2020).

#### *Parent-Child Relationships at Other Ages*

It seems unequivocal that parental rejection produces disadvantage in the lives of LGB youth. Whether blatant or subtle, the negative consequences of this rejection have the potential to reverberate throughout the life course. However, there is very little evidence on what happens in the parent-child relationship either before or after the critical stage of adolescence (Reczek, 2020). Internationally, the empirical literature on the family relationships of LGB persons has tended to focus on discrete points in time and an understanding of "how family relationships unfold and accumulate across the life course" has been missing (Reczek, 2020, p. 316). Taking adolescence as the starting point, our aim in the frst part of this chapter is to look both backwards and forwards to address these gaps in the literature.

#### **When Do Disparities in the Parent-Child Relationship Begin?**

Given that most LGB youth will come out in their early teens, it makes sense that the focus of the literature has been on this age group. Yet, there are plausible reasons why distance in the parent-LGB child relationship could emerge prior to adolescence. First, some LGB youth might withdraw from their parents prior to coming out if they expect that their parents will reject them. A substantial minority (15–36%) of LGB youth report being aware of their same-sex attractions before 10 years of age (Hillier et al., 2010; Robinson et al., 2014), so it is possible that distance in the parent-child relationship could be evident in earlier age years. Second, there is an association between non-heterosexuality in adulthood and gender non-conformity during childhood: on average, LGB people display higher levels of gender nonconforming behaviours during their childhood than heterosexual people (Xu et al., 2019). Some parents might reject their child because of their gender non-conformity or the perception that their child will grow up to be gender or sexuality diverse.

Identifying when disparities in the parent-LGB child relationship begin and how they develop across adolescence are important tasks. This knowledge can shed light on how disadvantage accumulates over the LGB life course and when and where we should intervene to prevent it. Yet, to our knowledge, no previous Australian studies have examined the parent-LGB child relationship across childhood and adolescence using longitudinal (prospective) data. In fact, there is only one international study that has done so. Xu et al. (2019) analysed data from the Avon Longitudinal Study of Parents and Children in the UK. They found an association between the quality of relationships with parents reported at 9 years of age and the sexual orientation reported by these same young people at age 15. Specifcally, adolescents who reported a non-heterosexual orientation had reported poorer relationships with their parents during childhood, on average, than adolescents who reported a heterosexual orientation. This association remained after levels of gender non-conforming behaviour during childhood were accounted for.

#### **How Do Disparities in the Parent-Child Relationship Track Across Adulthood?**

The second important gap in the literature that we address in part one of this chapter concerns how relationships between LGB people and their parents track across adulthood. It is possible that relationships improve over time as parents adjust to and accept their child's sexuality diversity. There is some evidence that getting legally married or becoming a parent can normalise and legitimise a child's nonheterosexuality in the eyes of their parents and lead to closer, more supportive intergenerational relationships (Power et al., 2012; Reczek, 2020). However, in a minority of cases these same events lead to parental rejection (Reczek, 2020). More than half of the lesbian and gay people in Carastathis et al.'s (2017) study continued to experience some level of rejection from family members years after disclosing their sexual orientation. Studies originating in the Netherlands (Fischer & Kalmijn, 2021) and Germany (Hank & Salzburger, 2015) have found that LGB adults are less close to their parents than heterosexual adults, although these differences are modest in size. On the other hand, Australian research has found that parents are more likely to make fnancial transfers to their adult child if their child is lesbian, gay or bisexual rather than heterosexual (Perales & Huang, 2020).

#### **What Are the Implications for Emotional Wellbeing and Mental Health?**

Finally, in part one of this chapter we investigate the implications of relationships with parents for the emotional wellbeing of LGB people across the life course. As we discussed earlier, the evidence on this for children and adolescents is robust, but researchers have yet to establish if parents matter to the same extent over time. On the one hand, the strong evidence that family ties shape health across the lifespan in the general population suggests that they do (Reczek, 2020). On the other hand, LGB adults may come to compensate for a lack of close, supportive relationships with their parents through relationships with partners, friends, and the LGB community. Consistent with this, Australian research found that LGB parents felt less connected to their families of origin but more connected to their friends than heterosexual parents (Power et al., 2015). Meanwhile, both groups reported similar levels of psychological distress, suggesting that among LGB people any negative impacts arising from a lack of support from their parents was mitigated by increased support from their friends.

#### *Generating New Empirical Evidence for Australia*

To answer these questions, we utilised data from two nationally representative samples of Australians. To investigate parent-child relationships and emotional wellbeing across childhood and adolescence we analysed data from the *Longitudinal Study*  *of Australian Children* (LSAC). In keeping with our previous analyses of these data (Perales & Campbell, 2020; Perales et al., 2020a), we compared two groups of young people: a heterosexual group (n = 4477, 82.5%) comprising those who reported being attracted to the other sex only, and an LGB group (n = 952, 17.5%) comprising those who reported being attracted to the same sex, both sexes, or who were unsure (questioning). Reports of sexual attractions were taken from the most recent survey in 2018, when young people from LSAC's younger cohort (the socalled "B" cohort) were aged 14–15 years and those from the older cohort (the socalled "K" cohort) were aged 18–19 years.

From ages 10/11, children in LSAC have reported on their relationships with their parents. Specifcally, they have been asked 8 questions from the Inventory of Parent and Peer Attachment (Armsden & Greenberg, 1987), measuring the extent to which they feel emotionally close to their parents, accepted by them, trust them and feel able to talk to them about their problems. We calculated the mean of the 8 items for each young person, with scores ranging from 1 (low on closeness) to 4 (high on closeness). From ages 10/11, children have also been asked to report on their emotional wellbeing using the emotional problems subscale of the Strengths and Diffculties Questionnaire (Goodman, 1997). Specifcally, they have been asked 5 questions measuring the extent to which they have felt worried, unhappy, nervous, and scared over the past 6 months. Mean scores for the 5 items were transformed to range from 0 (poorest emotional wellbeing) to 100 (best emotional wellbeing).

To investigate the parent-child relationship and emotional wellbeing across adulthood we analysed data from the *Household, Income and Labour Dynamics in Australia* (HILDA) Survey. Participants have been asked about their sexual orientation in Waves 12 (2012) and 16 (2016) of the HILDA Survey. Again, we divided the sample into two groups: one comprising those people who identifed as heterosexual (n = 17,579, 96%), and another comprising those people who identifed as lesbian, gay or bisexual in at least one of the two waves (n = 740, 4%).

Every year in the HILDA survey, participants rate how satisfed they are with their relationships with their parents on a scale from 0 (completely dissatisfed) to 10 (completely satisfed). They also report the state of their mental health over the past 4 weeks, including how often they have felt nervous, happy, or down, using 5 items from the SF-36 Mental Health Scale (Ware & Sherbourne, 1992). Total scores were transformed to range from 0 (poorest mental health) to 100 (best mental health).

#### *Results for Relationships with Parents Over Time*

Figure 12.1 below shows the average level of closeness to parents for heterosexual and LGB youth between ages 10 and 19. The frst thing to note is that the average score on closeness to parents for LGB youth is high: on a scale of 1–4, it never drops below 3. This suggests that most LGB youth have a good relationship with their parents across childhood and adolescence. However, on average, LGB youth were

**Fig. 12.1** Closeness to parents among heterosexual and LGB youth ages 10–19. Note: Predicted closeness to parents by age for LGB and heterosexual youth based on regression model with random effects, controlling for child's sex. (Source: Longitudinal Study of Australian Children. Waves 4 (2010) – 8 (2018), Cohorts "B" and "K")

less likely to feel close to and supported by their parents than heterosexual youth at all ages. While one-third (33.5%) of heterosexual youth rated their closeness to their parents at the maximum score of 4, only one-ffth (20.4%) of LGB youth did so. Meanwhile, LGB youth were twice as likely as heterosexual youth to rate their closeness to parents at a score of 2 or under (8.3% vs. 4.2%).

Irrespective of a young person's sexual orientation, it is normal for some distance to develop in the parent-child relationship during adolescence (Seiffge-Krenke et al., 2010). Consistent with this, levels of closeness to parents declined for heterosexual youth from 12 years onwards, while for LGB youth this decline began at 11 years and was more pronounced. This earlier and steeper decline in the parentchild relationship is consistent with arguments that some LGB youth are likely to withdraw from their parents as they become increasingly aware of their sexual orientation and contemplate the repercussions of disclosing it. In other cases, stress on the parent-child relationship may come later as LGB youth begin exploring their sexual and romantic feelings, engaging in sexual activity, and disclosing their sexual orientation to others. Average closeness to parents hit a low point for both groups at 17 years, before trending back upwards. However, at 19 years, when our period of observation ends, LGB youth continued to rate their relationships with parents as less close and supportive than heterosexual youth on average.

Turning now to adulthood, Fig. 12.2 shows predicted levels of satisfaction with relationships with parents for heterosexual and LGB adults between the ages 15 and 55. We chose 55 years as our end point because a substantial proportion of parents

**Fig. 12.2** Satisfaction with relationships with parents among heterosexual and LGB people ages 15–55.

Note: Predicted satisfaction with relationship with parents by age for LGB and heterosexual people based on regression model with random effects, controlling for respondent's sex. (Source: Household, Income and Labour Dynamics in Australia Survey. Waves 1(2001) – 19(2019))

will be deceased after this time, making results less reliable and more diffcult to interpret. Average levels of satisfaction were high for both groups—generally between 7 and 8 out of a maximum of 10—while differences between the two groups were modest. One in four (25.5%) heterosexual people reported being "completely satisfed" with their relationships with parents compared to one in fve (19.8%) LGB people. Meanwhile, LGB people were somewhat more likely than heterosexual people to report low levels of satisfaction with their relationships with parents (i.e., a score of 0–4: 12.2% vs. 7.5% respectively).

Consistent with our analyses of the LSAC data, we found that LGB adolescents in the HILDA Survey reported lower levels of satisfaction with their relationships with parents than heterosexual adolescents, a gap that persisted across their 20's. In fact, for LGB people the average level of satisfaction with relationships with parents never rose above that seen in adolescence (approximately 7.5 out of 10). This suggests that, if some LGB individuals do experience an improvement in their relationship with their parents over time, this must be offset by others who experience a deterioration. For both heterosexual and LGB people, satisfaction dropped to its lowest levels between the ages of 40 and 45, before increasing somewhat during their 50's. The gap between the two groups was smaller at this time than it was during adolescence.

#### *Results for Associations Between Relationships with Parents and Mental Health/Emotional Wellbeing*

Figure 12.3 shows the associations between closeness to parents and emotional wellbeing for LGB and heterosexual youth between the ages of 10 and 19. Higher values on the vertical axis refect better outcomes (i.e., greater emotional wellbeing). LGB and heterosexual youth who reported high levels of closeness to their parents reported similarly high levels of emotional wellbeing at age 10. However, following the top two lines across the graph, we can see that their trajectories diverged after that. Heterosexual youth who were close to their parents reported consistently high levels of emotional wellbeing across adolescence, while LGB youth who were equally close to their parents reported decreasing levels of emotional wellbeing as adolescence progressed.

The two lines at the bottom of the graph represent the predicted outcomes of LGB and heterosexual youth reporting low levels of closeness to their parents. At age 10, these two groups reported similar, moderate levels of emotional wellbeing. Among the heterosexual youth who did not feel close to their parents, predicted levels of emotional wellbeing declined gradually across adolescence from 57 to 47 points. Meanwhile, for the LGB youth who were not close to their parents the decline in emotional wellbeing was steeper, dropping from 52 to 30 points across adolescence.

**Fig. 12.3** Emotional wellbeing by closeness to parents among heterosexual and LGB youth ages 10–19.

Note: Predicted levels of emotional wellbeing by age and closeness to parents for LGB and heterosexual youth based on regression model with random effects, controlling for child's sex. (Source: Longitudinal Study of Australian Children. Waves 4(2010) – 8(2018), Cohorts "B" and "K")

Altogether, there are three key takeaways from this analysis. First, closeness to and support from parents is important for the emotional wellbeing of all youth regardless of their age or sexual orientation. Second, average levels of emotional wellbeing decrease for LGB youth across adolescence irrespective of how close they are to their parents. Third, LGB youth reporting low levels of closeness to their parents have the steepest drop in emotional wellbeing and, by the end of adolescence, exhibit substantially lower levels of emotional wellbeing than any other group. Thus, the combination of sexuality diversity and poor relationships with parents should be considered an important marker of vulnerability among young people.

Figure 12.4 is comparable to Fig. 12.3 and shows associations between satisfaction with relationships with parents and mental health in LGB and heterosexual people from ages 15 to 55. Again, higher values on the vertical axis refect more positive outcomes (i.e., better mental health). Consistent with what we observed in Fig. 12.3, the poorest mental health at age 15 was reported by LGB youth who were completely dissatisfed with their relationships with their parents. This group had a predicted average mental health score of 52 out of 100. Heterosexual youth who were dissatisfed with their relationships with parents fared somewhat better with an average predicted score of 59, yet they still had substantially poorer mental health than heterosexual youth who were completely satisfed with their relationships with parents (predicted score = 76). LGB youth who were completely satisfed with their relationships with parents fell in between with an average predicted score of 69.

**Fig. 12.4** Mental health according to satisfaction with relationships with parents among heterosexual and LGB people ages 15–55.

Note*:* Predicted mental health by age and satisfaction with relationship with parents for heterosexual and LGB people based on regression model with random effects, controlling for respondent's sex. (Source: Household, Income and Labour Dynamics in Australia Survey. Waves 1(2001) – 19(2019))

Similar to what we observed in the LSAC sample, heterosexual people in the HILDA Survey who were completely satisfed with their relationships with parents maintained consistently high levels of mental health over time. For LGB people who were completely satisfed with relationships with parents, average mental health scores also remained high. However, at no point did they reach parity with levels observed among their heterosexual counterparts.

Meanwhile, people who were completely dissatisfed with their relationships with parents experienced an improvement in their mental health as they aged, regardless of their sexual orientation. This suggests compensation and resilience: people who have negative relationships with their parents may compensate by forming close relationships with friends, partners, colleagues, and community, and over time their relationships with their parents may matter less for their emotional wellbeing. Nevertheless, the improvement in mental health among LGB people who were completely dissatisfed with their relationships with parents was less pronounced than it was for their heterosexual counterparts. By age 55, LGB people who were completely dissatisfed with their relationships with parents remained the most vulnerable group, experiencing a mental health penalty of around 17 points compared to heterosexual people who were completely satisfed with their relationships with parents.

#### *Summary of Findings from Part One*

Our frst set of results for this chapter has shown that inequalities between LGB and heterosexual people in the quality of relationships with their parents emerge at the onset of adolescence and persist across adulthood. Meanwhile, the emotional wellbeing of all LGB youth declines over the course of adolescence and this drop is especially pronounced for LGB youth with very low levels of closeness to and support from their parents. Satisfaction with relationships with parents becomes less important for shaping the mental health of LGB people as they move through the life course. However, at all ages, LGB people with the poorest-quality relationships with parents experience the worst mental health and emotional wellbeing of all groups. In part two, we extend our focus across the generations and consider whether the less favourable relationships with parents we observed in part one play any role in shaping LGB people's desires and expectations to become parents themselves.

