+ All Categories
Home > Documents > Parent and child wellbeing and the influence of work and family ... · Improving the lives of...

Parent and child wellbeing and the influence of work and family ... · Improving the lives of...

Date post: 21-Feb-2019
Category:
Upload: leminh
View: 213 times
Download: 0 times
Share this document with a friend
157
Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence of work and family arrangements: a three cohort study LYNDALL STRAZDINS 1 NINA LUCAS 1 MEGAN SHIPLEY 2 REBECCA MATHEWS 1 HELEN BERRY 1 BRYAN RODGERS 3 ANNA DAVIES 1 1 National Centre for Epidemiology and Population Health, Australian National University (ANU) 2 Department of Families, Housing, Community Services and Indigenous Affairs 3 Australian Demographic & Social Research Institute, ANU
Transcript
Page 1: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

Improving the lives of Australians

Social Policy Research Paper No. 44

Parent and child wellbeing and the influence of work and family arrangements:

a three cohort study

LYNDALL STRAZDINS1

NINA LUCAS1

MEGAN SHIPLEY2

REBECCA MATHEWS1

HELEN BERRY1

BRYAN RODGERS3

ANNA DAVIES1

1 National Centre for Epidemiology and Population Health, Australian National University (ANU)2 Department of Families, Housing, Community Services and Indigenous Affairs

3 Australian Demographic & Social Research Institute, ANU

Page 2: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

© Commonwealth of Australia 2011

ISSN 1833‑4369

ISBN 978‑1‑921647‑98‑7

This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced by any process without prior written permission from the Commonwealth available from the Commonwealth Copyright Administration, Attorney‑General’s Department. Requests and inquiries concerning reproduction and rights should be addressed to the Commonwealth Copyright Administration, Attorney‑General’s Department, Robert Garran Offices, National Circuit, Barton, ACT 2600 or posted at <http://www.ag.gov.au/cca>.

The opinions, comments and/or analysis expressed in this document are those of the authors and do not necessarily represent the views of the Minister for Families, Housing, Community Services and Indigenous Affairs or the Australian Government Department of Families, Housing, Community Services and Indigenous Affairs, and cannot be taken in any way as expressions of Government policy.

Refereed publication

Submissions to the department’s Social Policy Research Paper series are subject to a blind peer review. This series is recognised by the Australian Research Council’s (ARC) Excellence in Research for Australia Ranked Journal List of refereed journals.

Acknowledgements

We thank and acknowledge Tamar Oz for assistance with descriptive analyses and Sharryn Sims for assistance with information on measures and scales and for providing editorial assistance. We are deeply indebted to Cheney Brew for her calm and professional editing and formatting of the report. Thanks also to Tanya Caldwell and Mark Clements for their expert statistical advice. Finally, we thank Kim Wisener, Tamara Blakemore, Ruth Ragless, Jean Gifford, Justine Gibbings, Susan Garner and Jonette McDonnell—FaHCSIA staff, past and present—who provided input, expertise and helpful suggestions. Please note all views expressed in this document are those of the authors and do not reflect FaHCSIA or Government views.

This paper uses unit record data from the Household, Income and Labour Dynamics in Australia (HILDA) survey and Growing up in Australia: the Longitudinal Study of Australian Children (LSAC). Funding for these data collections is by the Australian Government Department of Families, Housing, Community Services and Indigenous Affairs (FaHCSIA). The HILDA survey is managed by the Melbourne Institute of Applied Economic and Social Research (MIAESR) and LSAC is conducted in partnership between FaHCSIA, the Australian Institute of Family Studies (AIFS) and the Australian Bureau of Statistics (ABS). Source information for LSAC has been compiled using information from the AIFS user guide and the AIFS rationale document.

Further information about the LSAC study can be found at <www.aifs.gov.au/growingup/data.html>.

Further details on the HILDA survey can be found at the website: <www.melbourneinstitute.com/hilda>.

The research reported in this paper was completed under FaHCSIA’s Social Policy Research Services Agreement (2005–09) with the Family and Community Health Research Unit (FCH) at The Australian National University. The views expressed in the report are the views of the researchers only and hence do not reflect the views of FaHCSIA, the Australian Government, AIFS or MIAESR.

Blakemore, Strazdins & Gibbings (2009) gave permission to use their socioeconomic position (SEP) measure. We thank and acknowledge the contribution of FaHCSIA, especially Tamara Blakemore and Justine Gibbings, to the study.

For more information

Research Publications UnitResearch and Analysis BranchAustralian Government Department of Families, Housing, Community Services and Indigenous AffairsPO Box 7576Canberra Business Centre ACT 2610Phone: (02) 6244 5458Fax: (02) 6133 8387Email: [email protected]

Page 3: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

iii

CONTENTS

ContentsExecutive summary vii

1 Introduction 11.1 Changing families 11.2 Changing jobs 31.3 Parent and child wellbeing 61.4 Work and parent wellbeing 91.5 Parents’ work and children’s wellbeing 101.6 Summary of key issues 121.7 Conceptual framework 131.8 Research questions 15

2 Sample and measures 172.1 Datasets methodology 172.2 Methodology for current study 192.3 Caveats 252.4 Parent and child wellbeing measures 272.5 Measures of employment arrangements and parent work conditions 292.6 Family characteristics and sociodemographic measures 32

3 Families across the cohorts: sociodemographic characteristics and wellbeing 373.1 Overview and aims 373.2 Methods 373.3 Family characteristics and socioeconomic demographics 383.4 Sociodemographic differences by disadvantage 403.5 Employed parents’ wellbeing 453.6 Children’s wellbeing 473.7 Summary 50

4 Mothers’ and fathers’ experience of work 534.1 Overview and aims 534.2 Do mothers and fathers work in jobs with similar conditions? 544.3 Breadwinner and dual‑earner families across the three cohorts 584.4 Breadwinner and dual‑earner work conditions 614.5 Gender equality in the workforce: family time as a resource 634.6 Summary 65

5 Work and wellbeing 675.1 Overview and aims 675.2 Methods 685.3 Wellbeing and employed mothers’ and fathers’ work hours and conditions 695.4 Parents’ work hours and conditions and children’s wellbeing 715.5 Parent and children’s wellbeing 735.6 Summary 73

6 Vulnerable families 756.1 Overview and aims 756.2 Methods 766.3 Jobs in low‑SEP and lone‑mother families 786.4 Work hours and conditions by socioeconomic position and family type 786.5 Low‑SEP families and lone mothers and job conditions 836.6 Interactions with socioeconomic position 836.7 Interactions with family type 906.8 Summary 92

Page 4: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

iv Social Policy research Paper No. 44

List of tablesTable 1: Comparison of employed mother and father sub‑samples against ABS data 23Table 2: Employed mothers: sub‑samples in cohort analyses 26Table 3: Family characteristics of employed mothers and fathers 39Table 4: Employed mothers’ and fathers’ family characteristics by socioeconomic position:

within‑cohort comparisons 41Table 5: Employed mothers’ and fathers’ socioeconomic characteristics by socioeconomic position:

within‑cohort comparisons 42Table 6: Employed mothers’ family characteristics by family type: within‑cohort comparisons 44Table 7: Employed mothers’ and fathers’ wellbeing 46Table 8: Child wellbeing across cohorts 48Table 9: Child wellbeing: within‑cohort comparisons by socioeconomic position of employed

mothers and fathers 48Table 10: Child wellbeing by mothers’ family type: within‑cohort comparison 49Table 11: Infants’ and preschoolers’ wellbeing by mothers’ earner arrangement:

within‑cohort comparison 50Table 12: Employed mothers’ and fathers’ work hours and conditions by cohort, with comparisons

across cohorts 55Table 13: Intercorrelations of job conditions for mothers and fathers by cohort 57Table 14: Family characteristics by earner arrangements: within‑cohort comparison 59Table 15: Socioeconomic characteristics by mothers’ and fathers’ earner arrangements:

within‑cohort comparison 60Table 16: Mothers’ and fathers’ work hours and conditions by earner arrangements 62Table 17: Mothers’ and fathers’ wellbeing by work hours and conditions: summary of linear

multiple regression analyses 70Table 18: Association between children’s wellbeing and mothers’ and fathers’ work hours and

conditions: summary of linear regression analyses 72Table 19: Associations between child and parents’ wellbeing scores: summary of

multiple linear regressions 73Table 20: Employed mothers’ and fathers’ work hours and conditions by socioeconomic position:

within and across‑cohort comparisons 82

7 Work and family wellbeing—what does it mean? 957.1 Families across the cohorts 957.2 Mothers’ and fathers’ jobs—different or equal? 967.3 Work and parent and child wellbeing 977.4 Vulnerable families 987.5 Limitations 987.6 Conclusion 99

Appendixes Appendix A: Missing data analysis 101Appendix B: Technical appendix—study strata, clustering and weighting 103Appendix C: Employed mothers on leave 107Appendix D: Scale items 109Appendix E: Work and parent wellbeing: full regression models 111Appendix F: Work and child wellbeing: full regression models 115Appendix G: Parent wellbeing and child wellbeing: full regression models 121Appendix H: Parent and infant wellbeing by socioeconomic position: summary regression models 125Appendix I: Parent and infant/child wellbeing by family type: summary regression models 129