#### **Part 2: LGB People's Desires and Expectations for Parenthood**

Parenthood is a role of great personal signifcance and a highly valued life goal for many people, irrespective of their sexual orientation. Heterosexual and LGB people report valuing parenthood equally, believing that the associated rewards justify the work and costs involved (Riskind & Patterson, 2010). Regardless of gender composition, couples with parenting desires report the same motivations for wanting to have children (Santona et al., 2021). These include feelings of happiness, affection, fulflment, positive benefts for the entire family, and a wish to live symbolically through one's children after death.

Yet despite placing an equal value on parenthood, studies from countries including Italy, Israel, and the UK have consistently found that lesbian women and gay men report lower parenting desires, intentions and expectations than their heterosexual peers (Baiocco & Laghi, 2013; Gato et al., 2020; Leal et al., 2019; Shenkman, 2012; Tate & Patterson, 2019). For example, an analysis of nationally representative data from the US found that childless heterosexual women and men were almost four times as likely to express parenting desires as childless lesbian women or gay men (Riskind & Tornello, 2017). Meanwhile, the parenting desires of bisexual people did not differ from those of heterosexual people. This may be because bisexual people are more likely to form long-term partnerships with and/or marry a member of the other sex rather than the same sex (Gates, 2015). Their experiences of forming a family may more closely resemble those of heterosexual people than lesbian or gay people as a result.

Sexual orientation has further been found to shape correlations between parenting desires and intentions/expectations. Among those who reported a desire to have children, lesbian women were just as likely as heterosexual women to intend to become a parent (Riskind & Tornello, 2017). However, gay men who desired to have children were signifcantly less likely than heterosexual men to intend to fulfl those desires. In fact, gay men in the US were more than six times as likely as heterosexual men to express that they wanted but did not intend to have children (Riskind & Tornello, 2017). Similarly, in an Israeli sample, more than two-thirds of gay men reported a strong desire to have children (Shenkman, 2012). Yet less than a third of those men wanting to have children believed they had a high chance of becoming a parent in the future.

#### *Barriers to Parenthood*

The lower parenting desires and expectations reported by lesbian and gay people can be largely if not completely attributed to the unique barriers to parenthood that they face. Pathways to parenthood are characterised by more practical complications for same-sex couples than for different-sex couples. For example, same-sex couples must make decisions around method of conception, which member of the couple will be the biological parent and donor anonymity (Perales et al., 2020b). These decisions can be further complicated by the legal framework in which couples fnd themselves. In Australia, for example, commercial surrogacy is illegal. Gay men wishing to conceive a child in this way must therefore either fnd an altruistic surrogate or access commercial surrogacy in another country that allows it, such as the US1 (Perales et al., 2020b).

In addition to these practical barriers, lesbian and gay people face psychosocial barriers to parenthood. Heteronormative ideology positions the nuclear, heterosexual family as the ideal, while lesbian and gay identities are constructed as incompatible with parenthood. In one Australian study, participants read hypothetical vignettes and consistently rated heterosexual parents as more responsible, emotionally stable, competent and nurturing, as well as better role models for their children, than lesbian or gay parents (Morse et al., 2008). These beliefs are reinforced through media representations and interpersonal interactions in community and health care settings (Perales et al., 2020b). Gay men are especially likely to be stigmatised as parents due to the historical confation of homosexuality with paedophilia and the notion that men are naturally less nurturing than women (Perales et al., 2020b). Several international studies have found that gay and lesbian people anticipate more stigma upon becoming parents than heterosexual people, and that this accounts for their lowered desires and expectations to have children (Gato et al., 2020; Leal et al., 2019; Shenkman, 2021). In a heteronormative and homophobic society, many gay and lesbian people assume that their sexual orientation forecloses the possibility of becoming a parent (Hayman et al., 2015; Murphy, 2013).

#### *The Importance of Support from Own Parents*

Another psychosocial barrier to parenthood that gay and lesbian people face is lower expectations of social support upon becoming parents. As we demonstrated in the frst part of this chapter, LGB people report lower support from, closeness to, and satisfaction with their relationships with parents than heterosexual people on average. These disparities are evident across the life course and could contribute to lowered parenting desires and expectations among lesbian and gay people. Irrespective of sexual orientation, people who feel closer to and expect greater support from their parents and other family members are more likely to wish to become parents themselves (Baiocco & Laghi, 2013; Gato et al., 2020). Crucially, support from friends and the LGB community may not always compensate for a lack of support from family when lesbian or gay people become parents. Becoming a parent is still seen as a heteronormative act within segments of the LGB community (Gato et al., 2020). Lesbian women have reported being rejected by friends upon becoming a mother, with one participant in an Australian study stating, "there are a proportion of the lesbian community who are very anti-children" (Hayman et al., 2015, p. 398).

<sup>1</sup>Even then, it is currently illegal for residents of the ACT, NSW and QLD to enter into commercial surrogacy arrangements overseas.

Relationships with parents may shape lesbian and gay people's parenting desires and expectations for reasons other than expectations of support. As we discussed earlier in the chapter, being rejected by parents can lead to the internalisation of homophobic beliefs and feelings of self-hatred in LGB people (Carastathis et al., 2017), and these feelings may contribute to lower parenting desires and expectations. Consistent with this premise, Shenkman (2012) found that sexual orientation self-acceptance was a signifcant, positive predictor of expectations for fatherhood among gay men. Meanwhile, perceived support from family and signifcant others buffered the impact of minority stressors (such as internalised homophobia and felt stigma) on parenting desires and intentions in an Italian sample of lesbian and gay people (Scandurra et al., 2019). This suggests that being accepted and supported by parents may help lesbian and gay people see themselves as worthy of having children and capable of overcoming the practical barriers involved.

#### *Associations Between Parenting Desires, Expectations, and Mental Health*

When social structures interfere with a person's aspirations to have children, there may be negative consequences for their mental health and wellbeing (Gato et al., 2020). Existing evidence supports this proposition. For example, a relationship between expectations for parenthood and depression has been reported in two Israeli studies. In their sample of 183 gay men, Shenkman (2012) found that greater expectations of fatherhood were associated with signifcantly lower depressive symptoms. This fnding was replicated in a sample of gay men, lesbian women, and heterosexual men and women: those with higher expectations for parenthood had lower depressive symptoms and higher average of happiness and life satisfaction (Shenkman & Abramovitch, 2021).

Analysing data from approximately 1500 gay men in the US, Bauermeister (2014) found that the association between aspirations for fatherhood and depressive symptoms was heavily dependent on the legislative context of the state in which men lived. Discriminatory legislation (such as the banning of same-sex marriage, same-sex adoptions or second-parent adoptions) represent barriers to fulflling parenting desires for same-sex couples. Such legislation is also likely to inform gay men's perceptions of how much support they will receive and how much stigma they will face should they become a parent. Consistent with these arguments, Bauermeister (2014) found a negative relationship between aspirations for parenthood and depressive symptoms among gay men living in states without discriminatory legalisation (i.e., higher aspirations = lower depressive symptoms), but a positive relationship between these two constructs among men living in states that had banned same-sex marriage (i.e., higher aspirations = higher depressive symptoms).

#### *New Analyses for Australia*

In this next section of the chapter we present new evidence on the parenting desires and expectations of childless Australians according to their gender and sexual orientation. In addition, we examine the role of satisfaction with relationships with parents in predicting parenting desires and expectations, and the associations between parenting desires, expectations, and mental health. To do this we used data from the HILDA Survey. To be consistent with prior international studies, we only analysed data from respondents who were childless at the time in which the survey was administered.

As mentioned earlier in the chapter, sexual orientation was measured in waves 12 (2012) and 16 (2016) of the HILDA survey. In these waves, participants aged 15–54 years were also asked to pick a number between 0 and 10 to show how they felt about having a child in the future. The response scale ranged from 0 (defnitely don't want children) to 10 (very much like to have children). To create a dichotomous measure of desire to have children, we followed the approach of Shenkman (2012) and assigned a score of 1 (wants children) to those who responded with 7–10, and a score of 0 (ambivalent/doesn't want children) to those who responded 0–6. Participants aged 15–54 years were also asked about their expectations for having a child in the future. The response scale ranged from 0 (very unlikely) to 10 (very likely). To create a dichotomous measure of expectations, we gave a score of 1 (expects to have children) to those who responded with 7–10, and a score of 0 (unsure/doesn't expect to have children) to those who responded 0–6.

Given that people experience different barriers to achieving parenthood according to the intersection of their gender and sexual orientation, we created six groups for analysis: lesbian women (n = 106, 1.2%), gay men (n = 188, 2.1%), bisexual women (n = 160, 1.8%), bisexual men (n = 68, 0.8%), heterosexual women (n = 3557, 40.1%) and heterosexual men (n = 4803, 54.1%). We used the same measures of satisfaction with relationships with parents and mental health (SF-36) as we did in part one of the chapter.

#### *Results for Parenting Desires and Expectations*

Figure 12.5 shows the probability of wanting to have children by gender and sexual orientation, controlling for age. The predicted probability of wanting to have a child was highest for heterosexual women (71%) and heterosexual men (66%), followed by lesbian women (51%), bisexual men (50%), and bisexual women (44%). In our sample, gay men were the least likely to report a desire to have children in the future (26%).

We subsequently examined the predicted probability of expecting to have a child for those who reported wanting a child, looking across gender and sexual orientation and controlling for age. Results are displayed in Fig. 12.6 below. Among

**Fig. 12.5** Desire to have children among childless Australian adults 15–54 years. Note: Predicted probability of wanting to have a child in the future based on logistic regression model with random effects, controlling for age. (Source: Household, Income and Labour Dynamics in Australia Survey. Waves 12(2012) and 16(2016))

Australians with a desire to become a parent, gay men were the least likely to expect to fulfll those aspirations (60%), which is consistent with the international literature (see e.g., Riskind & Tornello, 2017). Meanwhile, heterosexual men, bisexual men and heterosexual women were the most likely to expect to fulfll their parenting aspirations (93% each). Most lesbian and bisexual women who wanted children also expected to have them in the future (85% and 86% respectively).

#### *Results for the Role of Relationships with Own Parents*

Next, we looked at the association between satisfaction with relationships with parents and parenting desires. We found that the more satisfed a person was with their relationships with their parents, the more likely they were to report wanting to have children in the future. The probability of wanting to have a child for those who were completely dissatisfed with their relationships with their parents was 46.1%, compared to 73.7% for those who were completely satisfed. This association was the same regardless of gender and sexual orientation. As such, a more positive relationship with parents was equally important for shaping the parenting desires of bisexual and heterosexual people as for lesbian and gay people.

We then examined associations between satisfaction with relationships with parents and parenting expectations among those who reported a desire to have children

**Fig. 12.6** Expectation to have children among childless Australian adults 15–54 years with parenting desires.

Note: Predicted probability of expecting to have a child in the future among those who want to have one based on logistic regression model with random effects, controlling for age. (Source: Household, Income and Labour Dynamics in Australia Survey. Waves 12(2012) and 16(2016))

in the future. In this instance, we found that there was a signifcant interaction between gender and sexual orientation. Specifcally, we found that a positive relationship with parents was substantially more important for shaping the parenting expectations of gay men than any of the other groups. For lesbian women, heterosexual men and women, and bisexual men and women, parenting expectations were high (>70%) regardless of how satisfed they were with their relationships with their parents. For gay men, however, when satisfaction with parents was low, parenting expectations were also low (10%). As satisfaction with parents increased parenting expectations increased, such that gay men who reported being completely satisfed with their relationships with parents had a high probability (>70%) of expecting to have children in the future. These fndings are visualised in Fig. 12.7.

#### *Results for Associations Between Aspiration-Expectation Mismatches and Mental Health*

In our fnal analysis for this chapter, we examined how the match between an individual's parenting desires and expectations related to their mental health. We established earlier in the chapter that satisfaction with relationships with parents is associated with all of these variables. We therefore accounted for satisfaction with

**Fig. 12.7** Parenting expectations according to satisfaction with relationships with own parents among those with a desire to have children.

Note: Predicted probability of expecting to have a child in the future among those who want to have one based on logistic regression model with random effects, controlling for age. (Source: Household, Income and Labour Dynamics in Australia Survey. Waves 12(2012) and 16(2016))

relationships with parents in our analysis to identify if thwarted parenting desires contributed to mental health disadvantage over and above less satisfactory relationships with parents. We found mental health differences according to the match/ mismatch between parenting desires and expectations that were consistent across gender and sexual orientation groups. The highest levels of mental health were reported by people who both desired and expected to have children in the future (average predicted score of 73.5 out of 100), while the lowest levels of mental health were reported by people who desired but did not expect to have children in the future (average predicted score of 68.6 out of 100). The mental health of people without parenting desires fell in between (Fig. 12.8).

#### *Summary of Findings from Part Two*

In our second set of analyses for this chapter we have shown that childless LGB people in Australia are less likely to report a desire to have children in the future compared to their heterosexual counterparts, with gay men especially unlikely to report wanting children. Furthermore, gay men who do report wanting children are signifcantly less likely to expect to fulfll those aspirations than any other group. We found that this mismatch between desires and expectations is associated with

**Fig. 12.8** Mental health according to the match between parenting desires and expectations. Note: Predicted mental health score based on logistic regression model with random effects, controlling for age and satisfaction with relationships with parents. (Source: Household, Income and Labour Dynamics in Australia Survey. Waves 12(2012) and 16(2016))

diminished mental health compared to people who expect to fulfll their parenting desires. Last, we showed that disadvantageous family dynamics can accumulate across generations: people are less likely to express parenting desires the more dissatisfed they are with relationships with their own parents. In addition, gay men who wish to have a child are signifcantly less likely to expect to fulfll their desires the less satisfed they are with their relationships with their parents.

#### **Discussion**

In this chapter, we drew on the life course principles of trajectories, turning points, and linked lives to examine the role of family dynamics in producing disadvantage in the lives of LGB people in Australia. In part one, we presented new Australian evidence on the quality and signifcance of relationships between LGB people and their parents across the life course, adding to fndings reported in earlier chapters on parental engagement in education (Chap. 6) and parental support during emerging adulthood (Chap. 8) for broader populations. Prior research has focused almost exclusively on adolescence, which is justifed: adolescence is a critical time for the development of important attributes and life skills, and it is when trajectories in education, employment, health, wealth and family begin to take shape. Furthermore, it is when most LGB people begin disclosing their non-heterosexuality to others. Thus, both the risks and consequences of parental rejection are especially acute during this life stage. However, by ignoring the rest of the life course, we are missing important pieces of the puzzle (Reczek, 2020). We need to better understand when disparities in the parent-child relationship emerge, how they track over time, and their role in producing disadvantage at different stages of the life course. The fndings we have presented here constitute some initial steps in that direction.