Endnotes 133

references 135

Page 5: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

v

CONTENTS

Table 21: Employed mothers’ work hours and conditions by family type: within and across‑cohort comparisons 83

Table 22: Association between employed mothers’ psychological distress and mothers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions 84

Table 23: Association between employed fathers’ psychological distress and fathers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions 85

Table 24: Standardised beta coefficients for the effect of parent wellbeing, socioeconomic position and interactions on infant and preschooler wellbeing 87

Table 25: Association between infants’ irritability and employed mothers’ work hours by socioeconomic position: adjusted means (standard errors) and interaction 88

Table 26: Association between preschoolers’ SDQ (total) scores and employed mothers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions 88

Table 27: Association between infants’ approachability and employed fathers’ flexible work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions 89

Table 28: Association between preschoolers’ SDQ (total) scores and employed fathers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions 90

Table 29: Associations between employed mothers’ distress and employed mothers’ work hours and conditions by family type: adjusted means (standard errors) and interactions 91

Table 30: Standardised beta coefficients for the effect of parent wellbeing, socioeconomic position and interactions on infant and preschooler wellbeing 92

Appendix tablesTable A1: Number of parents with complete data in each cohort 101Table A2: Means (and standard deviations) for employed mothers with and without complete data

(infant cohort) 101Table A3: Means and (standard deviations) for employed fathers with and without complete data

(infant cohort) 102Table A4: Means and (standard deviations) for employed mothers with and without complete data

(preschooler cohort) 102Table A5: Means and (standard deviations) for employed fathers with and without complete data

(preschooler cohort) 102Table B1: Influence of mothers’ wellbeing outcomes on children’s outcomes: comparison of outcomes

from original, clustered and stratified, and weighted, clustered and stratified data 103Table B2: Influence of fathers’ wellbeing outcomes on children’s outcomes: comparison of outcomes

from original, clustered and stratified, and weighted, clustered and stratified data 104Table B3: Influence of mothers’ wellbeing outcomes on infants’ irritability: comparison of outcomes

from original, clustered and stratified, and weighted, clustered and stratified data 104Table B4: Influence of fathers’ wellbeing outcomes on infants’ irritability: comparison of outcomes

from original, clustered and stratified, and weighted, clustered and stratified data 104Table B5: Influence of mothers’ wellbeing outcomes on preschoolers’ SDQ scores: comparison of

outcomes from original, clustered and stratified, and weighted, clustered and stratified data 105Table B6: Influence of fathers’ wellbeing outcomes on preschoolers’ SDQ scores: comparison of

outcomes from original, clustered and stratified, and weighted, clustered and stratified data 105Table B7: Interaction between fathers’ self‑employment and socioeconomic position on fathers’

psychological distress (infant cohort) 105Table B8: Interaction between mothers’ job control and family type on mothers’ psychological distress

(infant cohort) 106Table B9: Interaction between mothers’ flexible hours and family type and psychological distress

(preschooler cohort) 106Table C1: Employed mothers on leave 107Table E1: Mothers’ psychological distress (K6) and job conditions—LSAC infant cohort 111Table E2: Mothers’ psychological distress (K6) and job conditions—LSAC preschooler cohort 112Table E3: Mothers’ mental health (SF‑36) and job conditions—HILDA adolescent cohort 112Table E4: Fathers’ psychological distress (K6) and job conditions—LSAC infant cohort 113Table E5: Fathers’ psychological distress (K6) and job conditions—LSAC preschooler cohort 113

Page 6: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

vi Social Policy research Paper No. 44

Table E6: Fathers’ mental health (SF‑36) and job conditions—HILDA adolescent cohort 114Table F1: Infant approachability (STSI) and mothers’ job conditions—LSAC infant cohort 115Table F2: Infant irritability (STSI) and mothers’ job conditions—LSAC infant cohort 116Table F3: Child behavioural and emotional problems (SDQ) and mothers’ job conditions

—LSAC preschooler cohort 116Table F4: Infant approachability (STSI) and fathers’ job conditions—LSAC infant cohort 117Table F5: Infant irritability (STSI) and fathers’ job conditions—LSAC infant cohort 118Table F6: Child behavioural and emotional problems (SDQ) and fathers’ job conditions

—LSAC preschooler cohort 119Table G1: Infant approachability (STSI) and mothers’ psychological distress (K6)—LSAC infant cohort 121Table G2: Infant irritability (STSI) and mothers’ psychological distress (K6)—LSAC infant cohort 121Table G3: Child emotional and behavioural issues (SDQ) and mothers’ psychological distress (K6)

—LSAC preschooler cohort 122Table G4: Infant approachability (STSI) and fathers’ psychological distress (K6)—LSAC infant cohort 122Table G5: Infant irritability (STSI) and fathers’ psychological distress (K6)—LSAC infant cohort 123Table G6: Child emotional and behavioural issues (SDQ) and fathers’ psychological distress (K6)

—LSAC preschooler cohort 123Table H1: Association between infants’ approachability scores and employed mothers’ work hours and

conditions by socioeconomic position: adjusted means (standard errors) and interactions 125Table H2: Association between infants’ irritability and employed mothers’ work hours and conditions

by socioeconomic position: adjusted means (standard errors) and interactions 126Table H3: Association between infants’ approachability and employed fathers’ work hours and

conditions by socioeconomic position: adjusted means (standard errors) and interactions 127Table H4: Association between infants’ irritability and employed fathers’ work hours and conditions

by socioeconomic position: adjusted means (standard errors) and interactions 128Table I1: Association between infants’ approachability and employed mothers’ work hours and

conditions by family type: adjusted means (standard errors) and interactions 129Table I2: Association between infants’ irritability and employed mothers’ work hours and conditions

by family type: adjusted means (standard errors) and interactions 130Table I3: Association between preschoolers’ SDQ (total) scores and employed mothers’ work hours and

conditions by family type: adjusted means (standard errors) and interactions 131

List of figuresFigure 1: Basic conceptual model of pathways between parent work and parent and child wellbeing 8Figure 2: Conceptual model of pathways between parent work arrangements and parent and

child wellbeing 14Figure 3: Families with infants, LSAC B cohort (n=3,266): flow chart of parents in each family

employment category with complete data 20Figure 4: Families with preschoolers, LSAC K cohort (n=3,206): flow chart of parents in each

family employment category with complete data 20Figure 5: Families with adolescents, HILDA survey (n=360): flow chart of parents in each family

employment category with complete data 21Figure 6: Household work hours per week by earner arrangements per cohort 64Figure 7: Mothers’ and fathers’ work hours per week by earner arrangements per cohort 65Figure 8: Proportion of mothers’ and fathers’ working part time, full time or long full‑time hours by

socioeconomic position, in each cohort 79Figure 9: Proportion of mothers and fathers in each type of employment tenure by socioeconomic

position, in each cohort 80Figure 10: Employed fathers’ psychological distress by self‑employment and socioeconomic position

(infant cohort) 86Figure 11: Employed mothers’ distress by job control and family type (infant cohort) 92

Page 7: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

vii

ExECuTIVE SuMMAry

Executive summary

Australian families work and raise children in different contexts from those of families 50 or even 30 years ago. Technology, globalisation, flexible and deregulated labour markets, and the unprecedented movement of mothers back into the labour force have rewritten the way work and family interconnect. At the same time, what constitutes adequate parenting has been redefined in novel and demanding ways. New knowledge is needed to understand what the convergence of these changes means for the wellbeing of parents and their children.

Work experiences and parenting expectations have been revolutionised. Most children now grow up within the confines of nuclear families focused on children’s developmental needs. Bringing up children was once the business of large, multigenerational families and their local communities and is now seen as the responsibility of parents. At the same time global competition, increased labour flexibility, and the shift to service and knowledge economies has placed new pressures on working life. Employed parents must combine working with providing care for children and are therefore facing increased expectations and demands from both work and family spheres.

Overview of this study

This study underlines how work and family life are not separate, but closely connected. The nature and quality of parents’ work is linked to children’s outcomes, via its impact on parent wellbeing. Access to good quality jobs (that is, those with some flexibility, security and control) may help optimise parents’ and their children’s wellbeing. Thus what happens at work and the sorts of jobs Australian parents have access to could have consequences across generations.

Using data from Growing Up in Australia: the Longitudinal Study of Australian Children (LSAC) and the Household, Income and Labour Dynamics (HILDA) survey, this report provides new evidence about the influence of work on parents’ and children’s wellbeing. It investigates this link for three cohorts of children (infants, preschoolers and adolescents) and for four different types of families (breadwinner compared with dual‑earner families and lone parent compared with two‑parent families). Differences between relatively disadvantaged and other families are also considered.

In investigating whether parents’ work impacts children’s wellbeing, this report raises important questions about the intergenerational consequences of parents’ work. It also explores variations in the relationship between work and wellbeing at different points in children’s development by examining this relationship in three cohorts of children. Further, by distinguishing different types of families, it recognises that some families may be more vulnerable to particular work conditions than others. Rather than simply considering children’s wellbeing when parents are either employed or unemployed, this report provides a detailed investigation of how the conditions of employed parents’ work impacts on their own and their children’s wellbeing.

We address four main themes in the report:

�� parents’ work experiences

�� parents’ work and parents’ and children’s wellbeing

�� work experiences and wellbeing at three stages of child development

�� work experiences and wellbeing for different types of families.

This study capitalises on unique, world‑first data from LSAC that assess conditions of parents’ jobs as well as work hours and income. Rare in studies of children, it gathers data on both mothers’ and fathers’ jobs and their

Page 8: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

viii Social Policy research Paper No. 44

wellbeing, and thereby allows pathways from parents’ work and wellbeing to be traced through the analyses. The addition of the third age cohort, the adolescent cohort drawn from the HILDA survey, enables us to make preliminary comparisons of work, families and wellbeing across three different child ages.

Parents’ work experiences

Working time as a gendered experience

Employed mothers work fewer hours than do employed fathers, and their hours are modulated according to their children’s age and partners’ working status. Couple mothers tend to work part time when their children are infants or preschoolers, and work more hours when their children reach adolescence. Couple mothers also work fewer hours when their partner is employed. In contrast, breadwinner mothers’ hours remain stable at around 30 hours per week in all cohorts.

Fathers’ work hours show a different pattern. Fathers’ work time is not influenced by age of children or whether their partner is employed or not. Across all three cohorts fathers work longer hours than mothers (including breadwinner mothers), tending to work full‑time or long full‑time hours irrespective of their child’s age or partner’s employment status.

Part‑time work remains the main option for employed mothers in our sample. Even in the adolescent cohort, part‑time work is still the most common work pattern, although the proportions are lower. Part‑time work does not appear to be a viable option for fathers. Irrespective of children’s developmental stage or partners’ work hours, few fathers work part time.

Our findings therefore indicate that mothers and fathers hold down jobs in the context of different time resources. Looking across the cohorts, mothers do not experience the same ‘flexibility’ from their partners’ work hour commitments that fathers do. Mothers work around the care needs of their children and their partners’ jobs while fathers do not appear to do so, at least to the same extent. This finding raises questions about how to achieve greater gender equity in the workforce, especially in the context of Australia’s long full‑time work hours relative to most other Organisation for Economic Co‑operation and Development (OECD) countries. The capacity to put in long hours is a feature of many full‑time jobs, particularly senior jobs. Our data point to a ‘time gap’, which means that, irrespective of their qualifications or capabilities, mothers are much less likely to have the time resources to hold down full‑time, especially senior full‑time, jobs. While the focus of public debate has been on increasing fathers’ time commitments to caring, a complementary focus for gender equity and participation initiatives could be to tackle long work hour expectations for full‑time jobs.

dual earners and breadwinners: different experiences for mothers and fathers

Being the breadwinner (sole income earner) is common for fathers when their children are young, but many two‑parent families appear to shift to dual‑earner arrangements with older children. Proportionally, twice as many dual‑earner families in the adolescent cohort have both parents working full time compared with the infant and preschooler cohorts. This suggests father‑as‑breadwinner arrangements may be relatively short term, linked to children’s development and care needs. In a small proportion of two‑parent families, mothers are breadwinners. The proportions of such mothers remain fairly stable across cohorts, apparently independent of the age of their children.

Proportionally more breadwinner families report financial hardship than do families in which both parents are employed. This hardship gap is particularly marked when mothers are the breadwinners, appearing to widen as the children grow up. One in four breadwinner mothers report financial hardship in families with an infant, compared with one in two in the adolescent cohort.

Page 9: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

ix

ExECuTIVE SuMMAry

Parents’ work and children’s wellbeing

Children’s wellbeing is linked to their parents’ work circumstances. Some of these links are direct: we find poorer wellbeing among infants with mothers in casual employment and with poor job control, compared with infants whose mothers have more favourable work circumstances. Indeed, mothers’ work circumstances contribute at least as much to understanding infant wellbeing as do other established factors such as household income and chronic maternal illness. Infant and preschooler wellbeing is enhanced by having fathers who are self‑employed or who have flexible work hours.

Infants may also benefit from having a parent working less than full‑time hours. Although infants of fathers working very long hours are less irritable than those of fathers working standard hours, many of these fathers have partners who work less than full‑time hours. Controlling for their partners’ work hours removes this association. This finding highlights a need to consider work hours as a family, rather than an individual, resource.

Although some links between parents’ work circumstances and child wellbeing are direct, more often the relationship is indirect: work circumstances affect parents, which then influence children’s outcomes. Unfavourable work circumstances are associated with higher distress in parents, which is in turn linked to poorer wellbeing in their children. Over and above income and hours of work, job security is associated with fathers’ and mothers’ mental health and with fathers’ work‑related wellbeing. For parents of infants and preschoolers, greater job control and flexible hours are linked to less parental distress.

Our findings also suggest that children’s wellbeing could be further optimised if parents had jobs that gave them flexibility in work hours, some security about their job future and some control over the tasks they do. Ensuring jobs contain these three minimal conditions could reduce potential negative trade‑offs parents face between earning income and maximising their children’s wellbeing.

Work experiences and wellbeing: variations by family type and socioeconomic circumstances

Our study indicates that socioeconomic disadvantage is closely linked to both work experiences and wellbeing. Work hours, conditions and type of employment contract are related to socioeconomic circumstances, which are also systematically related to differences among family types. Across all cohorts, parents (particularly mothers) from more disadvantaged circumstances are less likely to have secure jobs and less likely to have jobs that give them control and flexibility compared with more advantaged parents. Importantly, we find evidence that these systematic differences in job security, control and flexibility not only persist but increase as children grow older. Thus, disadvantaged parents are more likely to work in jobs with poorer conditions, for longer, and we also find suggestive evidence that their wellbeing could be more vulnerable to any adverse impacts from these jobs.

Work hours

Our analyses compare families in the bottom quartile of socioeconomic position with more advantaged families. Parents living in families that enjoy more favourable socioeconomic circumstances work slightly more hours per week than do those with less favourable circumstances (three to four hours more per week for mothers and around two hours more for fathers). Fathers in these more favourable circumstances work similar hours in each cohort. In contrast, disadvantaged fathers work longer hours when their children are adolescents than they do when their children are infants or preschoolers. Over one‑third of fathers from disadvantaged families work long full‑time hours when their children are infants and this proportion increases to more than one‑half among fathers of adolescent children.

Page 10: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

x Social Policy research Paper No. 44

Regardless of socioeconomic circumstances or family type, mothers of adolescents work more hours per week and are more likely to work full time than mothers of infants and preschoolers. Mothers living in poorer socioeconomic circumstances work fewer hours than do their more advantaged peers. Lone mothers are more likely to work full time when their children are preschoolers than when their children are infants, suggesting that lone mothers may return to full‑time work relatively quickly after the birth of a child.

Job tenure

Regardless of socioeconomic circumstances or family type, rates of job security are higher among mothers of adolescents than they are among mothers of infants and preschoolers. Fathers follow a similar pattern, with rates of job security increasing, particularly among fathers in more favourable socioeconomic circumstances.

Overall, socioeconomically disadvantaged parents are more likely to be employed casually, more likely to report feeling insecure about their job and less likely to be employed permanently than are their more advantaged peers. This job security disparity is apparent regardless of their children’s age and is particularly marked for mothers. We find a similar pattern with job control. Relatively disadvantaged lone‑mother or two‑parent families are more likely than more advantaged peers to report poor job control. Couple mothers and mothers living in more favourable socioeconomic circumstances report steadily increasing levels of job control with older cohorts of children. Job control does not show a similar increase among lone or relatively disadvantaged mothers.

Working conditions and parental distress

We find consistent links between work conditions and employed mothers’ and fathers’ distress. Couple and lone mothers both report less psychological distress when they have some control over their job tasks. When jobs do not offer mothers such control, lone mothers report relatively higher levels of psychological distress than do couple mothers. Mothers with inflexible work hours report higher levels of distress than do those with flexible hours, regardless of socioeconomic circumstances. Among mothers of preschoolers, the trend towards increased distress when facing inflexible work hours is more pronounced for those experiencing disadvantage than for those with more favourable socioeconomic circumstances. Though this trend does not reach statistical significance, it is consistent with other findings indicating greater vulnerability to work‑related adversity among disadvantaged mothers of preschoolers.

Among fathers from families in disadvantaged socioeconomic circumstances, self‑employed fathers report higher rates of distress than do fathers in other employment categories. This pattern is reversed among fathers living with more favourable socioeconomic circumstances, who report slightly lower distress when they are self‑employed than when they are not. This may be because disadvantaged fathers have poorer quality self‑employment or may be more vulnerable to risk factors associated with self‑employment (such as job insecurity). This finding requires further research.

Summary of main findings

Our study points to three key findings:

�� The quality and conditions of employed parents’ jobs matter to their own and their children’s wellbeing.

�� Socioeconomic disadvantage is persistently linked to poorer access to good quality jobs, especially for mothers.

�� Mothers (but not fathers) modulate their work hours according to their children’s age and partner’s work, and hold down jobs in the context of having partners with heavy work‑time commitments. Mothers’ employment is therefore conducted in a different ‘time context’ to fathers’.