The good news is that, at all ages, the majority of LGB people in Australia report a positive relationship with their parents. The observed differences between the LGB and heterosexual groups were modest, which is consistent with recent international literature (Fischer & Kalmijn, 2021; Hank & Salzburger, 2015). It is also consistent with documented increases in the proportions of LGB youth in Australia who disclose their sexuality to and receive support from their parents (Hillier et al., 2010). As Australian society has become less heterosexist and homophobic, parents are less likely to reject their LGB children. In our view, this is a powerful example of how changing social structures can change people's lives.

Despite this progress, a minority of LGB people still experience negative relationships with their parents. These disparities emerged early, with a turning point evident at age 11, and they persisted across the life course. We must therefore view childhood as a critical time for intervention if we are to prevent these disadvantageous family dynamics from developing. We agree with the conclusions of other Australian scholars in this feld that parents must be provided with both education and support as children approach adolescence, so that they are prepared in the event that their child discloses a non-heterosexual orientation (Hillier et al., 2010). This could be done through the inclusion of information on sexuality diversity in mainstream material on adolescent health that parents are likely to access, as well as dedicated education programs for both parents and children. Most sexuality and relationship education provided in schools fails to meet the needs of LGB youth (Hillier et al., 2010; Robinson et al., 2014). Improving these programs and expanding their scope so that they reach parents, too, should be a priority for policymakers.

As we looked further into the life course, what we saw was arguably a story of resilience. While average satisfaction with the parent-child relationship remained fairly constant for LGB people across adulthood, the mental health impacts of feeling dissatisfed with this relationship declined with age. This is consistent with the accounts of LGB Australians who—with time, strength and effort—found ways to foster their wellbeing in the face of family rejection (Carastathis et al., 2017). They did this by drawing on friends, colleagues and other members of the LGB community—often referred to as "families of choice"—for the acceptance, emotional support and sense of belonging missing in their families of origin. In particular, connecting with other sexuality diverse people helped normalise their experiences and challenge any homophobic views that they had internalised, thereby growing their confdence, optimism and self-acceptance (Carastathis et al., 2017).

In the second part of this chapter, we presented new Australian evidence on parenting desires and expectations across gender and sexual orientation groups. Consistent with prior international evidence (e.g., Riskind & Tornello, 2017), we found that gay men and lesbian women were less likely to express parenting desires than heterosexual men and women. This is unsurprising given the unique barriers that lesbian and gay people face in their journey to become parents. It also aligns with qualitative Australian evidence indicating that some gay and lesbian people assume that their sexual identity is incompatible with becoming a parent and do not consider the possibility of having children as a result (Hayman et al., 2015; Murphy, 2013). Our fndings support the view that heteronormative social structures that privilege heterosexuality (e.g., through legislation and the design and provision of family services) act to suppress parenting desires among some lesbian and gay people. The role of existing legislation in facilitating versus constraining access to parenthood—for instance, via adoption and surrogacy—should be revisited in light of our fndings (see also Kazyak et al., 2018).

Interestingly, bisexual men and women were also less likely to express parenting desires than heterosexual men and women in our sample, a fnding that diverges from recent US evidence (Riskind & Tornello, 2017). Given that bisexual people are much more likely to be partnered with a member of the other sex than the same sex, we anticipated that their expectations regarding parenthood would more closely resemble those of heterosexual people. However, one factor that we had not considered was relationship quality. Bisexual men and women, including those in othersex couples, have been found to report signifcantly lower levels of relationship quality than heterosexual men and women (Perales & Baxter, 2018). A poor quality relationship with one's partner may deter some bisexual people from pursuing parenthood, a proposition that future research should investigate.

The vast majority of people in our sample who wanted to have children in the future were confdent of achieving their aspirations, with one exception: gay men. Gay men were more likely to report a mismatch between their parenting desires and expectations, with two-ffths of those who wanted to have children not expecting to become a parent in the future. Given the strong personal signifcance that parenthood holds, thwarted parenting desires are an important form of disadvantage in their own right. In addition, we found that they are associated with a signifcant penalty with regards to mental health. While statistically signifcant, this difference was arguably modest. However, meaningful disparities in mental health can result from the accumulation of multiple, small blows over time, and thwarted parenting desires are just one disadvantage that LGB people (in particular, gay men) are at risk of experiencing across the course of their lives.

Our analyses also shone light on how disadvantageous family dynamics in the lives of LGB people can accumulate over time. We established in part one of this chapter that, compared to heterosexual people, LGB people feel less close to and less supported by their parents across adolescence, and less satisfed with their relationships with their parents throughout adulthood. In part two of this chapter we found that the less satisfed people were with their relationships with their parents, the less likely they were to desire to have children of their own. Furthermore, gay men who did express parenting desires were substantially less likely to expect to fulfll those aspirations if satisfaction with their relationships with their own parents was low. We proposed that internalised homophobia, a more negative self-concept, higher expected stigma and lower expected social support could all be mechanisms underlying this association. This is something that future studies could examine.

#### *The Way Forward*

The analyses that we have presented in this chapter represent just the frst steps towards developing a better understanding of how family dynamics accumulate and produce disadvantage in the lives of LGB Australians. They provide a broad, descriptive view—albeit with the advantage of coming from large, nationally representative samples—and further research is needed to build on the foundations we have laid here. One fruitful avenue for further inquiry would be to examine individual (as opposed to average) life course trajectories of parent support and mental health among LGB people, to identify factors that promote resilience versus vulnerability. Further empirical attention could also be directed towards important transitions in the lives of LGB people, and how this effects their relationships with their parents. For example, preliminary research suggests that getting married or becoming a parent can bring some LGB people closer to their parents (Power et al., 2012; Reczek, 2020). In addition, the role of relationships between LGB people and other family members, such as grandparents and siblings (see Perales & Plage, 2020), in shaping outcomes remains something of a blind spot in this literature. Most crucially, we must ascertain how experiences of parental and family rejection during childhood and adolescence impact LGB youth's transitions into adulthood and their trajectories in other life domains (e.g., education, employment). There may always be a small proportion of LGB people who are rejected by their families. Having a solid grasp of how this impacts individual lives is the frst step towards creating effective policy interventions to remedy disadvantage.

#### *Conclusion*

In this chapter, we have demonstrated that LGB people are disadvantaged not only by the actions of strangers who exclude them through inequitable legislation, make angry phone calls to talk back radio, or glare at them as they walk down the street, but also through the actions of their own parents. This is a clear example of the life course principle of linked lives in action: social structures have the power to shape our most important, personal relationships and, through these, our mental health and wellbeing. If we are to achieve equality of outcomes irrespective of sexual orientation, we must not only support and educate parents, but continue to challenge and dismantle the heteronormative social structures that produce this disadvantage in the frst place.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 13 Ageing and Loneliness: A Life Course and Cumulative Disadvantage Approach**

**Jack Lam, Catherine Dickson, and Janeen Baxter**

Loneliness is the discrepancy between desired and actual social relationships, both in terms of quality and quantity (Cacioppo et al., 2015). Loneliness has received much attention in recent years, and has been especially highlighted by the COVID-19 pandemic (Luchetti et al., 2020; van Tilburg et al., 2020). Older adults are one population of concern for loneliness, as age and loneliness have a U-shaped relationship, with young adults and young-old adults (65–85 years old) experiencing the highest prevalence of loneliness (Nicolaisen & Thorsen, 2017; Pinquart & Sorensen, 2001). An Australian study found that loneliness prevalence among Australian older adults closely resembles that found overseas in Europe and Great Britain, with 7% of older Australian adults in their Western Australian sample severely lonely, and a further 31.5% sometimes lonely (Steed et al., 2007). A more recent study using data from the nationally-representative *Household, Income and Labour Dynamics in Australia* (HILDA) survey reports that 16.1% of older adults (65+ years) reported loneliness, agreeing with a single-item statement 'I often feel very lonely' (Kung et al., 2021). Loneliness is signifcant through its associations with a range of health and wellbeing outcomes (Cohen-Mansfeld et al., 2016) including higher all-cause mortality (Holt-Lunstad et al., 2015; Leigh-Hunt et al., 2017; Luo et al., 2012; Rico-Uribe et al., 2018) and dementia (Rafnsson et al., 2017).

This chapter draws on interview data with older adults to examine how experiences throughout the life course may explain variation in later-life loneliness. This

J. Lam (\*) · J. Baxter

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia

e-mail: j.lam@uq.edu.au; j.baxter@uq.edu.au

C. Dickson

Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia e-mail: catherine.dickson@uq.edu.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_13

shows that loneliness, like other social or health outcomes, is not only the result of current circumstances and conditions, but may be traced back to earlier life events. Thus, the reduction of later-life loneliness requires not just examining circumstances at the later life stage but may be supported by intervention earlier in the life course. Literature on the correlates of loneliness to date has focused predominantly on contemporaneous factors, such as partner status, living arrangements, health and socioeconomic status (Cohen-Mansfeld et al., 2016; De Koning et al., 2017; Pinquart & Sorensen, 2001). Some studies also pay attention to how experiences of life-changing events, such as widowhood or divorce, may also lead to subsequent changes in loneliness (Dahlberg et al., 2015; Peters & Liefbroer, 1997). Lim et al. (2020) describe a conceptual model in which entry into loneliness depends on 'triggers'; that is, life events associated with a change in social identity.

While there is robust evidence of the signifcance of these factors, unexplained variation remains which may be better explained by a life course perspective. This is because even in longitudinal studies of loneliness, researchers may often overlook the importance of stability of family and work life conditions and circumstances, given their inherent focus on change, or the observation of varied events (Dahlberg et al., 2015; Dykstra et al., 2005; Victor & Bowling, 2012). The timeframes focussed upon in such research, ranging from 5 years (Newall et al., 2014), to 20 years (Wenger & Burholt, 2004), and 28 years (Aartsen & Jylhä, 2011), as well as a focus on changes observed in later life, may also overlook factors or events that may be of signifcance up to that point in the adult life course.

Taking a qualitative, inductive approach, we highlight important explanatory factors of later-life loneliness. Our fndings point to four themes that are signifcant for understanding later-life loneliness, namely: (1) history of familial support; (2) formation and maintenance of social networks; (3) relocation and migration; and (4) widowhood and separation. These highlight the importance of continuity of social ties and the impact of change on later-life loneliness. In the next section, we start to build the case for examining earlier life infuence. We draw on the life course perspective and cumulative disadvantage theory to underscore the value of taking a broader view of the adult life course to understand later-life loneliness.

#### **Literature Review**

A body of research has investigated the prevalence, correlates, and predictors of loneliness among older adults (65+ years old). The literature describes a multifaceted relationship between old age and loneliness that is mediated by changes in physical, social and psychological circumstances that occur during aging (Cohen-Mansfeld et al., 2016). Loneliness is an important issue and has been found to relate to a number of socio-demographic factors, including partnership status (Boger & Huxhold, 2020), health (Burholt & Scharf, 2014), educational attainment (Bishop & Martin, 2007), and economic status (Hawkley et al., 2008).

Being partnered is a protective factor against loneliness, although the quality of the relationship is a signifcant moderator. Having a supportive partner is associated with lower levels of loneliness while having a strained partner relationship is associated with higher levels of loneliness (Shiovitz-Ezra & Leitsch, 2010). Studies of older adults' social networks consistently fnd that having a diversity of social ties is related to the lowest levels of loneliness (Litwin & Shiovitz-Ezra, 2011). Better health, higher income and wealth as well as higher educational attainment are also related to lower levels of loneliness (Cohen-Mansfeld et al., 2016; Savikko et al., 2005). Higher income and educational attainment may afford an individual the resources to access and participate in more social activities, offsetting the risk of loneliness (von Soest et al., 2018).

Prospective studies have also recently begun to investigate how the experience of specifc events or life stage transitions may relate to changes in loneliness, particularly those typical in later life. Changes in marital status, health, social networks or living arrangements are associated with increased loneliness, because they frequently lead to losses to independence, mobility and social participation (Victor & Bowling, 2012). The loss of a supportive partner through widowhood for instance is strongly associated with increased loneliness (Dahlberg et al., 2015). Declines in health status, such as in functional health, vision and hearing, are associated with increased loneliness, possibly because they create barriers to social interactions (Savikko et al., 2005; Dykstra et al., 2005). Negative changes may have a stronger impact than positive ones, with health declines associated with a greater magnitude of increase in loneliness when compared with improvements in health and their association with decreases in loneliness (Dykstra et al., 2005). Overall, partner status, relationship quality, health, and economic, social and educational resources are all key factors associated with loneliness.

#### *The Life Course Perspective and Mechanisms Linking Life Course Disadvantage and Later-Life Loneliness*

The life course perspective suggests that the trajectory of a person's life is shaped by the sequence and timing of life events. These are infuenced by their historical and geographical location, and importantly for the study of loneliness, also by their own agency, and relationships with others—'linked lives' across their life course (Elder & Giele, 2009). Under this perspective, it can be expected that the cumulative effect of life events over the life course, such as childhood positive and negative experiences, partner status, parental status, bereavement, and migration experiences, infuence the experience of later-life loneliness in ways that contribute to, but may be distinct from, the current situation of the older adult.

Cumulative disadvantage theory states that negative events or conditions can compound over the life course. We understand loneliness as a phenomenon in relation to the existence and quality of a social network, fltered by an individual's perception. Thus, cumulative disadvantage is expected to be relevant to loneliness in two ways; through its infuence in shaping a supportive social network across the life course, and through the development of psychological skills for coping with loneliness-provoking social contexts, such as a sense of agency and self-effcacy (Fry & Debats, 2002; Gerino et al., 2017). Early inequalities may not only limit the direct opportunities for development of appropriate social and psychological skills, but also infuence other life domains. Social advantage increases exposure to social opportunities, such as stable living and working arrangements which maintain stable social relationships, or suffcient resources and good health to enable high levels of social participation. These circumstances may then afford people further opportunities for social support and connection. Conversely, disadvantage in one of these life domains may increase the risks of psychological and fnancial stresses, placing the individual at greater risk of isolation from their social network. Examining loneliness among older adults as a function of their life course experiences provides the beneft of describing the mechanisms linking past events and present circumstances.

Longitudinal and retrospective research linking a range of childhood and adult experiences with later-life loneliness adds credence to investigating loneliness through cumulative disadvantage. These studies tend to focus upon diminished later life social resources as the key domain of disadvantage relevant to loneliness, suggesting that disadvantage compounds as a result of either maladaptive social skills (Ejlskov et al., 2020; Hensley et al., 2012), or social censure or confict (Zoutewelle-Terovan & Liefbroer, 2018; Wu & Penning, 2015). Negative childhood experiences, including poverty or lack of positive engagement with parents, are associated with loneliness in later life (Merz & de Jong Gierveld, 2016; Kamiya et al., 2014). These retrospective studies suggest, as a mechanism, developmental defciencies which persist into adulthood, such as a lack of parental investment in childhood social development which affects the social skills at their disposal (Case et al., 2005) or the development of maladaptive attachment behaviours (Andersson & Stevens, 1993). These mechanisms of impaired social abilities originating in childhood are not overcome, but continue to affect the individual throughout their life, although these studies do not investigate to what extent they affect subsequent life events. The extant studies which have considered social or other adversities that occur during adulthood (Ejlskov et al., 2020; Peters & Liefbroer, 1997) are limited because they are most interested, not in the accumulation of disadvantages, but in loneliness effects which remain for older adults independent of their current social situation. They fnd that previous experiences of damaged or lost relationships are associated with increased loneliness independent of current social relationships. This suggests that continuity in relationships is protective against loneliness. A cumulative disadvantage perspective highlights that inequalities like low income, poor education, or poor health may directly infuence available social opportunities across the life course, thus signifcantly differentiating individuals' risk of loneliness by the time that they reach old age.

Divergence from the normative script of life events, such as delays or non-entry into partner relationships and parenthood, is also related to later-life loneliness. This confrms the relevance of applying a life course perspective to the study of loneliness, with its interest in the sequence and timing of events. The suggested mechanisms consist of social censure or lack of support from surrounding social networks, or life-long lack of a common source of social or material support (Zoutewelle-Terovan & Liefbroer, 2018; Van Humbeek et al., 2016). However, the loneliness associated with singlehood may be moderated by the role of choice or constraint, with those who never married by choice expressing satisfaction in their later life, while those who were never-married by constraints—such as caring responsibilities or economic factors—experience loneliness and discontent with their unmarried status (Timonen & Doyle, 2014). Other proposed moderating factors are the role of stability throughout the life course (Dykstra & de Jong Gierveld, 2004), or the practice of 'anticipatory socialization' to maintain the personal and social resources required for wellbeing (Koropeckyj-Cox, 1998). This suggests that where single people have exercised agency over their own social circumstances, they are less susceptible to loneliness.

These fndings regarding human agency introduce an important counterpoint for avoiding a determinist interpretation of the accumulation of disadvantage. The life course perspective affrms that within their constrained circumstances, people plan and make choices (Elder & Giele, 2009). Resilience can be understood as a deliberate, purposive response based on recognition of one's structural circumstances (Schafer et al., 2009). In relation to loneliness, where the person perceives themselves to be in an adverse social situation, resilience consists of being able to identify corrective action to counter the adversity, and being able to activate appropriate resources to do so (Schafer et al., 2009). These may be practical, social, or psychological resources, as older adults seek to make changes within their activities or social networks or reframe their experiences within the context of their life, in order to reduce their loneliness. Resilience is not a magic bullet, because it depends upon the ability to mobilise resources. Cumulative disadvantage theory highlights that events which diminish these resources can be expected to have ongoing, compounding effects exposing the individual to greater risks of social losses with fewer opportunities for social gains. As a result, we may expect older adults to recall a range of adulthood factors that play a role in their understanding of later-life loneliness.

#### **Data and Methods**

This chapter draws on data collected as part of a broader research project "Understanding Daily Activities in Later Life" which aimed to provide up-to-date evidence on the experience of daily activities and loneliness for a group of older Australians. It has a specifc focus on understanding the patterns and rhythms of daily life and how these relate to older Australians' wellbeing. It was designed and conducted with the assistance of a home care organisation in Southeast Queensland, Australia. This mixed-methods project comprised an initial survey that was conducted to collect data on health, social interactions, family structure and personal relationships. A total of 182 older individuals (aged 65 years plus) returned the selfreport questionnaire. From this sample, a subset of 50 individuals who represented a range of socio-demographic characteristics was then selectively contacted for follow-up one-on-one semi-structured interviews. Ethics approval was obtained from the University of Queensland.

Interviews took place at the location of choosing of respondents. The majority took place in home residences, with the exceptions being in the common area of 7 retirement villages or the lobby of apartment buildings. The interviews ranged from around 30 minutes to 2 hours. During the interview, we attempted to get a sense of the daily lives of the respondents, posing statements such as "Tell me about your day yesterday," "Tell me about your week this past week" and "Tell me about your relationships and support". Given that the interviews were semi-structured, respondents also at times discussed their life histories, providing a broader overview of how they understood their current circumstances in relation to past events. They likely felt further encouraged to refect in this way by the questions "Have you experienced any major life events recently? Has that changed anything in your life?". For example, in one interview, a participant described his whole family life history, from the time of birth while describing this alongside a photo album of photos since his childhood, documenting various events along the way. Although this is one extreme, as the respondent was engaged in a project around his family ancestry, other participants also brought up discussions around past life events.

The mean age of the respondents was 82, and the overwhelming majority were women. About half were widowed (50%), a quarter married (28%), and the remainder divorced, separated, or never married. The average respondent reported being 'reasonably comfortable' in terms of their fnancial situation and reported being somewhere between 'good' and 'moderate' in terms of their health.

Interviews were recorded with the consent of the participants and audio fles were transcribed and uploaded onto a qualitative software, NVivo, which we drew on to further analyse the data. We undertook inductive coding to explore themes that might arise from the interview data. The lead author and two undergraduate research assistants coded several of the same interview transcripts and met at the initial stages to establish inter-rater reliability. As the analysis progressed, the research team met and established ffteen distinct themes. For the purposes of the current chapter, we focused on data with mentions of past events and family life history to understand how these are signifcant factors for the respondents' current circumstances. While respondents often focussed discussion on present-day activities and relationships, such as around care arrangements, discussion of the past was quite common around themes such as family, friendships, support network, and community. From these, our data highlights for older adults in our sample the importance of different factors that relate to their loneliness.

#### **Results**

Our fndings highlight the importance of social network and ties over the life course for understanding later-life loneliness. While prior research examining the impact of life events on later-life loneliness highlights the role of change (via an observed event), it has paid less attention to the importance of stability and history of family social support. Our data however highlights this as a crucial factor. Not only is stability of social ties highlighted, but the role of human agency is instrumental in shoring up family support as well as in creating a support network. Consistent with previous research, our fndings also showed that the experience of events such as migration and relationship dissolution impact on later-life loneliness through the disruption of ties. While existing studies have primarily focused on the occurrence of such events in later life, for example the observation of events at age 50 or older as is common in most gerontological studies, our study highlights how events earlier in the life course are also of signifcance and have longer term implications. Our fndings show the importance of four themes: history of familial support, formation of social networks, relocation and migration, and widowhood and separation.

#### *History of Familial Support*

Support from the family unit throughout the life course was linked to decreased later life loneliness. Many participants associated the support, or lack thereof, of their family across the life course with lesser, or greater, feelings of loneliness in later life. Several participants identifed how they developed a familial culture of support across their life history. One participant detailed the culture of open and honest communication that had developed in their family from a history of encouraging conversation with one another, "We've always believed in family discussion and I think that's a lot of it where they've picked up things from me" (Female, Aged 86, Widowed, Living Alone). The participant detailed how this led to a "very, very close" family relationship, which did not leave her feeling at all isolated. Other participants echoed this sentiment, beneftting socially from close relationships with their family members.

Another participant indicated that their family had become geographically distant over the years, and throughout their family history there had been few occasions where the family had come together as adults. The participant stated, "It just doesn't kind of work that way in our family." (Female, Aged 88, Married, Living with Spouse). Yet, after a traumatic health event that led to the hospitalisation of her and her spouse, the family converged from all over the world and were immediately present to assist. This family event shifted her perception of her connection with her family. She stated, "It taught me about my family and the care and the fact that each one contributed some of their own specifc gifts to me at that time, their specifc caring gifts." This highlights both that across her family history she had cultivated

social ties that could be called upon in times of need, and that this life event acted as a catalyst for her family network to connect with her differently, changing her perception of the social support available to her. A widowed woman aged 88 and living alone, also highlighted the social support she received from close relationships with all of her children, especially in the aftermath of her husband and brother passing away. Another widowed woman (77, living alone) described a sequence of events which had left her a self-described "hermit". Like the others, after the death of her spouse, her daughter had provided her with a sense of available support, as she also dealt with health issues that limited her ability to leave the home for social engagements. It was only after her daughter's subsequent move to another state that she felt a lack of support, describing herself as "a ship in the ocean; drifting", demonstrating the compounding infuence of successive life events.

Other participants identifed alternate positive outcomes from vastly different family histories. One female participant, aged 75, talked of her life as a single person, having never been married. She claimed this as being a large mitigating factor in avoiding loneliness in later life. She attributed loneliness to the isolation of widowhood, and stated, "They haven't got [the same skills as me] because I've been by myself and I'm single." (Female, Aged 75, Never Married, Living Alone). The skills developed across the life course as a single person were identifed as steeling her against events in later life that would, speculatively, cause loneliness in others. Another woman, aged 75 and living with her spouse, credited her upbringing in a rural setting with an ability to "just get on with it" in older age.

Some participants tended to attribute loneliness to distance from family, both emotionally and physically. A participant described a series of accusations that had been levelled at her by her daughter. This caused a discrepancy between her ideal relationship with her daughter, and her actual relationship, "It's not as warm as it used to be, or as warm as I'd like it to be." (Female, Aged 82, Married, Living with Spouse). Another participant, aged 77, widowed and living alone, spoke of the physical distance from her cousins. After the death of her parents early in life, she was raised by her aunt, and these cousins were essentially adoptive sisters. However, in later life, one of these cousins was situated in Sydney and the other had recently moved into a nursing home. The participant implied that this geographical isolation from her closest family heightened her feelings of loneliness.

There were also those that accounted for stronger feelings of loneliness and isolation through descriptions of their early life history. A 100 year-old lady, widowed and living alone, spoke of the deaths of her parents and sister. With these deaths, her relationship with her nieces soured and left her isolated from family members.

#### *Formation of Social Networks*

A second theme of the fndings was the role of the formation of social networks. In accordance with the literature, the formation and maintenance of social networks across the life course proved to be a strong theme through the project's interviews.

There were a variety of perspectives provided as to how these networks developed, and how they are best maintained in later life.

Several participants discussed forming and maintaining social networks through formal social, hobby or activity groups. A female participant, aged 91, widowed and living alone, spoke of the importance of cultivating a strong social network early in the life course and carrying it into old age. The woman anecdotally evidenced this with her craft, church, and tennis groups that she had regularly attended for many years. She stated, "I think it's something you need organised before you get retired. Just think about it when you're young … Get into something, because it does help immensely."

There were other accounts of the importance of formal socialisation. One woman, aged 76 and widowed, spoke of the social benefts she had received from a book club that she had been a part of for the best part of a decade. Another woman, aged 88 and married, continued to engage with choirs and craft groups into later life. She reported a beneft from being able to utilise skills and interests developed earlier in the life course. The participant advocated, much like others, for older people to "make an effort to join a group" in order to curb isolation.

Similarly, another participant, a 79 year old male living with his spouse, encouraged older people not to be forced into being a part of programs they have no interest in, but to, instead, maintain connection with people they enjoy the company of. His perspective was grounded in a "men's group" that he had formed with some acquaintances earlier in life, as well as networks through church and exercise. The men's group was explicitly identifed as being a "mutual help society", which provided an important source of social support. He attested to the social power of being a part of formal social groups, stating, "Get into the groups … which allow you to understand other people like you." There is a clear consensus through all of these accounts that there is a social beneft to seeking out formal networks of others that align with one's interests or hobbies.

There was another identifable theme of building networks through local community. One widowed woman, aged 76 and living alone, spoke of the friendships she had maintained in the tight area around where her children had gone to school. Another participant spoke of the social value brought to her by neighbours she had when her children were young. It is clear that the community situation had played an important role in the formation of social networks across the life course. When discussing the current structure of her social networks, one female participant stated, "I've worked with either in nursing or we've worked in the same organisation. So, we have been friends for many years. I've been [in this house] since 1966." (Female, Aged 80, Divorced, Living Alone). This outlines the importance of life course consistency in maintaining a social network.

The conclusion implied or stated by many of these individuals is that a diverse and active social network, vital for the mitigation of loneliness in later life, is a product of developing connections across the life course. A female participant succinctly summarised this point in answering whether her current friend group was more new or old friends, "It's over a lifetime really. You make new friends and then they sort of become old friends … It's a matter of coordinating everything. That's

more diffcult than making new friends."(Female, Aged 83, Widowed, Living Alone). She makes it clear that developing new friends across the life course ensures a pool of established social support in later life, but made the point that maintaining these networks is more complex than establishing them. Similarly, a 66-year old female participant describes, "around the 90s I met a group of friends who I've maintained… I know other people, but I tend to like just seeing good friends occasionally." This reliance on the maintenance of longstanding connections for close connections in older adulthood highlights the importance of stability and the risk posed by late-life social network disruptions.

There were others, however, that were less optimistic about maintaining and developing social connection in later life. Some participants indicated that a breakdown of socialisation and growth of isolation were inevitable in older age. An example of this comes from a woman, aged 81, widowed and living alone, who detailed how her social group had broken down after her husband passed away and other members began to move into retirement living. These events, widowhood and retirement living, are closely linked to the later life course, leading to increased social isolation in later life. Another participant outlined the diffculty faced by already isolated older people, "Sometimes if they haven't got a network of friends when they've been ft and healthy, then it's very hard to get a network going when you're not well and not going outside the house. So I think it starts a lot earlier." (Female, Aged 83, Widowed, Living Alone).

#### *Relocation and Migration*

We also found that migration was a key event that affected feelings of loneliness. There were frequent suggestions from participants that relocation, both international and in-country, as an event from any point across the life course, played a large part in increased isolation and, consequently, increased feelings of loneliness in later life. For instance, one widowed woman, aged 92 and living alone, detailed losing social connections after moving to South-East Queensland from elsewhere in Australia. She stated, "We used to live in Sydney. Once you move, you lose contact with friends." She discussed the adverse effects of that relocation in relation to the impacts of the death of close friends in Queensland. This suggests that she felt that she had decreased social support, due to migration, in those times of grieving. Another participant detailed a near-parallel story. She stated that friendships formed through church groups and raising children had "kind of dwindled" after moving to the city from a relatively nearby town (Female, Aged 81, Widowed, Living Alone). One other woman, aged 86, widowed and living alone, spoke of leaving wellestablished social networks in New Zealand and in New South Wales to move closer to family in Queensland. She spoke about not being able to click with similar social groups as she had earlier in life, and stated, "I don't think I have any friends. I had a lot of friends." The timing of relocation during the life span was signifcant to participants, with another participant articulating, "When you move to someplace when you're older you don't have the same friendships really." Conversely, other participants highlighted that residing in the same area for 60+ years had supported long-term friendships (Female, 80 Years Old, Divorced, Living Alone), with one woman noting that 12 friends came to her birthday party, expressing pride in these long-standing friendships (Female, Aged 74, Divorced, Living Alone). Although they came from "up the coast, down the coast" to attend, it is clear that her longterm friendships were supported by her network remaining within travelling distance over many decades.

There was a recurring theme of people relocating in order to be geographically closer to their children. In these cases, there were mixed accounts of people feeling separated from their developed networks while also feeling increased social support from family. One female participant, aged 83 and widowed, spoke explicitly of experiencing this. The relocation, reportedly, led to conficting experiences of losing easy access to her social networks developed through the life course, but having increased social support available to her from her family. Others discussed the infuence family had on this decision, considering the opinion of their family members more important than any material or social connections with long-term friends (Female, Aged 90, Widowed, Living Alone). One 86-year old widow identifed her daughter's move overseas as a "trigger" for her own move to Brisbane, to be near other family members (Female, Aged 86, Widowed, Living Alone).