Parents’ employment circumstances are related to their own and their children’s wellbeing and there is considerable variation around the details of this relationship. Some features of jobs affect mothers but not

Page 11: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

xi

ExECuTIVE SuMMAry

fathers, or vulnerable parents but not those living with more favourable socioeconomic circumstances, and so on. Some features of jobs show direct links to children, while other pathways are indirect, through their impact on parents. Despite these subtleties, the main story from this research is that parents’ employment circumstances may have impacts on their own wellbeing and on the wellbeing of their children.

Insecure jobs, inflexible hours and poor job control are associated with an increase in parental distress. Parents’ unfavourable employment experiences may undermine children’s wellbeing—largely because they erode parents’ wellbeing. Conversely, ensuring that jobs give parents some flexibility, control and security may help optimise parents’ and children’s wellbeing, yielding both short‑term and long‑term benefits to families.

Negative employment experiences and conditions appear to be particularly damaging for vulnerable parents—lone mothers and parents in unfavourable socioeconomic circumstances—who already live with underlying disadvantage. These are also the people who are most likely to encounter poor employment circumstances. The greater the pressures on parents, the greater is the potential harm to children’s wellbeing, especially if, as our findings suggest, underlying disadvantage might interact with poor employment circumstances. Positive employment circumstances for parents could deliver clear advantages to children’s wellbeing and development, particularly to disadvantaged children.

Future research

Future research should extend and address some of the limitations of this study. We analysed data for employed mothers and fathers separately, an essential first step in understanding families, work and wellbeing over the life course. However, in couple families, mothers’ and fathers’ work decisions and arrangements are closely interlinked, and treating families as a unit would reveal the dynamics of work and family between parents.

Our study points to the wider impact on children of work and family strain. As both work hours and child‑rearing responsibilities have increased in recent years, further studies could investigate the strategies parents want to use, and the constraints they face, in handling these competing responsibilities.

Possible policy implications

Maximising and supporting workforce participation is a national priority and policy objective that is essential to support Australia’s ageing population. Also critical to Australia’s social and economic future is the full development of children’s mental, physical and academic capabilities. Jobs, parents, children and family life lie at the centre of these two priorities. This study identifies what aspects of work may serve to promote parent and child wellbeing, thereby optimising the benefits of employment for families and the economy.

Little large‑scale epidemiological research has investigated the relationship between parental employment circumstances and child wellbeing, certainly not in the detail contained in this report. In this respect, we present new knowledge for inclusion in policy debate about child development and wellbeing.

First, our findings point to the close interlinking of employment and family policy. Our findings suggest that improving job security, access to flexible working arrangements and opportunities for control over work could optimise the wellbeing of all employed parents. Second, our findings suggest that strategies to optimise the wellbeing of children—all children—would include improving their parents’ working arrangements and reducing the extent that work places strains on family life. Third, even when employed, vulnerable parents do it tough. Added to other hardships, they are more likely than other parents to encounter poor employment circumstances and these circumstances appear to hit them harder. The children of vulnerable parents are also vulnerable, largely because their parents are.

Our research suggests that there may be ways to short‑circuit such compounding jeopardy: vulnerable parents (and therefore their children) may be particularly responsive to supportive employment circumstances. Raising minimum standards would ensure that even the most vulnerable are employed under more flexible conditions, with better job control and security. However, our findings also suggest that wages and income support

Page 12: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

xii Social Policy research Paper No. 44

remain a critical part of the picture; even when employed, many lone mothers still report high rates of financial hardship.

A number of other key findings hold important clues to work–family policy development. Job control, flexible hours and job security have implications for mental health. Mothers tend to hold less secure and poorer quality jobs than do fathers, and this inequality occurs regardless of the age of their child. Flexible hours are important for both mothers’ and fathers’ wellbeing, contrary to popular perceptions that flexibility is most beneficial for mothers. It may be also that fathers’ jobs are worked in a more ‘time‑rich’ context than are those of mothers, particularly when the mother in a two‑parent family is not in paid employment, or is working part time.

Further, our findings suggest that work hours would most appropriately be conceptualised at a household or family level, rather than at an individual level. Work programs and job design would benefit from recognising that mothers do not have access to the same time resources as fathers; full‑time work, as it is currently configured, may embed time‑based gender discrimination where long hours are required or expected.

Policy advances in addressing any of these employment‑related factors could be expected to reduce parental distress. Our research indicates that any such advances would, in turn, flow on to improve the wellbeing of children.

Page 13: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

1

INTrOduCTION

1 Introduction

Australian families now live and work in entirely different contexts to families 50 or even 30 years ago. Technology, globalisation, the shift to service and knowledge work, flexible and deregulated labour markets and the unprecedented movement of mothers back into the labour force have rewritten the way work and family interconnect. New knowledge is needed to understand what this convergence might mean for the wellbeing of parents and their children. This study aims to provide new evidence on this interplay.

How can families avoid trade‑offs between earning income and protecting and investing in wellbeing? How are children faring and can some jobs optimise family wellbeing as well as support family income and social inclusion? To answer these questions we introduce children’s wellbeing into our analysis of work and family, asking whether children also have a stake in the interconnections between work, families and wellbeing. These analyses raise important questions regarding the intergenerational impacts parents’ work may have. Families change as children develop so the way jobs relate to wellbeing may also vary over the life course. We consider children’s (and their parents’) wellbeing according to three child developmental stages—infancy, preschool and adolescence—and ask if jobs promote parent and child wellbeing differently in three cohorts of families spanning these three stages. Instead of asking whether it is better for mothers or fathers to work or not, our study seeks to discover when work matters to parents’ and their children’s wellbeing and which sorts of families are most affected.

Finally, this report extends current thinking about families and employment by developing a detailed investigation of the conditions of parents’ work. Much of the research on employment and wellbeing has compared families with jobs to unemployed families. Clearly, jobs benefit families by delivering income, and jobless families are disadvantaged. But not all jobs are equal. The quality of the job, the conditions of work and the income delivered may also be important drivers of how families fare. We consider three aspects of jobs linked to management practices or industrial relations policies—job control, security and work hours (including flexible start and stop times)—and ask if these are additional, critical ingredients for balancing work and family life.

This report is based on analyses of two large national studies of Australian families: Growing Up in Australia: the Longitudinal Study of Australian Children (LSAC) (the infant and preschooler cohort) and the Household, Income and Labour Dynamics in Australia (HILDA) survey (a selected, adolescent cohort). We wish to acknowledge and thank all those involved in study funding, design and data collection. We also wish to state that our current cross‑sectional design does not enable us to draw causal conclusions, but provides cross‑sectional evidence on the patterns among the elements of work, families and wellbeing.

1.1 Changing families

Over the past few decades, Australian families have changed and so have the pressures they experience. These changes form the backdrop to questions of how work may influence family wellbeing and the context for the families analysed in this study.

Changing care and work arrangements

The Australian economy has moved from its agrarian beginning to a post‑industrial service economy. Paid work is no longer located in homes (with the exceptions of self‑employed and teleworkers) and so child care and employment rarely occurs in the same physical location (Glass 2000). The rise of the breadwinner wage model at the turn of the twentieth century created one solution to work and care giving responsibilities. The different locations were divided along gender lines, with mothers at home in charge of care giving and fathers away from home engaged in work (Glass 2000). Under a breadwinner model, workers who were parents

Page 14: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

2 Social Policy research Paper No. 44

(that is, fathers) were relatively free from care giving responsibilities because their wage supported their spouse to run the home and raise children.

This solution eventually became unstable. Wages for working‑class men were rarely high enough to support the whole family (hence the secondary labour market of working‑class women and children) and mothers without partners were unable to support themselves or their children (Glass 2000). The breadwinner model inadvertently entrenched gender inequality by making financial independence more difficult for mothers and tied care to non‑participation in the workforce. Consequently, mothers began to return to the labour force. For example, since the 1980s, employment increased among mothers with dependent children from 46 per cent to 72 per cent (Cass 2002; Murphy 2002). The majority of mothers in two‑parent families are now back in paid work by the time a child goes to school (ABS 2000) and more mothers are returning to work before their children are at school. In 2000, 45 per cent of Australian women with children aged under 4 years were in the paid workforce compared with 29 per cent in 1985 (OECD 2002).

Uncertainty about jobs and finances linked to globalisation (Edwards 2001) and labour shortages due to population ageing have continued the pressure for mothers to remain in the labour force. In the context of high housing prices, Australian mothers’ wages have now become an integral component of most middle and working‑class family incomes (Cass 2002). Like other Western nations, dual‑earner families (where both mothers and fathers are employed while raising children) are now the most common Australian family type (ABS 2002; Dex 2003; Marshall 1993; US Department of Labor 1999).

Changing childhoods and child rearing

Perhaps less obviously, children’s places in family and community life, as well as the tasks and costs of parenting, are being redefined. Compared with a century ago, family size has reduced, due to better survival rates of infants and children and the use of contraception. Despite this shift, having fewer children has not meant that parents have become less invested in child care.