There was also evidence of later-life impacts from immigration and relocation earlier in the life course, though there was less indication that this impacted feelings of loneliness than for later-life relocation. A woman, aged 85 and widowed, spoke of the experience of moving throughout Africa and developing social networks with her husband. She implied that, while possible, maintaining any frm handle on these connections became diffcult with relocation to Australia and then to South-East Queensland in later life. Another participant spoke of feelings of loneliness and associated them with an inability to develop social networks over the life course due to a transient life of constant relocation. This man, aged 79, separated and living with his child, talked of moving all over Australia for work throughout his life. He identifed some social connections formed in later life, but he expressed that his only long-term social support fgure, his wife, had been lost with an Alzheimer's diagnosis and subsequent move to a support facility. Two further participants talked about their immigration experience earlier in life, from Germany (Male, Aged 87, Married, Living with Spouse) and England (Female, Aged 86, Widowed, Living Alone). While they showed some signs of connection to their respective places of origin, there was no clear indication this impacted on their feelings of loneliness.

A fnal participant associated relocation with a decrease in loneliness. This was an outlying case that incorporated a variety of factors including re-partnership and a move to be closer to his partner's family. His outlook on this major shift was that any people of importance would work to stay in touch, "We shifted to start a new life, and if they want you, they'll come after you, and you've got to go with that." (Male, Aged 83, Married, Living with Spouse). With this framing, he assigns himself a non-active role in the maintenance of social ties. This attitude may be protective against loneliness because reduced connections are thus not understood as an avoidable lack or internalised as the result of a personal failing. However, it is also possible that he had simply developed a suffciently supportive social network post-relocation.

#### *Widowhood and Separation*

Participants revealed that a history which involved a separation event could drastically affect social connection and feelings of isolation in later life, in a variety of ways. Several participants attributed feelings of loneliness to the deaths of their signifcant others. These feelings were commonly associated with the deep connection that forms between spouses over the life course. Participants highlighted the connection formed in 71 years of marriage (Female, Aged 95, Widowed, Living Alone) and, when asked if there were any wishes that they had, expressed the desire to have a husband back (Female, Aged 82, Widowed, Living Alone). One participant succinctly summarised this sentiment, stating, "You need someone who understands you and knows your weak points and your good points and your… everything." (Female, Aged 92, Widowed, Living Alone).

There were implications of widowhood on broader social networks. An 81-yearold, widowed woman who lived alone, noted how, since the death of her husband, her connections with her social group had gradually deteriorated. She stated, "[Connection with my social group] has disintegrated, because, frst, my husband went … So yeah, that's sort of something you can't control any longer." The way in which she talked portrayed a perception of inevitability surrounding this decline in social interaction. Another 91-year-old participant described how she felt most alone after the death of her spouse because she no longer had someone to talk to about her day when she returned to an empty house, and this diminished her enjoyment of social outings.

Divorce was also described by some as a factor that impacted on their loneliness. One divorced woman, aged 81 and living alone, described the isolation that resulted from her marriage breaking down earlier in life. When asked about her current mental state, the participant described having experienced deep depression and resentment caused by the separation that made her earlier life "unmanageable". However, she framed this experience as one that had built her resilience in the face of recent confict with her daughter and a relocation: "When you're a few disasters over the way, you learn a bit." She describes actively making choices to maintain her friendships, such as ensuring that her new residential address was close to public transport, and does not feel lonely in the present.

Interviews indicate that modelling loneliness as a linear function of widowhood events is not a certainty. Several participants detailed the complexities of their relationships with their signifcant other, indicating that their separation had actively beneftted them. One widowed woman, aged 86 and living alone, detailed that her 60 year marriage had been more of a long-run "pal", and the marriage never resulted in the participant feeling love. The interview made it clear that, while there was clearly a sense of loss, the loss of her partner did not impact on her as would be normatively expected of a typical widowing. Another perspective was brought by a widowed woman, aged 76, who accounted for her continued unmarried status with an unhappy marriage to her late husband. The participant described an unwillingness to pursue any future partnership with men as, she perceived, the patriarchal nature of marriage is overly controlling. She described her situation, "You're lonely, but it's peaceful." She identifes feelings of loneliness, but utilises the term outside of its usual negative connotation.

Finally, one male participant, aged 83 and living with his spouse, detailed how re-partnership had mitigated feelings of loneliness in the wake of widowhood. After re-marriage, the newlyweds moved to a different area and found it easy to rebuild a social network as a couple. The man indicated that, in general, married couples seem to be happier and less lonely than those on their own.

#### **Conclusion**

The insights offered in this chapter demonstrate how a life course perspective can contribute to advancing our understanding of later-life loneliness. While the contemporaneous correlates of loneliness are well established, and research has indicated an association with earlier life events, the qualitative accounts in this study extend this body of knowledge by describing which life events are relevant, and the processes by which they may lead to loneliness among older adults. The principles of the life course perspective provide a useful lens for interpreting these fndings, particularly linked lives and human agency, which were applied throughout to explain the four key themes within the fndings. These themes were the formation of social networks, migration and relocation, a history of familial support, and widowhood and divorce. Across these themes, the fndings provide further detail to three broad observations about the nature of loneliness; that social support develops over the life course and is diffcult to replace at later stages, how individuals' perceptions of their circumstances shape their loneliness, but also sometimes empower them to take action, and the varied effects of relationship disruptions, particularly with family or a partner.

Participants valued the qualities of 'old' friendships, tied to a life history. This emphasis on long-standing relationships for supporting valued social connections highlights the diffculty of supporting older adults who have, for various reasons, already become socially isolated. Studies of older adults' social networks have found robust associations between diverse- or friend-focused social networks and the lowest levels of loneliness, which support the participants' emphasis on seeking out a variety of interest-based social groups (Litwin & Shiovitz-Ezra, 2011). This mosaic of social connections may be disrupted in cases where childhood or adulthood events affected the development of social skills, or where signifcant past life events disconnected friendships, such as through relocation or confict (Case et al., 2005; Ejlskov et al., 2020; Hensley et al., 2012). Two participants attributed their past circumstances to having shaped their expectations and coping mechanisms, and thus having better prepared them, mentally and emotionally, for their current life circumstances. However, participants' discussion of life events primarily centred on past events which were continuing to infuence their current circumstances. When participants discussed relocation, they focused on previous networks lost, and social gains from the move centred predominantly on family. This suggests that older adults are less likely to form new friendships, particularly after a loss of old ones. This fnding aligns with socioemotional selectivity theory, which suggests that as people age, they refne their social connections, focussing on strengthening existing close relationships and pruning more distant or less rewarding connections (Löckenhoff & Carstensen, 2004). The theory suggests that when people perceive their remaining time as limited, they switch from future-oriented goals seeking novelty, to present-oriented goals seeking emotionally meaningful experiences.

Across the older adult participants' responses, active and ongoing socialising during earlier life stages emerged almost as an assumed prerequisite to successful socialising in late life. However, the relationship between social isolation and loneliness is not linear, and experience of loneliness has been correlated with childhood and adulthood events even where the individual is not socially isolated in their older age (Ejlskov et al., 2020). According to the loneliness literature, the mediator may be how comfortable individuals feel with their current situation, and how much it matches the expectations that they have for their life. It may be that the participants who highlighted the need to seek out social situations to develop a diverse social network are those who have expectations of maintaining a large social network in old age. For these participants, life events which create discontinuities are likely to instigate loneliness. Conversely, some older adults may be satisfed with a small number of close connections. Across the sample, participants appeared to be highly conscious of their own agency and to feel empowered to make changes in response to perceived defcits in their friendships. Our sample of older adults were all community-dwelling, in affuent suburbs, and so likely were well-resourced to combat loneliness in their own lives. However, with the exception of relocating, they did not discuss taking actions to improve their family relationships.

Regarding family relationships, participants' loneliness was similarly related to the distinction between desired and actual circumstances. Family relationships have many normative connotations of intrinsic support, care and permanence, and where these were met, participants expressed satisfaction, while where confict or separation from family members meant that these relationships failed to have these intrinsic qualities, participants expressed dissatisfaction and loneliness. Partnership and its dissolution, through widowhood or divorce, appear as a complex issue in the interviews. While there is a strong body of knowledge that having a partner is protective against loneliness, this is moderated by the quality of the relationship, and is generally stronger for men than for women (Dykstra & de Jong Gierveld, 2004; Shiovitz-Ezra & Leitsch, 2010). These empirical fndings are refected in the interviews, as some participants identifed increased loneliness following widowhood due to the loss of a close emotional relationship, but conversely several female participants indicated that marriage had been an unfulflling or traumatic experience

with lasting effects upon their emotional wellbeing and related to subsequent life choices not to repartner.

Stability and continuity of relationships arose as a key factor in prevention of loneliness. This suggests that people's ability and inclination to make new social connections decreases, or simply that relationships strengthen with time, making longstanding ones more valuable in prevention of loneliness. The value of linked lives across the life course was most represented where participants described loss of these connections, through relocation and geographical distance, bereavement, family estrangement and relationship dissolutions. Furthermore, we can understand the importance of stability through cumulative disadvantage, as losses create the social and emotional context for subsequent diffculties. Participants referred to stable relationships which later provided support, lessening the impact of diffcult circumstances, and conversely, discussed how lack of support, through migration and bereavement, increased their diffculties in coping with later negative circumstances.

Our fndings are limited to a particular sample of older adults in Australia, and future studies drawing on a larger, more diverse and representative sample could re-examine the fndings reported in this chapter. Future research might also examine the relative impact of different life events on later life loneliness and consider a longer span of the life course in relation to later life loneliness. Our fndings also suggest the importance of greater knowledge and education on the benefts of maintaining strong social support and ties. Similar to health messaging on exercise and diet, health messaging could alert people to the importance of maintaining social networks and ties to reduce loneliness and increase wellbeing.

**Acknowledgements** We are especially grateful for the time of the research participants for speaking with us as well as the support for recruitment from the home care organisation. We would also like to acknowledge the support of the research assistants who took part in this project, for assistance with recruitment, data collection and analysis.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 14 Infuencing Social Policy on Families through Research in Australia**

**Tim Reddel, Kelly Hand, and Lutfun Nahar Lata**

For researchers to maximise the potential impact of their work, they must ensure that politicians and civil servants are aware of their fndings. This means learning how to communicate effectively with government and discovering the entry points into the policy-making process. (former British politician, David Blunkett, 2000 in a speech to the Economic and Social Research Council)

Justice Henry Bournes Higgins, President of the Commonwealth Conciliation and Arbitration Court, in handing down his famous Harvester judgement in 1907 establishing the 'Living Wage' researched and referred directly to the study of adequate wages conducted by Seebohm Rowntree (1901) in York, England in the late nineteenth century. This seminal labour law decision was perhaps one of the earliest examples in Australia of research informing directly public policy, and especially for families. The framework established by the Harvester judgement had clearly defned policy goals: a living wage to support a family with two children which continued as the foundation of wages (and family) policy in Australia until the early 1970s when it was replaced by a system of equal pay for equal work (McDonald, 2014, p. 130).

Research and evidence takes many and varied forms, often not integrated and strategic but driven by ad hoc opportunities in both public policy and research communities. Despite signifcant investment from the government in research it is often not released or acted upon—why? For example, the commissioning approach by the government for the research can be critical to determining release and/or nonrelease, or the form and content of the fnal research product is too complex and not conducive to public engagement or policy translation. This chapter explores both

Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia e-mail: tim.reddel@uq.edu.au; l.lata@uq.edu.au

K. Hand

© The Author(s) 2022

T. Reddel (\*) · L. N. Lata

Australian Human Rights Commission, Melbourne, Australia e-mail: kelly.hand@humanrights.gov.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_14

these broader systemic issues about the role of research in public policy and more specifcally key social policy episodes where research has infuenced social policies for families in Australia.

In this chapter we focus on how research can best be translated to real world outcomes in social policy that affect families. We discuss issues related to the measurement of disadvantage and how policymakers deal with this and develop policy to reduce family disadvantage and their impacts on different social groups. We present the research-policy relationship and discuss how research has infuenced social policy to solve family disadvantage issues. We conclude by suggesting that research should be directed towards the most pressing social problems to fnd social solutions.

The chapter builds on earlier chapters and other research by examining selected social policy episodes where research from academia, public sector, civil society groups and beyond have had an impact on the key national and state based policy systems central to this book's themes about the transmission of social and economic inequalities within families over the life course and across generations. We endeavour to draw out the implications of this research for social policies designed to support families and address the many manifestations of disadvantage in Australia.

#### **History and Context—Social Policy Research on Families and Disadvantage**

The preceding chapters of this book have illustrated that the experiences of families are at the heart of our lived experience. Whether these experiences of disadvantage are positive or negative, our current and past experiences of family have the potential to shape our futures. As highlighted in these writings there is a strong research evidence base historically around the importance of family in shaping outcomes for children, but also how families can provide both support and safety or be places of fear and danger (Perales et al., 2016).

As Chap. 2 of this book highlighted, cycles of disadvantage are numerous and complex, including defning cycles of disadvantage using a systems approach that incorporates life course theory and ecological systems theory, supplemented with substantive theories from disciplines such as sociology, developmental psychology, welfare economics, social policy and political theory. Ecological systems theory draws attention to individuals' interactions with their environments and recognises that these environments are embedded in social institutions, times and places. In addition, life course theory recognises that human development occurs across people's lives and that their lives are organised according to socially and historically specifc categories such as infancy, childhood, adolescence, young adulthood, middle age and later life. Some common events and transitions signal movement over the life course, and early events matter for later life events and outcomes. The life course perspective also emphasises the importance of 'linked lives', where individuals' lives infuence and are infuenced by others.

A life course and systems approach refect an increasing recognition in academic and public policy thinking that disadvantage is a multi-dimensional concept. As the Productivity Commission's 2013 working paper stated 'it (disadvantage) is about 'impoverished lives' (including a lack of opportunities), not just low income. Poverty, deprivation, capabilities and social exclusion are different lenses to view and measure disadvantage' (McLachlan et al., 2013). This defnition refects many historical and contemporary discourses about disadvantage in which the so-called standard approach to disadvantage—an income measurement approach to poverty—has been challenged and adapted. The importance of addressing the nonmonetary and relational aspects of advantage and disadvantage such as human capability and opportunity, the strength of community networks and civic engagement, and the impact of linked problems such as unemployment, poor housing, crime, bad health and family dysfunction on individuals and places have been highlighted (see Saunders et al., 2007).