The shift away from an agrarian economy and household production has meant that children are no longer part of a productive unit, able to contribute to family resources (Glass 2000). Instead, providing care for children (in economic terms) places opportunity and financial ‘costs’ on parents who must trade‑off time in the labour market for time spent care giving (or purchase child care to enable labour market work). Children are now dependent on their parents for longer periods of time, because children must spend more time in education garnering the skills needed for knowledge and service jobs. Further, cultural understandings and expectations of ‘good parenting’ have intensified. Drawing on theories of attachment, child development and mental health, parenting has become an increasingly skilled and intensive endeavour.

Furthermore, the context of child rearing has changed. For example, parents must filter or monitor information from mass media or the internet that previous generations did not encounter. Increased fears for children’s safety and more traffic have led to closer parent supervision of children in their neighbourhoods. Meanwhile parents and children live in ‘obesogenic’ environments where energy dense food is readily available and children’s consumption must be monitored if obesity is avoided. All this requires a time investment and most responsibility for children’s safety, health and behaviour is seen to reside with parents (Glass 2000). Mothers’ re‑entry into the labour market and the dual‑earner family has therefore occurred in tandem with new notions of children’s needs and ‘good enough’ parenting.

Contemporary fatherhood is also being redefined as active and engaged, which may partly be driven by mothers’ labour force participation. The amount of involvement fathers have with their children is influenced by when their partner enters the workforce after having children (Gottfried & Gottfried 2006). Once this pattern was established it is sustained into their child’s adolescence (Gottfried & Gottfried 2006). Thus, employed mothers and fathers must manage potentially opposing pressures, maintaining their work force participation while meeting the considerable expectations, investments and responsibilities of being a contemporary parent.

Page 15: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

3

INTrOduCTION

The transformation of family time

Economic pressures may be relieved by working, but both paid work and providing care for children takes time. Three decades ago in the US, when most two‑parent families had mothers at home, the time devoted to paid work (by fathers) averaged 45 hours per week. Now, as dual‑earner families become much more common, the time devoted to paid work by both parents averages 82 hours per week (Jacobs & Gerson 2004). In Australia the figure is closer to 75 hours per week (Drago, Tseng & Wooden 2004). Being a parent adds an additional two hours unpaid work a day to both mothers’ and fathers’ total workloads compared to non‑parents (Craig 2002). If time is considered a resource, then parents compete in the labour market with different time resources they can commit to working. Mothers’ workloads (combining paid and unpaid work) average 9 hours a day and fathers’ 9.2 hours a day compared with childless women who work 7.13 hours and childless men 7.27 hours a day of either paid or unpaid work (Craig 2002). Once secondary activities (combining two or more activities, such as cooking dinner and minding children) are also included, mothers’ total workload becomes even higher (Craig 2002).

The way families manage these ‘time loads’ tends to be gendered. Parenthood increases the time men spend in paid work, with fathers working on average 1.8 hours more per day than childless men (Craig 2002). Interestingly, fathers’ increased allocation of time to paid work persists over the life course. Fathers’ workplace participation (both part time and full time) stays virtually the same regardless of the stage of their child’s development (ABS 2006). The opposite is true for women, with mothers in paid work for about one hour less per day compared with childless women (Craig 2002). Mothers also appear to adjust their work hours and participation according to their children’s age (Gottfried & Gottfried 2006).

Parenthood leads to an increase in time spent in domestic tasks (cleaning, shopping, cooking) for mothers while decreasing it for fathers (Craig 2002). Mothers cut back on paid work to enable them to devote time to raising children and running the house, so their allocation of time to these activities varies over the life course (Craig 2002).

Although time pressures affect all parents, they are especially acute for two groups of parents—mothers with a partner who are employed full time and employed lone mothers. Adding together paid and unpaid work, the total time demands on full‑time employed parents during the working week is 14 hours per day for mothers and 12 hours per day for fathers (Bittman & Craig 2005). Child care helps but does not solve the problem because parents still strive to preserve time with their children (Sayer, Bianchi & Robinson 2004). What is lost is access to free or discretionary time. Among two‑parent families, full‑time employed mothers are the most time‑poor and this time pressure is related to poorer mental health (Roxburgh 2004). However, a recent report using the LSAC data found that employed lone mothers were even more likely to report time pressures (Baxter et al. 2007). For example, nearly eight out of 10 full‑time employed lone mothers were often or always rushed for time, compared with six out of 10 full‑time employed couple mothers and about five out of 10 full‑time employed fathers (Baxter et al. 2007).

Having children affects career advancement of mothers and this negative impact is strongest when women have younger children (Jones & Dickerson 2007). Thus time constraints may be one of the underlying determinants of the ‘glass ceiling’ that women, especially mothers, encounter. For many mothers, especially lone mothers, inability to meet long‑hour expectations attached to senior or highly paid jobs may result in an inadvertent but subtle form of inequity and exclusion, and this may be one reason women are under‑represented in senior roles in this country.

1.2 Changing jobs

At the same time that families are increasing their workforce participation, the nature and conditions of jobs available to them are changing. Global competition, technological advances and the shift to a service and knowledge economy, coupled with increased labour flexibility and the deregulation of the labour market, have led to more polarised labour markets in many developed economies (Grzywacz & Dooley 2003).

Page 16: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

4 Social Policy research Paper No. 44

Job control and job insecurity

Australia, like other developed nations, has achieved an unprecedented standard of living. The shift to service and knowledge economies has occurred in tandem with technological changes and new ideologies of management (Green 2006). The experience of working has changed and there is now a greater need for jobs to be secure (Glass 2000). Technology has led to the automatisation of many jobs, resulting in job loss for some low‑skilled workers. Globalisation has made it easier for business to move across nations, placing downward pressure on wages and conditions for their workers. Consequently, some segments of the labour force are vulnerable to job insecurity because of business closure and relocation offshore.

Some have argued that there has been a steady transfer of economic risk from employers to employees over the past few decades (Hancock 2002), which may increase perceptions of insecurity. Wages have been decentralised and there has been a move from universally legislated to workplace‑negotiated conditions. The ‘flexibilisation’ of the labour market has resulted in the growth in casual employment and increase in contracting‑out work. Downsizing has also affected high‑skilled workers, leading to increased workloads and the speeding up of effort and pace (termed ‘work intensification’). Australia has one of the highest proportions of casual or contingent employees in the Organisation for Economic Co‑operation and Development (OECD) with 27 per cent of the workforce employed in this way, compared with 19 per cent a decade ago (ABS 1999). These trends mirror those of other developed nations. For example, temporary and casual work have risen steadily in the European Union (EU) from 3.9 per cent of the workforce in 1983 to 9.3 per cent in 1991 and 13.3 per cent in 2004 (Weiler 2005).

The combined influence of globalisation, technology and the shift to services has been linked to a more polarised labour market in terms of pay, skill level and control over work hours and tasks (Grzywacz & Dooley 2003). High‑skilled knowledge workers tend to be given greater levels of control at work, including work hours, in return for commitment, responsibility and innovation (Green 2006). On the opposite end of the scale, low‑skilled workers may be subject to inflexible, just‑in‑time or on‑call (employer‑controlled) work, as well as surveillance and scientific management of their work enabled by computerisation and automation (Green 2006). Some Australian parents, especially the low skilled, are likely to be working in insecure jobs, or in jobs that do not allow them a say in how they do their work. However, very few studies have examined job insecurity and job control among employed parents.

hours and intensification: additional time pressures

At the same time that parents are facing increased expectations and responsibilities for raising children, work times and demands have also changed. One hundred and fifty years ago, long work hours were a feature of the Australian economy. It was not until 1856 that an eight‑hour day was legislated and average hours gradually declined up until the early 1980s, when Australian full‑time workers averaged less than 40 hours per week (ABS 1985). The 1970s marked a turning point for work time. Global competition accelerated because of widespread financial deregulation and international movement of capital. At the same time downsizing occurred across public and private enterprise while flexible and insecure employment increased in many developed economies (OECD 2007). In Australia, employees began to work longer and harder and average full‑time work hours began to increase (Quiggin 1999). Now the average Australian full‑time employee works nearly 42 hours per week (ABS 2006).

There has also been an increase in the number of Australian employees working long hours (50 hours or more a week) over the last 20 years (ABS 2006) and a high proportion of these workers are fathers (Goward et al. 2005). In 1985, 22 per cent of male and 9 per cent of female full‑time employees worked 50 or more hours per week (ABS 2006). In 2005 the percentage of those working very long hours rose to 30 per cent of men and 16 per cent of women (ABS 2006). Unlike poor job control or casual employment, long hours are more commonplace in high‑skilled jobs (Gray et al. 2004). Indeed, low‑skilled jobs are much more likely to be part time and therefore involve fewer work hours.

Long work hours do not always translate into more remuneration. Work hour growth is a reflection of increased workloads and work intensification, with one‑third of the Australian labour force putting in unpaid overtime

Page 17: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

5

INTrOduCTION

(Watson et al. 2003). Ninety per cent of overtime is worked by full‑time employees and a large proportion of those regularly working overtime are fathers (ABS 2006). Across European nations work intensity has increased for the past 20 years, with more people working harder, at high speed and under a great deal of tension (Green 2006). Recent surveys have documented an increased pace and intensity of work in EU countries (Houtman 2005; Merllié & Paoli 2001; OECD 2004) and data available for Australia indicates similar trends here (Green 2006; Houtman 2005).