A multi-dimensional approach to disadvantage is conceptually sound. There are, however, a number of practical challenges impacting on political, policy and community discourses in Australia (and more broadly). Measuring disadvantage is not a simple task. Research by the Productivity Commission (2018) has highlighted that no single metric is suffcient to give a defnitive answer to the seemingly straightforward question: have inequality, economic mobility and disadvantage in Australia risen, fallen or remained steady in recent years? This is the case because these concepts are multidimensional, and they link to each other—and to broader notions of wellbeing—in complex ways. To address this problem, the Commission focussed on measuring three elements of disadvantage: poverty, material deprivation, and social exclusion. While there are numerous data sources available to support the measurement (or at least estimate) the prevalence of these three elements, integrating them into a coherent disadvantage framework remains a challenge. These measurement limitations also infuence community understanding, public debate and public policy impact. Notwithstanding numerous inquiries, research and reports by a Royal Commission, parliamentary committees, non-government organisations and research bodies, policy making and community debate regarding the causes of and solutions to disadvantage have been narrow, cautious and limited in impact (see Smyth, 2014). A case in point is the annual release of the statistical report of the *Household, Income and Labour Dynamics in Australia* (HILDA) Survey. The 15th iteration of the survey report was released in 2020 and measures many dimensions of disadvantage including the intergenerational transmission of income poverty and economic and subjective wellbeing. The survey has been funded and supported by the Australian government since 2001 and is highly regarded for its scope and rigour. A challenge remains, however, to convert this empirical research into an impactful narrative that can infuence community views and policy making in a measurable way.

Given this research and historical legacy it is not surprising that 'the family' is a key point of public policy intervention for addressing a multitude of social and economic problems confronting population groups and communities. Social policy should ameliorate disadvantage for people and places and also be good

economic investment promoting economic participation and productivity (see McClelland & Smyth, 2014, p. 113). From the 1960s, social policy research has traditionally not been well integrated with the economy and often seen as unproductive—'let markets grow and welfare provision address equity later'—(Smyth & Buchanan, 2013).

Constructs such as family functioning and wellbeing are seen as key points of social policy analysis and intervention to address disadvantages including increased unemployment, anti-social behaviours, poor physical/mental health and intergenerational impacts and building individual capability and improving life choices (Sen, 1999). Improving family functioning, generating protective factors and mitigating risk factors are key policy levers in mediating the impacts of disadvantage and improving individual wellbeing (Department of Family and Community Services, 2001; Heady, 2006; Gorecki & Kelly, 2012). This approach, particularly Sen's theory of capability i.e., 'a persons' capacity to be whom they want to be' point to a more integrated view of wellbeing and a joining up of economic and social policy. Heady (2006) talks about capabilities and functioning in four domains of life: the fnancial domain, the employment or labour market domain, the health domain, and the family and social domain leading to improved individual and family wellbeing. In the Australian context this capabilities approach has been infuential in key public policy initiatives such as Commonwealth Treasury's wellbeing framework of the early 2000s and non-government social policy development by major players (and from different ideological perspectives) such as the Brotherhood of St Laurence and the Cape York Institute (see McClelland & Smyth, 2014, p. 117).

In Australia family policy is seen as the preserve of both Commonwealth and State and Territory governments, but not in any coordinated way. Policy makers seek to impact family wellbeing and the ways that families engage with the economy, labour markets, education and other systems through a variety of policy levers—such as income support payments and subsidies, the tax system and statutory mechanisms that seek to promote child safety and engagement with education. While many policies have universal aspects, many are largely focused on strengthening the individual capabilities of, and promoting positive outcomes for, families experiencing disadvantage. To date, there have been limited attempts by governments to develop national strategies to promote the wellbeing of Australian families. For example, The Howard government established the Stronger Families and Communities Strategy in 2000 with the aim to create new partnerships to strengthen families and communities by investing in prevention and early intervention responses and community capacity to solve local problems.1 More recently, the *National Framework for Protecting Australia's Children (2009–2020)* and the *National Plan to Reduce Violence against Women and their Children (2010*–*2022)* are respectively partnerships between the Commonwealth and State/Territory

<sup>1</sup> https://www.dss.gov.au/about-the-department/publications-articles/corporate-publications/ budget-and-additional-estimates-statements/2000-01-budget-and-additional-estimates/a-strongerfamilies-and-communities-strategy

governments to reduce child abuse and neglect, and coordinate prevention and other actions to reduce violence.

Recent research exploring the patterns and predictors of disadvantage in Australia consistently place families at the centre of this research and tell a story of the experience of disadvantage earlier in life being a strong predictor for experiencing disadvantage later in life. Further the importance of community and place in predicting and addressing disadvantage remains a consistent theme. However, the stories presented within the data about families and the recommendations made are often at odds with dominant policy discourses that promote behavioural change to improve family functioning and instead suggest more complex structural and community level solutions are required to work in tandem with building the capabilities of individuals and families.

This is particularly the case when looking to understand and address the issue of entrenched or persistent disadvantage. In these circumstances families are framed as the key site of both the transmission of disadvantage as well as the key point of intervention. Understanding the role families play in this transmission is critical. Factors such as employment, education and family type are strong predictors of disadvantage and changes in these statuses can impact on pathways in or out of poverty (Ananyev et al., 2020). However, using a life course approach, Vera-Toscano and Wilkins (2020) argue that 'the intergenerational transmission of poverty is not the sole responsibility of families' and that instead a lens that takes into account public policy levers and both national and global contexts need to be used to understand the causes of, and levers to address entrenched disadvantage within families. Factors such as low family income, for example, intersect with other factors such as housing affordability and unemployment rates to impact life opportunities (Vinson & Rawsthorne, 2015). Solutions to address the complex nature of entrenched disadvantage therefore need to also need to be multifaceted and target both individuals and families as well as the systems and contexts in which they live (CEDA, 2015).

#### *Community, Place and Family*

A common theme in research on disadvantage for families is the role of community or place in both understanding disadvantage and in addressing it (Vinson & Rawsthorne, 2015; Payne & Samarage, 2020; Ananyev et al., 2020). Living in a community where there are high rates of poverty or other indicators of disadvantage is a strong predictor of experiencing persistent disadvantage and addressing disadvantage at the community or place based level is seen as an important pathway in moving people out of entrenched disadvantage. It can be argued that research related to place has infuenced government policy approaches. Both at the Commonwealth and State and Territory levels, place-based policies and programs are seen as effective ways of addressing the complex nature of disadvantage experienced by families through looking at their physical and social environment and the service systems they engage with rather than looking solely at the issues they face as individuals (Centre for Community Child Health, 2011; Dart, 2018; Victorian Government, 2020). Over the past decade, place-based approaches have been increasingly rolled out as policy responses to complex social problems across Australia. These approaches are driven by the notion of local answers to local solutions and have consultation and shared decision making as core components of these models. However, place-based approaches in the Australian context have been characterised by trials, pilots, time limited programs and a narrow focus on human service delivery rather than broader policy design (Reddel, 2002).

#### *Australian Social Policy and 'the Family'*

This scan of key social policy research reports (and of course many others) highlights that poverty and disadvantage are contested terms, impact differently on individuals, families, households and communities and have had varied levels of policy infuence and impact. In order to provide some form of social and income protection to Australian families, the Australian government began a national system of (limited) care for its citizens in 1901 (Stanton, 2001). The frst social security program was passed in June 1908 that "provided for the introduction of means-tested fat-rate age and invalid pensions, fnanced from the government's general revenue" (Stanton, 2001, p. 3). Following this, a number of social policy reforms have been introduced to provide social and economic support to Australian families (see Table 14.1).

This historical scan, while long, is by no means complete, but does highlight that the Australian social welfare system reformed signifcantly during and after the Second World War and this reform is related to the 1941 review of Australia's social policies by the House of Representatives Joint Standing Committee on Social Security (Shaver, 1987). To improve post-war life, especially for families, the committee recommended several additional measures including widow's pension in 1942, unemployment beneft in 1945, and established the Commonwealth Employment Service (Marston & Staines, 2020). Australian families have benefted from the introduction of a series of social policies and programs since the early 1900s (see Table 14.1) that have been aimed (with varying degrees of impact) at contributing to the "cost of bearing and raising children, redistributing resources over the life cycles, alleviating child poverty and boosting family earnings, promoting equity within the tax system, redistributing within families and relieving unemployment and low income traps" (Whiteford et al., 2001).

These events and reforms have, however, not benefted all citizens equally. First Nations families were not even entitled to receive most of these benefts until 1959. For example, First Nations peoples were not eligible for unemployment benefts until 1959 although unemployment benefts were introduced in 1945. Even after 1959, First Nations peoples had to address several personal related criteria such as "having a fxed address, completing forms written in English (which was often a second or third language), and providing documentation like birth certifcates"


**Table 14.1** Selected Australian social policy reforms impacting on families (from 1900s to present)

Sources: Adapted from Arthur (2015), Herscovitch and Stanton (2008), Marston and Staines (2020), Parliament of Australia (2013), Stanton (2001) and Yeend (2004)

(Marston & Staines, 2020, p. 219). In addition, their unemployment benefts were placed into their protectors' bank accounts (Marston & Staines, 2020).

First Nations peoples also received less wages in comparison to non-Indigenous Australians until 1965. The 1965 Equal Pay reform played a key role in declining the Indigenous Australians' employment opportunities as their labour was not cheaper any more. Consequently, nearly half of the First Nations workforce was unemployed by 1976 (Sanders, 2012). Within these circumstances, social security benefts played an important role in providing incomes for Indigenous families in areas where labour markets were particularly weak and seasonal (Altman, 2011). Yet, to this date, Indigenous Australians have lower incomes on an average in comparison to their non-indigenous counterparts (Staines, 2017; Marston & Staines, 2020). Similarly, First Nations peoples still must meet several behavioural conditionality requirements to access social security benefts (Marston & Staines, 2020).

This book's contributions have highlighted a range of recent debates and controversies impacting directly on family dynamics over the life course where research and broader social policy have been critical factors. In broad terms these include population ageing and families; family formation and functioning; poverty and disadvantage—individuals, households, place and mobility; and work and family. The current COVID-19 pandemic is and will of course continue to have signifcant implications for families and social policy research. Agility and fexibility for researchers and policy makers will need to be the standard operational model for their relationship.

#### **How Has/Should Research Infuence/d Social Policy on Families and Disadvantage?**

This current challenging context has highlighted many longstanding tensions and opportunities for the research-policy relationship. It should be self-evident that not all research activities are infuential or impactful in contemporary social policy development for families, other cohorts and places. The Australian Research Council's (ARC) Excellence in Research for Australia (ERA) and Engagement and Impact (EI) Final Report (2021) canvasses at a high level some of the complex relationships, incentives and barriers to research impact and engagement between researchers and end users. Importantly, the ARC recognises that engagement and research impact are related but not the same. There is, however, a need for a more granular understanding and analysis of the researcher's and end user's experience of the research-policy impact process. There continues to be a danger that greater attention to research impact (and engagement) by academics and other researchers will 'simply result in strategic manoeuvring to present good impact stories' rather than produce better community outcomes (Hughes, 2016).

'The vectors in the 'golden triangle' of research, policy, and practice are rarely intuitive.2 Those who do research do not necessarily understand policy or practice; those who work as practitioners, judicial decision makers, or clinicians do not necessarily understand policy or research; and those who develop and refne policy do not necessarily understand research or practice' (Smyth, 2011). The challenge of translating research into policy and practice is a long standing one. While it has been seen to gain greater importance in recent years, with signifcant interest in the development of knowledge translation and impact as a discipline in its own right, there is little evidence that an emerging 'knowledge translation' movement has managed to shift the use of research in policy making. Barriers include the confdence of researchers to develop translational materials as well as a lack of resourcing to undertake such work (Newson et al., 2015). Tsey et al. (2019) and his collaborators sum up the need for pragmatism. Flexibility and a 'learning by doing' approach to research-policy impact:

Assessing research impact is best approached as a 'wicked problem' for which there are no perfect templates. It requires fexible, transparent, collaborative learning-by-doing approaches in order to build the evidence base over time. At these formative stages of the evolving impact agenda, government assessment criteria must allow researchers the maximum fexibility to demonstrate the true impact of their research. (Tsey et al., 2019, p. 182)

#### *How Can Effective Research-Policy-Practice Translation Work?*

While knowledge translation and research impact on policy and practice is now popular, it is not necessarily a new idea. Looking back historically, Charles Lindblom specifcally challenged social scientists to think more creatively and practically about 'usable knowledge' sources to better defne and solve policy problems (Lindblom & Cohen, 1979). They emphasised the benefts of engaging with the 'common' knowledge and expertise of service delivery practitioners, administrators, community groups and public commentators. Harvard academic Carol Weiss (1986) argued that research can be utilised in many different ways: as a process of enlightenment that flls the 'well of knowledge' from which all may draw; as a way of lubricating the machinery of policy development by solving problems; and as a way of raising awareness and exerting pressure for action. She has emphasised the importance of ideas, arguing that:

More often, it is the ideas and general notions coming from research which have had an impact … [they] are picked up in diverse ways and percolate through to offce-holders in many offces who deal with the issues.… Because research provides powerful labels for previously inchoate and unorganised experience, it helps to mould offcials' thinking into categories derived from social science. (Weiss, 1986, p. 218)

<sup>2</sup>The phrase 'the "golden triangle" of research, policy and practice' is borrowed from Alan Hayes (2004, p. 7).

Leading Australian social policy researcher, Barbara Pocock (2005) has argued that 'many researchers take the JD Salinger approach to publication, that is they pack their academic article, put it on the bus, and send it out into the world to fnd its own readership and effect. Unfortunately, not all research products are the social science equivalent of *Catcher in the Rye* and many sink without trace' (Pocock, 2005, p. 136). Powerful tensions exist between these two approaches—between *doing research and talking about it* (and we would add acting on research to inform policy and practice). Pocock proposes a research-policy impact and translation framework based on what she calls the '*Technologies of Infuence: Transmission mechanism for research*' addressing a variety of audiences, communication and engagement processes and mechanisms that are tailored to various policy actors including politicians, bureaucrats, media, scholars and the 'public' (2005, p. 137).

Another distinguished social policy researcher, Peter Saunders (2011) refects on his long research career and offers two refections on the relationship between research and policy. The frst is the 'importance of the conventional hallmarks of research quality—independence of thought, conceptual sophistication and methodological rigour' as a foundation for determining what research gets used by policy makers. The second is the need for researchers to both understand the context and substance of a particular policy activity but also how to disseminate their fndings effectively by identifying better ways of connecting with policy makers and practitioners (Saunders, 2011, pp. 250–251). Morton (2015, p. 407) drills into researchpolicy impact and translation conundrum by highlighting the key challenges as timing, attribution and diffculties in addressing context. She proposes a more interactive model of research-policy impact that understands both processes and outcomes through the dimensions of interaction, dissemination, policy making and adoption.

It is also important to acknowledge that these very worthwhile refections are mostly part of the concluding comments of the author's writings and could be seen (perhaps harshly) as an 'afterthought'. This is of course mostly not intentional, but it might be a metaphor for the broader problem we are wanting to address. It is a confusing landscape and academics, policy makers, and broader public policy stakeholders need to refect and where necessary reframe our traditional approaches to research, analysis, policy impact and engagement. The ARC' 2018 *Engagement and Impact Assessment* exercise refected a broader institutional desire by researchers and the university sector to improve the social, economic and environmental impacts of research across all academic disciplines.3

Social science researchers and their academic institutions have argued for many years that their disciplines need to do more to solve real world practical problems (Western, 2019). 'End users' of research including governments, industry, philanthropy and civil society are increasingly concerned with policy outcomes, greater use of data, evidence, multiple forms of expertise, co-design

<sup>3</sup> International interest in greater policy impact from research evidence can be seen in Sasse and Haddon's (2018) study for the UK's Institute for Government on improving links between government and academia.

with stakeholders and better policy translation of research from their engagement with academics. Co-production rather than one-off consultations and siloed conversations are necessary ingredients to more effective research translation and impact (Blomkamp, 2018).

A central argument is that the public value and utility of academic and commissioned research should give (but not necessarily be captured by) greater attention to understanding individual and population level needs, identifying, defning and measuring outcomes as the basis for improved policy design, implementation and evaluation. As Western (2019, p. 21) argues the 'fundamental goal (for university research), however, is to design elements of the research ecosystem to produce outcomes that all participants (policy makers, service deliverers, researchers, civil society organisations and other stakeholders) value'. Consequently, researchers need to listen and build partnerships rather than expecting to talk and be heard by stakeholders, especially government.

To inform discussion and thinking, an initial set of criteria building on this literature, to support solution based research-social policy-practice impact is proposed:


Importantly, this indicative and challenging criterion argues that solution based policy research impact is not necessarily the same as good engagement or even effective research to policy translation and communication. These dimensions are important but not always suffcient to solve complex and 'wicked' policy problems. As this book's contributions including this chapter's historical scan have

highlighted, developing policy responses to the multiple dimensions of family dynamics and the transmission of (dis)advantage over the life course has been and is complex and iterative.

#### **How Has Research Infuenced Social Policy and Life Course Approach to Understanding Families and Disadvantage?**

Given this complexity, applying the preceding criteria to an assessment of how research has infuenced social policy and a life course approach to family dynamics and disadvantage will necessarily be incomplete. This book's contributions cover a wide range of related themes and issues—early childhood, Indigenous families, refugees/migrant families, adolescence, schooling/higher education, emerging adulthood, labour market participation, marriage and partnering, parenthood, LGBTIQ+ families, ageing and loneliness. All of these contributions have referred directly to the authors' own research together with other scholarship utilising a variety of methodologies with key fndings directed at improving policy development, program design and the broader reform of social structures.

This scholarship should be considered alongside other more applied research. For example, government funded and authorised research bodies such as the Australian Institute of Family Studies (AIFS) aim to balance quality research into the wellbeing of families with communicating fndings to policy makers, service providers and the broader community.4 AIFS's research portfolio traverses 'pure' research with more applied policy research and commissioned program evaluations including family law reform, family relationships, marriage formation and breakdown, child protection and social security system (Saccotelli & Muldoon, 2018). Statutory research bodies such as AIFS value their independence but also proximity to government. AIFS has undertaken a number of signifcant evaluations during its history including the evaluation of the 2006 Family Law Reforms—this evaluation contributed to further changes to the 2012 Family Violence Amendments to the *Family Law Act 1975. Growing Up in Australia: the Longitudinal Study of Australian Children* (LSAC) is also a fagship study for the Institute. Undertaken in partnership with the Department of Social Services, LSAC has followed 10,000 children and their families since 2003 to examine children's developmental pathways in the contexts of their family and the communities in which they live. AIFS also plays an important role in the translation of research evidence for the child and family welfare sector, as well as building this sector's capability to evaluate the effcacy of their own practice. Statutory based research institutes such as AIFS are a rarity in Australia, especially those focused on social research. Policy and program related research and evaluations are more often directly commissioned by government agencies to third parties such as universities or consultancies. High profle

<sup>4</sup> https://aifs.gov.au/about-us/what-we-do

evaluations over a number of years of national programs such as the Communities for Children initiative and the Cashless Debit Card trial were carried by a mix of university-based researchers and consultancy bodies. All were intended to develop an evidence base for the further development of the respective program and to varying degrees were used by both critics and advocates to argue a particular view. The various evaluations of the Cashless Debit Card, for instance, have been the most politically contentious and debated. The Commonwealth government insists on the program's merit while many community advocates, researchers and political opponents argue that the evaluations have been methodologically fawed and the Card's income management framework is inherently unjust (Klein, 2020).

As the indicative solution focussed criteria in the previous section highlighted, policy related research and evaluation are often politically contested with inherent tensions between rigorous research processes and short term political and policy 'success' variously defned. However, this tension and at times disconnection should not be simply characterised as a contest between 'virtuous researchers' and pragmatic and short sighted policy makers and politicians. A commitment to longer term research and policy development focused on structural reforms to address the long standing inequalities and disadvantages experienced by individuals, families and communities should be a fundamental goal of social science research. However, solution focussed social and policy research requires iterative approaches which can bring together multiple disciplines and views, and as we have highlighted 'a policy problem is not solved' until it is implemented accountably, adaptively effectively'.

#### **Conclusion: Future Opportunities and Challenges**

This chapter has examined selected historical and contemporary social policy episodes that have been engaged in the dynamics of families in Australia and the research and evidence that has infuenced their development and implementation. Beginning with the Harvester decision in 1907 we have traversed court judgements together with legislative, policy and program reforms and controversies. While there has been, and continues to be signifcant research and evaluation activities associated with these episodes, their policy infuence or impact is not easily assessed. More integrated approaches to data and evidence including the role of big-data complemented by qualitative methodologies can support more experimental policy research to address the 'real world' needs of individuals, families and communities. For example, better linkages between long standing child-family centric data sets such as LSAC and other administrative social security, education and health related data as part of co-designed intervention strategy for a local community or specifc cohort offers promise. More strategically, the largely ad hoc approach to social policy research and policy practice for families over the life course requires reform. The promise of a future social policy research system offered by what United States researchers Ron Haskins and Greg Margolis (2015, p. 239) called an evidence based movement "of thousands of evidence-based social programs that address each of the nation's most important social problems and that under the onslaught of these increasingly effective programs, the nation's social problems will at last recede". This challenge and opportunity should be embraced by policy makers, researchers and community advocates, especially as we respond to the COVID-19 pandemic with its unpreceded disruption and unavoidable dilemmas for individuals, families and communities.

**Acknowledgements** The authors thank Phillip Brown and Brenton Philp for their helpful comments and conversations that helped shape this chapter.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

# **Chapter 15 Emerging Directions and New Challenges**

#### **Stephen R. Zubrick, Rennie Lee, Janeen Baxter, Jack Lam, and Jenny Povey**

We commenced this volume with the observation that an individual's family background exerted a substantial—and defning—effect upon opportunities, expectations, and outcomes in their development across the life course. In this concluding chapter we refect that these effects are transmitted intergenerationally and, as our colleagues show, they span developmental life stages, occur in specifc institutions, and manifest differentially within specifc populations. The global South is relatively sparsely represented in inequality research, and the contributions in this volume are focussed upon Australian research fndings providing new additions to a feld of research that is, with notable exceptions, focussed on Europe or North America. We use our fndings to pose new questions and conclude by summarising emerging directions for Australian life course researchers.

#### **Disadvantage Across Developmental Stage, Institutions and Populations**

When viewed within developmental stages across the life course, disadvantage plays out in numerous ways. In childhood our colleagues show that slower rates of cognitive and academic growth emerge in response to disadvantages in

R. Lee · J. Baxter · J. Lam · J. Povey (\*)

S. R. Zubrick

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Telethon Kids Institute, University of Western Australia, Perth, WA, Australia e-mail: stephen.zubrick@telethonkids.org.au

Australian Research Council Centre of Excellence for Children and Families over the Life Course, Institute for Social Science Research, The University of Queensland, Brisbane, QLD, Australia

e-mail: rennie.lee@uq.edu.au; j.baxter@uq.edu.au; j.lam@uq.edu.au; j.povey@uq.edu.au

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8\_15

developmental circumstances. They show that some disadvantages cluster differently for some children and differentially delay onward capabilities that then compound in effect over time. In adolescence, a time where increasing autonomy and independence is accompanied by critical decision points about education and vocation, parental engagement in children's education—irrespective of family and school context factors—exerts a positive infuence in tertiary study aspiration. Gratifyingly, where this parental engagement is diminished or missing, it can be provided by teachers, service providers or community members suggesting avenues for countering this aspect of disadvantage. By ages 18–25, typically characterised by rapid increases in decision-making and autonomy, our authors fnd that extended dependence and co-residential support have now extended the period of emerging adulthood. Young people from privileged backgrounds appear to have more opportunities to invest through emerging adulthood. Young people with high-income parents receive co-residential and fnancial support longer than young people with low-income parents. So, while autonomy might be delayed owing to co-residential status for more privileged emerging adults, the savings achieved by them may translate into greater life-advantages relative to disadvantaged young people with shorter periods of co-resident status.

Disadvantage and its transmission are also structured by the institutions of family and marriage, education, labour markets and in government decision making. Within the institution of marriage, results in this volume support the adage "two heads are better than one," fnding those who marry to be the least disadvantaged while those who remain single appear to be the most disadvantaged with cohabitors somewhere in between. From single status, our colleagues show that transitions to cohabitation and to marriage each result in advantages relative to one another in employment, fnancial wellbeing, and subjective wellbeing. The institution of education, designed in part to counter the effects of disadvantage and provide upward social mobility, displays a large SES gap in university enrolment and participation. Where disadvantaged students do complete tertiary education, it takes a relatively long time—at least 4–5 years after graduation—for the average low SES graduate to achieve outcomes in job-security, satisfaction and fnancial prosperity comparable to the average high SES graduate. Labour markets are particularly problematic in imparting disadvantage to some and not others. This is seen especially in the conficts between workforce participation and parenthood. Women in disadvantaged households and those who have low skill levels, and thus low earning power, often experience the largest impacts offering a powerful mechanism for sustaining inequality in society, and most importantly, intergenerational transmission of inequality and disadvantage. Government institutions impart considerable instability into the lives of disadvantaged families and individuals. National strategies to promote the wellbeing of families are largely absent, children and family policies sit—at best—uneasily between State/Territory and Commonwealth jurisdictions. Dominant policy discourses most often fail to appreciate structural and community mechanisms that maintain disadvantage—or alternately offer solutions—in favour of presuming that most disadvantage can be addressed through individual behavioural change. Most of these "one size fts all" policies overlook contextual circumstances of specifc groups in places.

Finally, our colleagues show that disadvantage differentially concentrates and responds in specifc Australian populations. This is most vividly seen in respect of the Australian Indigenous population where disenfranchisement and historical injustices have produced, among other effects, persisting inequalities in autonomy, ownership, income, health, and education that have translated into shorter lifeexpectancies and intergenerational disadvantage. Our colleagues illustrate that attachment to Indigenous culture and cultural identity are individual assets that contribute to the health and wellbeing and buffer the negative effects of disadvantage in Indigenous children. Within the Australian lesbian, gay, and bisexual population structural stigma has decreased considerably but there is more to do. Large disparities in mental health, emotional wellbeing, self-harm and suicidal behaviours persist for LGB individuals. But our life course researchers show that as Australian society has become less heterosexist and homophobic, parents are less likely to reject their LGB children and as the life course progresses the emerging picture suggests resilience. Within Australian refugee populations, while challenges remain considerable, the picture of disadvantage is more hopeful: Children are outperforming their parents.

#### **Emerging Directions**

Our volume shows that social disadvantage is not equally experienced across all segments of Australia's population. As noted in Chap. 2, this may require future research to continue to push forward the life course framework to better account for the trajectories for groups such as those that differ by national origin, gender, sexuality and place. Several chapters in this volume illustrate that marital status, family dynamics, and socioemotional wellbeing differ by subgroups suggesting that life course principles need to be extended to fully account for the diversity in outcomes we observe across social groups. Persistent differences across social groups indicate the limits of life course theory for fully understanding these patterns. While we know that social groups experience different fates, we know less about how social events have different meaning by groups.

One way to address this is through a stronger focus on intersectionality to help us better understand how disadvantage is stratifed by social groups, such as race/ethnicity, gender and class. An intersectional lense posits that individuals occupy multiple social locations simultaneously, all of which may compound on one another to produce different types of privilege and discrimination (McCall, 2005). For instance, not all women experience the same outcomes; rather they will vary by their ethnic/racial and class backgrounds. Thus, if we pair a life course approach that acknowledges that where we end up has a lot to do with where we start, and where we start has a lot to do with the various social locations we occupy, this can give a richer and fuller understanding of individuals' trajectories.

A second arena that we believe is important for future inquiry is to consider the role of place. Place has played a central role in Australian social policy as disadvantage or advantage is often concentrated in geographical areas or regions (Vinson & Rawsthorne, 2015). Place provides residents with different opportunities, amenities, potential access to networks, and social norms, which may shape important life outcomes. As Fincher (2021) has outlined, it may be useful to re-theorise place not just as bounded geographical sites defned by census tracts or postcodes, but as "locales" or communities shaped by the identities of the people who inhabit the space. Such an approach shifts the emphasis away from viewing place as a static, bounded geographical site to recognise the dynamic and fuid features of places, and their social features such as the histories, identities, intentions and interests of the people and communities that live in them. Given the centrality of place in shaping Australian social life, future work that focuses on understanding how place may act as a form of stratifcation in and of itself in producing locational advantage and disadvantage is important.

Australia now has a diversity of longitudinal data sources. This is seen in the broad usage of data examined in this volume from the Longitudinal Study of Australian Children, the Longitudinal Study of Indigenous Children, the Household, Income and Labour Dynamics in Australia Survey, the Building a New Life in Australia study, the Longitudinal Surveys of Australian Youth, and the Australian Census Longitudinal Dataset, as well as qualitative data. While many of the data sources provide insights across the Australian population, there are now sources of data that focus on small and diffcult to recruit populations, notably Indigenous Australians, migrants, and refugees. Designed over the past 20 years, these are fundamentally new innovations in Australian data assets, and they refect an emergent direction in the thinking of Governments, researchers, policy makers—and indeed the Australian citizens that participate in these studies.

Further insights into life courses are enabled through provision of integrated and administrative data—this is becoming more prevalent and may open doors into fndings more relevant to policy and decision making. Such data bring with them additional responsibilities for managing confdential data as well as meeting the demands of their quantitative complexity. Methodological advances are evident—both in the design of longitudinal data sources themselves which now encompass cohort, panel, time series, and integrated administrative data. Australia's Multi-Agency Data Integration Project (MADIP) represents a considerable advance in national capacity to securely govern data integration and design safe access methods protecting privacy while allowing critically important planning and forecasting in health, welfare, and education and across governments. So too do individual state jurisdictions; notably, Western Australia and New South Wales have long standing data linkage capabilities that have pioneered the value and contribution that data integration can make in improving our understanding of how services are used, what they contribute, and in describing the health and wellbeing of Australian citizens.

Research in this volume also refects the statistical and methodological competencies that increasingly complex data demand. Latent class analysis, growth curve and random effects models are now typically applied to examine associations and

change. While these are used to reveal important effects, there is room for further innovation. For example, developments in relation to machine learning and predictive risk modelling that have potential to improve outcomes are further advanced in some disciplines and only slowly fnding their way into social science to examine complex outcomes such as child abuse (see, Cuccaro-Alamin et al., 2017) and suicide and suicidal behaviour (Chen et al., 2020).