Technological advances such as laptops, email, mobile phones and voicemail have also helped to erode the boundary between work and free time and added to the experience of overload (Galinsky 1999). Thus, some employed parents face the prospect of working in jobs with longer hours and/or in jobs with greater work pace and intensity, while also facing time pressures at home. A US study found a 10 per cent increase in the proportion of parents saying that their job requires that they work very fast between 1977 and 1997 and a 21 per cent increase in the proportion of people agreeing ‘I never have enough time to get everything done’ (Galinsky 1999). Many children are now growing up in households that may be income rich but characterised by busyness, rushing and time poverty. There is little research on how this may shape children’s aspirations and views of working and care giving.

Family‑friendly conditions

In the current study we consider a subset of family‑friendly conditions: access to flexible work hours and part‑time employment. We have not directly examined parents’ access to leave and any associations with parent or child wellbeing. However, we briefly reviewed a range of family‑friendly conditions in this section as they provided a wider context to the discussion on work, families and wellbeing.

Family‑friendly conditions can benefit business. They aid the attraction and retention of high‑quality employees who in turn add to business productivity, service standards and ultimately, competitiveness (Dickens 1999; Gray 2002). Flexible work arrangements and support for personal and family needs are both strong predictors of employee job satisfaction, commitment and retention (Bond et al. 2002). Additionally, supervisor support is related to increased willingness to work harder (Bond et al. 2002). These work conditions also greatly benefit families, society and the economy overall, often in the longer term (Dickens 1999).

The implementation and utilisation of family‑friendly conditions has varied across sectors, by sex and across nations. In contrast to most OECD countries, there have been no universally legislated formal family‑friendly conditions in Australia, except 12 months unpaid parental leave introduced in 1993 as part of the Industrial Relations Reform Act (Baird 2003). In 2001, this provision was extended to cover casuals who had been with the same employer for more than 12 months (Baird 2003). Australia’s first national Paid Parental Leave scheme started on 1 January 2011. It provides eligible working parents with 18 weeks of Parental Leave Pay at the national minimum wage. Although most Australian workplaces have at least some family‑friendly conditions (Whitehouse 2001), they are most widely implemented in the federal and state public sector (ACIRRT 2002; Earle 2002) or in private‑sector industries with high‑skilled employees (Gray & Tudball 2002).

The family‑friendly debate has rarely considered wider issues of low pay, poor job conditions (Dickens 1999) or the widening gap between low and high‑skilled jobs. Like higher wages, family‑friendly conditions have most often been given to employees who are valued the most, not employees who need them the most. For example, flexible work hours are more often a feature of high‑skilled knowledge jobs (Gray & Tudball 2002). Thus low‑skilled or socioeconomically disadvantaged parents may be more in need of family‑friendly conditions and the least likely to have them.

Further, there appears to be implicit, gender‑based constraints in access to and use of family‑friendly provisions. The proportion of fathers that take up leave for child care is low, even in firms recognised as ‘family friendly’ (Bittman, Hoffmann & Thompson 2004). For parents of children under 12 years, mothers (68 per cent) are much more likely to use work arrangements to assist with child care compared to fathers (27 per cent) (ABS 2001a). Additionally, 33.7 per cent of mothers used permanent part‑time work compared to only 1.9 per cent of fathers (Bittman, Hoffmann & Thompson 2004).

Page 18: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

6 Social Policy research Paper No. 44

Bittman, Hoffmann and Thompson (2004) argue that cultural norms of the father as a breadwinner and the mother as a care giver underlie fathers’ reluctance to use family‑friendly provisions. These norms are present in workplace cultures and shape the expectations of employers. Anecdotal evidence show that corporate culture does not view family‑friendly conditions as relevant to men and can even punish care‑giving fathers (Saltzman & Barry 1988). A survey by Meiers (1988, in Bittman, Hoffmann & Thompson 2004) found 41 per cent of managers in companies that had policies offering leave for men believed that men should not take time off for family reasons. So while there may have been an increase in formal flexibility in the workplace, this does not translate into fathers’ capacity to take time off work to provide care for children. Additionally, fathers appear to be less willing to bear the costs involved in care giving for children, such as loss of earnings and reduced career advancement.

Our review highlights several important points. First, most families are moving towards a dual‑earner arrangement, with both mothers and fathers in paid employment while raising children. This has increased income for families, but has dramatically altered the time available for caring and other unpaid work. At the same time expectations for parenting have changed, including time investments required to raise healthy children in a digital, sedentary and affluent society. So too have the nature and conditions of contemporary jobs changed. Over the past two to three decades labour markets have become more flexible and insecure, with increased workloads and pace (work intensification) and the growth of long‑hour jobs. Furthermore, there is evidence that labour markets are polarising into low‑skilled jobs with poor conditions and high‑skilled, long‑hour jobs with more flexible arrangements. There is virtually no research drawing all these elements together and examining the impacts on employed parents and their families. Nor is it known whether some jobs could optimise wellbeing for both parents and children. The next section briefly considers evidence on child and parent wellbeing and links wellbeing to parents’ jobs and work arrangements.

1.3 Parent and child wellbeing

Wellbeing is one way of assessing how parents and families are faring, and so can be used to track the success or otherwise of policies and intervention. Wellbeing is also closely linked to economic sustainability as productivity and participation depend upon people’s health and capability. Considering children’s wellbeing adds an intergenerational and future dimension; the future of Australian society and economy depends on the extent Australian children are resilient, educated and mentally and physically healthy, enabling them to participate and innovate once they grow up.

In the following section we briefly review available evidence on Australian parent and child wellbeing. Note, however, that most analyses of adult mental health do not distinguish between parents and non‑parents.

About one in 10 Australian adults aged 18 years or over reported a mental or behavioural problem that had lasted, or was expected to last, six months or more. Anxiety‑related and mood problems are the most commonly reported issues (ABS 2001c). Figures taken from the Child and Adolescent component of the 1998 National Survey of Mental Health and Wellbeing estimate that 14 per cent of Australian children aged between 4 and 14 years experience mental health problems (AIHW 2005). Attention deficit hyperactivity disorder (ADHD) was the most commonly reported disorder (17.8 per cent of boys and 7.9 per cent of girls) but may be overestimated due to diagnostic difficulties. In comparison, incidence of depressive disorder in children may be underestimated (3.7 per cent of boys and 2.6 per cent of girls) as this figure was based on parental reports and relied on parent ability to recognise and report child distress (AIHW 2005). Children in lone‑parent, step, blended and low‑income families were more likely to report mental health problems compared with children in two‑parent or in socioeconomically advantaged families (AIHW 2005).

There are few statistics comparing multiple aspects of wellbeing for Australian parents and children, making it difficult to ascertain how Australian families fare in comparison to other countries. In 2007 UNICEF assessed ‘economically advanced’ countries on six different aspects of child wellbeing (UNICEF 2007). The report contained statistics on Australian children in some of the wellbeing categories (there was incomplete data to fully include Australia). In terms of family and peer relationships, seven out of 10 Australian adolescents

Page 19: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

7

INTrOduCTION

(15 year olds) reported eating the main meal of the day with their parents ‘several times a week’ and just over one in two reported that their parents spent time ‘just talking to them’ several times a week. In a section on psychological and social aspects of ‘subjective wellbeing’, between 7.0 per cent and 9.0 per cent of Australian 15 year olds agreed with specific negative statements such as, ‘I feel lonely’ and ‘I feel awkward and out of place’. The report ranked Australia 19th out of the 25 countries included in the survey.

An OECD (2002) report on Australian literature on work–family balance found that many Australian parents working full time often felt that their employment hours ‘interfered negatively with family’. Dissatisfaction with work–family balance was highest among those working longer hours. In comparison with Denmark and the Netherlands, Australian males aged 25 to 54 years of age worked the longest hours of any age bracket across the three countries. Seventy per cent worked 40 hours or more per week compared with 52 per cent in the Netherlands and 42 per cent in Denmark. Similarly, 35 per cent of Australian women in the same age bracket worked 40 hours or more per week, compared with only 15 per cent in both Denmark and the Netherlands.

There are well‑established gender differences in the prevalence of anxiety and depression. The National Health Survey (ABS 2001c) shows that a greater proportion of women than men reported high (10.9 per cent compared with 7.2 per cent of men) and very high (4.4 per cent of women compared with 2.7 per cent of men) levels of psychological distress. Women showed higher rates of distress than men in all age groups except for 55 to 64 years, where rates of very high psychological distress were reported equally by men and women.

Low‑income and financial hardship are risk factors for parent depression (Mistry et al. 2002), which in turn play a critical role in children’s social development, mental health and wellbeing (Connell & Goodman 2002).

In 2001, mental and behavioural problems were more common among lone parents than parents with partners. Approximately 15 per cent of lone parents reported problems (ABS National Health Survey: mental health 2001c) compared with 7.3 per cent of parents in two‑parent families with dependent children. Rates of very high psychological distress showed a similar pattern. Just over 7 per cent of lone‑parent families with dependent children reported very high levels of psychological distress, compared with 2.8 per cent in two‑parent families with dependent children (ABS 2001c; Downey & Coyne 1990).

Similarly, socioeconomic disadvantage is associated with poorer overall child health including social and behavioural problems, injury, elevated blood pressure, asthma and negative health behaviours (Chen, Matthews & Boyce 2002; Propper, Rigg & Burgess 2007; SCARC 2004; Schoon, Sacker & Bartley 2003). A review on research into the link between socioeconomic position and child development found there is substantial evidence that children from low‑socioeconomic position (low‑SEP) families are at an increased risk of psychiatric disturbance, have poorer cognitive development, and academic and maladaptive social functioning (Bradley & Corwyn 2002). These health inequities can persist. Disadvantaged children are more vulnerable to compromised functioning and health in adulthood (Turrell et al. 1999).

linked lives: children’s stake in parents’ work

Children’s lives are intertwined with their parents’ lives. Elder (1985) describes each generation of a family as having life trajectories that are interlocking. Thus the resources, decisions, activities and events of one generation ‘bind’ to the life‑course trajectories and wellbeing of the other generations. The next section considers possible intergenerational linkages between employed adults’ work and their children’s wellbeing.

Work is a central activity of adult life (Green 2006). Work is more than a source of income; it shapes adult identity and engagement, and is a source of social connection and support. Work is also an important determinant of health. But few studies have considered the extent employed parents’ job conditions shape resources important to their children’s wellbeing. In the present study we focus on time resources (work hours and flexibility) and job security and control. Job security and control are both long‑established determinants of adult mental and physical health. In the following section we review a family resources theory of children’s wellbeing, before examining research on how jobs might impact family resources and, potentially, children.

Page 20: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

8 Social Policy research Paper No. 44

how children’s wellbeing is linked to their parents’: resource theory

A family resource theory would suggest that children’s wellbeing is a function of family resources that children depend on. This study’s focus is largely on parent wellbeing (measured by psychological distress) as the key mechanism linking jobs to children’s wellbeing. Other resources (which we also briefly review) are also important and appear in our analyses as covariates. Figure 1 illustrates the basic conceptual model used in this study.

Figure 1: Basic conceptual model of pathways between parent work and parent and child wellbeing

PARENT WORK PARENT RESOURCES CHILD WELLBEING

Time, income, human capital, psychological capital and social capital are major social and family resources that are relevant to child health and wellbeing (Zubrick et al. 2000) and may be key mechanisms linking parents’ work to children’s wellbeing. Jobs may affect these resources in different and contradictory ways. On the one hand, having a job delivers income and so is likely to be beneficial to parents and children. Jobs provide families with income, which gives parents purchasing power for resources such as food, housing, medical care and education (Zubrick et al. 2000). Hence, income is important to children’s development as it can help to provide environments conducive to learning and good health (Cooksey, Menaghan & Jekielek 1997). Financial hardship may also cause family stress and depression for parents, diminishing parents’ capacity for care giving, attention and supervision (Zubrick et al. 2000). Additionally, income can be used to help substitute for other resource areas that are pressured (Zubrick et al. 2000). For example, families that are time poor may purchase help with child care or housework (Zubrick et al. 2000).

On the other hand, there have been concerns about parents’ time with children and whether employment constrains care giving. The quality and quantity of available time (Zubrick et al. 2000) influences when care giving can be undertaken and the way in which it is done. The quality of time available for care giving is a function of the human and psychological capital of the parents (Zubrick et al. 2000). Time also influences other family resources as it is needed to be able to utilise these resources (Zubrick et al. 2000). Quantity of time is linked to other demands placed on parents, especially employment, raising questions about the influence of parents’ work hours and children’s wellbeing. Finally, jobs may also influence parents’ mental health and wellbeing, a central resource for children and a well‑established determinant of the quality of children’s care (Cooksey, Menaghan & Jekielek 1997).

Parents’ health is integral to their human capital and is a key resource for children. It also underpins access to others resources such as income. Without good health, parent employment can be curtailed. In terms of productivity and participation, poor mental health is one of the most limiting of all conditions. The estimated costs of depression to US business is in the region of US$33 billion annually (Druss, Rosenheck & Sledge 2000; Greenberg et al. 1996). Zubrick and colleagues (2000) view parent mental health and wellbeing as a psychological resource for children and include mental health, levels of stress and conflict within the family, family support and family cohesion. In this study we focus on parent mental health (psychological distress) and on a relatively novel element of family resources—parents’ work‑related wellbeing (work‑to‑family gain and work‑to‑family strain).

There is now a long‑established research literature on the critical role parents’ mental health and wellbeing, especially depression, plays in children’s own social development, mental health and wellbeing (Connell & Goodman 2002; Downey & Coyne 1990). Depressed parents are less spontaneous and more withdrawn, angry and sad (Downey & Coyne 1990; Lovejoy et al. 2000). Children of depressed parents are more likely than children of non‑depressed parents to have emotional or behavioural difficulties, poor physical health and impaired social or academic performance not attributable to any intellectual limitations (Lovejoy et al. 2000). Similarly, children’s wellbeing depends on the quality of family relationships (Cummings 1994; Repetti, Taylor & Seeman 2002; Sanders, Gooley & Nicholson 2000), especially the emotional tone and consistency of parenting that is linked to parent mental health. Parent–child interactions play a central role in determining children’s

Page 21: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

9

INTrOduCTION

wellbeing with potentially long‑term consequences of dysfunctional parent–child interactions (Dwyer, Nicholson & Battistutta 2003; Rodgers 1996; Rosenman & Rodgers 2006).

Interestingly, there is some evidence that the interplay between parent and child wellbeing may change over the life course. Greenfield and Marks (2006) found that elderly parents’ wellbeing was associated with the problems of their adult children, ranging from chronic illness, depression, alcohol or substance abuse, to legal problems. Fathers’ mental health is also a determinant of their children’s wellbeing but this interconnection has received far less research attention (Clarke‑Stewart 1978; Phares & Compass 1992). Interestingly the few studies on paternal depression find equally strong influences on children’s wellbeing as is found with maternal depression (Phares, Duhig & Watkins 2002). There is virtually no research examining how fathers’ work‑related wellbeing might impact on their children’s wellbeing.

Life‑course perspectives emphasise the way children’s wellbeing is a function of early experiences and socioeconomic disadvantage. Social inequalities can drive health inequalities and both can accumulate over the life course. Those who are more disadvantaged during childhood are more vulnerable to poor wellbeing, which then leads to compromised functioning and health in adulthood, in turn limiting participation and adult achievement (Keating & Hertzmann 1999). Social and health disadvantages can therefore be ‘transmitted’ (replicated) between generations, from parents to children (Bradley & Corwyn 2002).

1.4 Work and parent wellbeing

A wide range of work conditions have been linked to employed adults’ wellbeing, although relatively few studies have focused specifically on parents. Managing the interface between work and family can place additional strains on working parents, resulting in increased stress, depression and poorer health (Kinnunen & Mauno 1998), and there is limited evidence that some work conditions make it harder for parents to combine work with providing care for children. Long work hours, unpredictable or inflexible work schedules, night or evening work, low control and work pressure are associated with greater work‑to‑family conflicts (Grzywacz, Almeida & McDonald 2002). Very few studies have included children’s outcomes in analysis of the impacts on wellbeing.

Work hours and employed parents’ wellbeing

In a review of studies examining working hours and health, long work hours were linked to increased risk for cardiovascular disease, diabetes, illnesses leading to disability retirement, subjectively reported physical ill health and fatigue (van der Hulst 2003). There was also a relationship between long work hours and depression of the immune system, which may be due the association between long hours, work intensity and stress. Approximately 54 per cent of long‑hour workers describe their jobs as stressful compared with 39 per cent of those working standard hours (Wooden 2002).

Although long hours are mostly worked by men who are usually fathers (Goward et al. 2005), it appears that long hours affect women’s health to a greater extent. Women working long work hours face an increased incidence of stress‑related illnesses that men do not experience (Wooden 2002). This is most likely due to the double burden mothers face when they combine traditional family roles and paid work.

Work hours may impact parents’ wellbeing via the experience of time pressure. A US survey found distress in men and women is significantly associated with time pressure (Roxburgh 2004). Additionally, women who experience time pressure ‘everyday’ or ‘often’ are more likely to assess their health as poor (Goward et al. 2005). Social support at work and at home may help mitigate the negative impact of long hours and time pressure. Lower rates of depression were found for both men and women who had high partner social support. This was also the case for men who had high co‑worker social support (Roxburgh 2004).

Page 22: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

10 Social Policy research Paper No. 44

Job control, perceived insecurity and employed parents’ wellbeing

Work conditions such as low job control and job insecurity have long been recognised as making a significant, independent contribution to a range of adverse health outcomes, including ischaemic heart disease, depression, anxiety, sickness absence and poor self‑rated health (Ferrie 1999; Stansfeld, North & Marmot 1995). Job insecurity and low control are especially important to adult mental health, showing cross‑sectional and longitudinal associations comparable to other well‑established predictors such as adverse life events, major losses and relationships disruption (Cole et al. 2002; Marmot et al. 1999). Research in Australia has found consistent, strong, cross‑sectional associations of job insecurity, job control and high work demands with symptoms of anxiety and depression (D’Souza et al. 2003; Sanderson et al. 2005).