A substantial challenge to the many parties involved in life course studies is the planning and sustaining of these capacities. Longitudinal surveys and integrated administrative data provide an increasingly important source of information about Australians living their lives, over time and in place. The onward planning for sustaining some of these sources, refreshing others, and creating new sources poses both a challenge and opportunity to ensure that Australia has research and planning capabilities up to the task. Some worry that these data enable "surveillance of the poor" and reduce the complexity of disadvantaged life courses to numbers (Staines et al., 2020). While this is a legitimate concern, the research presented in this volume suggests that this propensity is both offset by the practical value of fndings that describe and address intergenerational disadvantage as well as by methodologies of narrative and qualitative studies of lived lives.

In reviewing the conduct of longitudinal studies there is still the need for researchers to implement designs that capture change processes and particularly those with better measures of growth and change in physical, cognitive, emotional, and social capabilities; and that reveal more about the constrained choices, decisions and actions taken (and not taken) by the child (and carers) as they grow to adulthood. With respect to the transmission of disadvantage, a more deliberate focus on dynamic mechanisms of change, rather than the status of an outcome at one point in time, is needed. Disadvantage is characterized by the impaired acquisition, poor accumulation, transformation or loss of human capital and human capability across the life course. Implementing study designs that defne and measure this acquisition, accumulation, transformation and loss of skill, knowledge and effort as individuals grow is essential in addressing underlying mechanisms of disadvantage.

Finally, it is important that our data sources do not circumscribe our understanding of the relational, systemic and institutional features of social disadvantage. Studies based on collecting information about individuals, as many social surveys do, encourage a focus on individual attributes and characteristics as the main drivers of disadvantage and shift attention away from institutions such as families, labour markets, schools and governments. But as many chapters in this volume show, inequalities are generated by the social positions that individuals occupy and the relationships between people, positions, and institutions, not by individuals acting alone (Tomascovic-Devey & Avent-Holt, 2019). Individuals have agency to shape these relationships but within bounded contexts and circumstances. It is important that we continue to foreground the social and relational nature of disadvantage and encourage policy solutions that go beyond changing individual behaviour to those that change the opportunity structures that defne social relationships.

#### **Final Thoughts**

One clear omission is that our book focuses on Australian society before COVID-19, with the exception of Chap. 6. The events of the last 2 years have brought into clear view many of the themes and issues underlying the chapters in this volume. COVID-19 has shone a spotlight on the social, cultural and economic cleavages that underpin our societies and made visible often unrecognised and unacknowledged disparities across individuals, groups and places in access to resources, opportunities and services. It has also brought into sharp view the consequences of increased risks and uncertainties, and how age, place and the timing of events are so consequential for opportunities and outcomes going forward. For many Australians COVID-19 is their frst experience of a truly global sudden disaster. While some older generations may have experienced the loss, disruption and devastating impacts of the Great Depression and two World Wars, for younger generations the experience of COVID-19 and the range of measures put in place to combat the pandemic is their frst encounter with global social disaster that has both short and long-term impacts on their lives.

To understand the full impact of the pandemic we will need a long view of the implications and outcomes of events and some groups will certainly fare worse than others (Settersen et al., 2020). Young Australians may be particularly affected. The uncertainties and day-to-day routine changes brought about by school closures, employment disruptions and lockdowns are not only affecting their wellbeing now but will certainly shape their onward life course journeys. Australia has fared comparatively well, to date, in terms of hospitalisation and mortality rates compared to many other countries and some of the early government responses to the outbreak raised unexpected opportunities for positive redesign of policy (Baxter et al., 2021). But the economic upheaval, job loss, school stoppages, business closures, education disruptions, travel restrictions, border closures and psychological consequences of virus outbreaks and the consequent sustained lockdowns is likely to have wide-ranging and long-lasting consequences, many of which are still largely unknown.

We hope that our work here highlights the value of a life course approach for studying these consequences and for understanding variations in outcomes across life course stages, places and populations. We believe that the value of such work will be greatly enhanced if the research is multidisciplinary and genuinely collaborative across sectors and agencies. Research that speaks only to colleagues in the academy and those with like-minded orientations will be much less useful in our view than research that is oriented outward to current policy challenges and speaks to governments and agencies that are tasked with designing solutions. Genuine multidisciplinary and cross-sectoral partnerships and collaborations that address research design, data collection, analyses, and translation of evidence into solutions is essential if we are to be successful in reducing social disadvantage and responding to emerging challenges.

#### **References**


**Open Access** This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

## **Index**

#### **A**

Aboriginal and Torres Strait Islander, 4, 6, 8, 57–67 Administrative data, 7, 18, 51, 150, 192, 316, 317 Adolescence early, 100, 121, 158, 233, 240, 255 late, 100, 114, 115, 117, 120, 121, 158 middle, 100, 121, 227 Adolescents, 9, 21, 28, 73, 78, 84, 93, 97–114, 117, 118, 120, 122, 161–163, 192, 223, 225, 228, 229, 231, 234–236, 253–257, 260, 272 Afterthought, 306 Anxiety, 58, 97–99, 101, 105–109, 117, 119, 121, 234, 254 Aspirations, 9–11, 25, 97, 98, 100, 101, 109–115, 117, 120–122, 141, 192, 196, 252, 266, 268, 270, 273, 314 Attitudes, homophobia, 255 Australia, v, vi, 3–6, 9–13, 17–30, 38, 48, 49, 53, 57, 58, 60, 65, 71–93, 97, 105, 106, 109, 115, 117, 133, 137, 139–150, 157–173, 177, 178, 181–184, 186–192, 194–196, 201–205, 209, 212, 214, 217, 228, 232, 233, 235–239, 251, 252, 257, 258, 264, 267, 270–272, 283, 288, 289, 293, 297–310, 315–318 Autonomy, 10, 99, 116, 120, 121, 159, 161, 167, 168, 173, 229, 314, 315

#### **B**

Big-data, 309 Birth, 6, 18–20, 23, 37, 48–50, 61, 86, 87, 89, 91, 133, 134, 188–190, 193, 196, 197, 226, 227, 230, 232, 284, 302 Bisexual, 251, 257, 258, 264, 267–269, 273, 315 Boomeranging, 173 Bullying, 72, 99, 101, 104, 105, 107–109,

#### **C**

Carastathis, 254, 255, 257, 266, 272

113, 121, 252

Caring/caregiving/caregiver/carer, 7, 21, 65, 78, 117–120, 179–181, 190, 193–196, 227, 283,


Child development, 23, 37, 40–42, 44, 48–54, 137, 226, 228

Child protection, 100, 115, 117, 236, 308


211–215, 218, 226, 314

Colonisation, 8, 58

© The Author(s) 2022

J. Baxter et al. (eds.), *Family Dynamics over the Life Course*, Life Course Research and Social Policies 15, https://doi.org/10.1007/978-3-031-12224-8 Community, 8, 11, 12, 22, 25, 26, 51–54, 58–60, 65, 66, 71, 72, 74–78, 81, 92, 93, 117, 122, 144, 150, 157, 187, 223–226, 229, 233–240, 252, 254, 257, 263, 265, 272, 284, 287, 297, 299–302, 304, 305, 308–310, 314, 316


#### **D**

Debt, 168–170 Defcit narrative, 9, 64, 66 Depression, 59, 92, 97–99, 105–109, 121, 203, 232, 266, 290, 318 Developmental circumstances developmental delay, 46, 47, 49 developmentally enabled, 54 low human capital, 46, 47, 52 overwhelmed, 53 resource poor non-English speaking, 47, 48 working poor, 46, 47 Developmental patterns declining, 45 fuctuating, 45 improving, 45 middle high, 45 stable low, 45 Developmental stage, 98, 224, 240, 313 Disability care, 21, 194 Disadvantage, 3, 6, 8, 10–12, 21, 22, 25–27, 37–54, 58, 62–66, 78, 82, 83, 90, 93, 97, 115, 120, 121, 133, 134, 137, 138, 140, 144–146, 149, 150, 158, 181, 184, 185, 196, 202, 204, 209, 214, 217, 218, 224, 226–230, 236, 237, 240, 251–274, 281–283, 298–309, 313–317 Disadvantaged backgrounds, 9, 38, 97, 99, 100, 114, 115, 121, 122, 134, 138, 139, 231 Divorce, 23, 61, 202, 280, 290–292 Drinking, 160, 161

#### **E**

Early Childhood Education and Care (ECEC), 37–39, 49 Early Years, 8, 18, 37–54 Ecological systems theory, 298

Economic security, 196, 202 Education, v, 2, 6, 7, 9–11, 18, 23, 25, 26, 30, 37–39, 41, 46–48, 51, 52, 58, 60, 62, 72, 75–78, 80, 82, 83, 87, 88, 93, 97–101, 103, 104, 106, 107, 109–114, 118, 121, 122, 135–138, 140, 141, 144, 146, 147, 151, 180–182, 185, 191, 192, 196, 204, 207, 208, 224, 227, 230, 237, 253, 255, 271, 272, 274, 282, 293, 300, 301, 309, 314–316, 318 Educational disadvantage, 135, 137 Effect size, 38, 39, 41, 42, 44, 103, 113 Elder care, 180, 194 Emergency funds, 11, 169, 212–214, 216, 217 Emerging adulthood, 2, 10, 11, 157–173, 203, 239, 240, 271, 308, 314 Employment job, 24, 178, 181, 184, 188, 190, 191, 318 End users, 304, 306 Engagement and impact (EI), 67, 304, 306 Ethnicity, 206, 315

Evidence based movement, 309–310

#### **F**

Familial support, 280, 285, 286, 291 Families, v, vi, 1–13, 17–30, 37–39, 41, 46–54, 58–66, 71–73, 75, 79–83, 87–89, 92, 98–102, 104, 107, 108, 111–113, 115, 117–122, 133–137, 139–141, 144, 146, 157, 159, 161, 163, 165, 171, 173, 177–197, 201–204, 209, 224–228, 230–236, 239, 240, 251–255, 257, 264–266, 271–274, 280, 283–286, 288, 289, 291–293, 297–310, 313–315, 317 Families of choice, 272 Family disadvantage, 298 Family formation, 181, 189–191, 193, 194, 304 Family functioning and wellbeing, 300 Family policy, 185, 188, 300, 314 Family relationships, 2, 82, 202, 256, 285, 292, 308 Female labour force participation, 177, 181, 185, 188, 189 Fertility, 181, 189–191, 193, 194, 196, 197 Financial support, 6, 21, 143, 159, 161, 163–171, 173, 237, 314 Financial wellbeing, 168–170 Formal social groups, 287 Fundamental goal, 307, 309

#### **G**

Gay, 11, 251, 252, 254, 255, 257, 258, 264–271, 273, 315 Gender, 3–5, 9, 12, 25, 27, 30, 41, 61, 63, 73, 77–79, 83–85, 87, 90, 92, 93, 104, 107–109, 113, 114, 142, 143, 160, 184, 188, 194–196, 205–208, 210–214, 217, 218, 224, 252, 256, 264, 267–270, 272, 315 Geographical remoteness, 62 Golden triangle, 305

#### **H**

	- 73–78, 81

#### **I**

	- 237, 300

Income tax, 21, 186


#### **J**

Job quality, 192, 193

#### **L**

	- 11, 251–274, 315

#### **M**


#### **N**

National Assessment Program-Literacy and Numeracy (NAPLAN), 48, 51

#### **O**

Opportunity pluralism theory, 137

#### **P**


Index

Policy and services, 50–52, 238 Policy impact, 12, 299, 306 Policy research, 298–304, 307–309 Poverty, 4, 5, 10, 27, 28, 159, 196, 197, 224, 234, 237, 282, 299, 301–304 Power, vi, 30, 196, 257, 274, 287, 314 Prevention, 37–54, 196, 293, 300, 301 Primary carer, 59–61, 98, 185, 194, 195 Public value, 307

#### **Q**

Qualitative research, 73

#### **R**

Race, 104, 138, 206, 315 Reczek, 256, 257, 272, 274 Refugee children, 9, 28, 71–93 Refugees, 4, 6, 7, 9, 18, 27, 28, 71–93, 150, 308, 315, 316 Relationship disruptions, 291 Relationship quality, 160, 203, 252, 273, 281 Relocation, 12, 280, 285, 288–293 Risk circumstances, 45, 46, 48–50, 115, 224, 232, 282 clusters, 41, 45 cumulative, 40, 228 models, 40, 50, 62, 111, 113, 228, 316 prediction, 40–42, 50 Riskind, 252, 264, 268, 272, 273

#### **S**

Same-sex couples, 264, 266 Same-sex marriage, 203, 251, 266 School attendance, 6, 8, 49, 52 School belonging, 99, 101, 105, 107, 113, 121–122 Self-concept academic ability, 102, 104, 105 Service use, 49, 50, 52 Single, 5, 7, 10, 11, 135, 139, 150, 163, 178, 184, 187, 189, 202, 205, 207, 209–213, 215–218, 225, 232, 234, 252, 283, 286, 299, 307, 314 Single parents, 5, 21, 25, 26, 28, 38, 39, 165, 166, 170, 171, 173, 178, 184, 187, 189, 197, 201, 202, 225, 234 Smalltalk, 235 Smoking, 41, 160, 161, 168

Social and emotional wellbeing, 8, 57–67, 98, 255 Social development, 282 Social disadvantage, v, vi, 3, 4, 7, 8, 11, 12, 17–30, 78, 100, 112, 122, 136, 202, 204, 205, 209–214, 218, 223–240, 315, 317, 318 Social Learning Theory, 224, 225 Social network, 71, 83, 92, 105, 109, 184, 280–283, 285–293 Social policy reforms, 302, 303 Social security program, 302 Social ties, 12, 20, 234, 280, 281, 285, 286, 289 Socio-economic background, 10, 136, 138–141, 143–146, 149, 150 Socioemotional selectivity theory, 292 Solution based policy, 307 Strengths and Diffculties Questionnaire (SDQ), 61–64, 80, 83, 84, 88, 91, 101, 258 Strengths-based discourse, 9, 66 Structural stigma, 251, 252, 315 Student life cycle, 133, 137–139, 149–151

#### **T**

Tax/taxation/tax and transfer systems, 186, 188, 189, 196, 197, 300, 302 Tornello, 252, 264, 268, 272, 273 Trajectories, 10, 18–20, 41, 72, 98, 114, 120, 135–138, 146–151, 192, 202, 252, 253, 261, 271, 274, 281, 315 Transition to adulthood, 26, 28 Translation framework, 306 Triple P, 235, 236, 238 Turning points, 11, 98, 135, 151, 252, 271, 272

#### **U**

Unemployment, 4, 25, 47, 83, 106, 177, 189, 190, 192, 195, 204, 207, 209, 234, 299–304

#### **V**

Vocabulary, 8, 40–45, 47–50, 52

#### **W**

Wealth, 7, 22, 23, 25, 29, 110, 159, 168–170, 214, 233, 253, 271, 281


#### **Y**

Young adults, 2, 9, 158–163, 165, 168, 169, 171, 193, 279