Work is one of the mechanisms that links social status and position to health inequality. Marmot and colleagues (1999) conducted a series of studies on British civil servants and found a clear gradient in health linked to job seniority. Employees in the most senior jobs had the best health, while those in the least skilled and lowest jobs had the worst, while employees in middle ranking jobs had moderate health outcomes. This gradient was a function of income and job control—the extent employees could control how and when their job was done or could influence decisions about their job (eds Marmot & Wilkinson 1999). This research would suggest that low‑skilled parents are more likely to work in jobs with poor conditions and may be at risk for poorer mental and physical health, potentially also eroding the resources that their children depend on.

In summary, epidemiological research demonstrates strong and consistent associations between job insecurity and control and adult mental health and wellbeing. Combinations of adverse conditions appear to be particularly harmful and there is also evidence that work conditions are one of the ways social inequity and exclusion translate into poorer health. Few studies have addressed the possibility that the influence of work conditions extend beyond individual employees, across generations, to their children. We briefly review existing research evidence below.

1.5 Parents’ work and children’s wellbeing

Parent work hours and children’s wellbeing

Most research has examined mothers’ work hours and children’s wellbeing. However, longer work hours do not automatically translate into less time spent with children, as parents strive to protect their time with children, rearranging their schedules and compensating for time lost (Sayer, Bianchi & Robinson 2004). In the US, despite increases in maternal time in paid work and reduction in number of children per family, married mothers’ time spent in caring for children has increased since 1985 and is now the same or higher than in 1965 (Sayer, Bianchi & Robinson 2004). The trend for increasing maternal time in high investment activities with children is even stronger (Sayer, Bianchi & Robinson 2004). More striking is the increase in married fathers’ time spent in both caring for children and teaching and playing with children (Sayer, Bianchi & Robinson 2004). The proportion of fathers who reporting spending time with their children has increased as well as the amount of time they spend with their children (Sayer, Bianchi & Robinson 2004).

Parents working evenings, nights or weekends report fewer opportunities for meals with their children, for reading together, playing and helping with schoolwork (Fagan 2001; Heymann & Earle 2001; La Valle et al. 2002). Work clearly influences the time parents have available for such activities and these types of activities are important for the transfer of resources to children. The quality of parenting is a key influence on children wellbeing (Repetti, Taylor & Seeman 2002) and when work impacts on parents’ wellbeing this can affect their parenting. It may be more important to consider parents’ mood and parenting when spending time with their children.

Children appear to be sensitive to the pace of time with their parents. If they experience time with their parents as rushed, they rate their parents much lower in terms of parenting, management of work and family and putting family life before work (Galinsky 1999). Longer work hours may also affect children in terms of parents’

Page 23: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

11

INTrOduCTION

energy levels. For example, Galinsky (1999) found that mothers’ employment was not predictive of children’s report of parenting, though children did wish that their parents were less affected by their work. Children did not generally wish for more time with their parents (about 10 per cent wish for more time with their mother, 16 per cent wish for more time with their father). Instead 34 per cent wished their mothers would be less stressed or tired from their work, while 28 per cent wanted the same for their fathers (Galinsky 1999).

Parents’ job control, perceived insecurity and children’s wellbeing

One possible mechanism linking work conditions such as job control and insecurity to children’s wellbeing is via parents’ mental health and mood. Parents’ mental health is a key determinant of their children’s wellbeing, and epidemiological research has established, cross‑sectionally and longitudinally, associations between low job control and perceived insecurity at work and adult mental health. But the extent this translates into poorer child wellbeing is virtually unknown with only a very few studies specifically examining parents’ job control and insecurity.

When parents feel frustrated, angry or distressed at work they are more likely to be punitive and harsh with their children (Galinsky 1999). Bringing work stress home (‘stress transmission’) has been associated with changes in wellbeing of employees’ spouses. Most of these studies are based on diary data, recording husbands’ and wives’ daily work stress, marital interactions and wellbeing. These studies generally indicate that work stress in one spouse is associated with increased marital tension and also poorer emotional wellbeing in the other spouse (although one study failed to find clear evidence of work stress transmission (Jones & Fletcher 1996). Similarly, parents with stressful jobs are more likely to be tense at home, which can translate into withdrawing from, or arguing with, their children (Almeida, Wethington & Chandler 1999; MacEwen & Barling 1994; Repetti 1989). Work stress, at least in fathers, affects their day‑to‑day interactions with children. After a high workload day, fathers tended to withdraw from their children. If angry or distressed from work, fathers also reported more irritable parent–child interactions (Repetti 1994). Poor quality work conditions, such as job insecurity and low job control, are also thought to negatively impact on family interactions (Cooksey, Menaghan & Jekielek 1997; Menaghan 1991). It is possible that these work conditions lead to reduced levels of self‑esteem and personal control as well as exacerbating psychological distress, making parents less able to show warmth to their children or provide a stimulating environment (Cooksey, Menaghan & Jekielek 1997). For example, fathers’ self‑esteem and interactions with their children (including showing warmth and acceptance) are linked to their experiences of control and support at work (Grimm‑Thomas & Perry‑Jenkins 1994).

Job security is an important part of maintaining strong families (Larson, Wilson & Beley 1994) but few studies have examined how children fare when their parents are insecurely employed or fearful for their job future. There is suggestive evidence that parents’ job insecurity may be associated with poorer academic performance in adolescent and young adult children (Barling, Zacharatos & Hepburn 1999).

Importance of developmental stages

Because families change as children develop, the way parents’ jobs relate to wellbeing may also vary over the life course. Do aspects of parents’ jobs such as their work hours, control or security promote parent and child wellbeing differently when children are infants, preschool aged or adolescents? The life‑course paradigm would propose that the timing of an event may be more significant than its actual occurrence, but there is virtually no research comparing parent working conditions or arrangements and parent and child wellbeing across different developmental stages (Elder 1995).

It is well established that children’s needs for care and their dependence on families are greater when they are young, placing pressures on parents’ time and energy. Both mothers and fathers spend more time per day performing child care tasks when their children are younger (Higgins, Duxbury & Lee 1994). There is also evidence that parenting demands vary across different development stages. For example, the time demands of raising children between 6 and 12 years old are largely focused on home chores and on errands (Higgins, Duxbury & Lee 1994).

Page 24: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

12 Social Policy research Paper No. 44

Thus time poverty, long work hours and family‑friendly conditions may be particularly important to families when children are young due to the increased care demands during this developmental stage. For example, mothers with young children experience significantly higher role overload and work–family interference than mothers with adolescents (Higgins, Duxbury & Lee 1994). Time pressures for dual‑earner families may be particularly high, placing dual‑earner parents with young children, especially mothers, at greater risk for poor mental health (Craig 2002; Roxburgh 2004).

As well as placing different demands on parents, children’s developmental age may shape their reactivity and sensitivity to parents’ mental health problems. For example, young children may be more sensitive to disruptions in routines and care settings and more reactive to parents’ irritability and inconsistent parenting arising from parent distress (Lovejoy et al. 2000). As a result, work demands may interact with care giving demands in different ways across different child developmental stages, thereby affecting parents’ mental health and wellbeing. Additionally, the links between parents’ mental health and children’s wellbeing may also vary, depending on children’s developmentally‑based dependence on their parents’ care.

1.6 Summary of key issues

Increasingly, mothers have remained in, or returned to, the paid workforce. Care demands have remained and perhaps even increased in an unprecedented era of intensive parenting. Despite needing time and energy for care giving, parents enter a labour market where work demands and expectations have increased over the previous few decades, especially for full‑time jobs. At the same time most fathers have not reduced their paid work hours. Thus, mothers continue to shoulder the main responsibility for raising children, while also working within a labour market where most jobs require high effort and increasingly long‑hour commitments (Goward et al. 2005).

Many Australian mothers have therefore opted for part‑time work as one way to manage family time pressures. Paradoxically this reinforces the ‘male breadwinner – female care giving’ family model as fathers continue to work longer hours and mothers find themselves managing a ‘second shift’ or ‘double load’ due to managing housework and child care on top of paid work (Goward et al. 2005).

What are the implications for gender equity of these gendered solutions to working and care giving and how are mothers and fathers faring? Are families with two employed parents showing signs of strain in their wellbeing relative to breadwinner families with one parent employed? In terms of wellbeing, does it matter if it is the mother or the father who is the sole earner? Do some work conditions such as flexible work times, control and security improve parents’ capacity to cope with the second shift and reduce potential trade‑offs between working and wellbeing? Does this also translate to better wellbeing for children and if so, at what ages and in what sorts of families? This study will supply new evidence to address these vital questions as a first step.

Low‑income families and lone‑parent families are two groups facing particular challenges for combining working with care giving. Because most lone‑parent families are headed by mothers, these challenges further underline issues of gender and social equity. For example, the low‑income workforce in the US is disproportionately made up of workers who are female, less highly educated, young and from minority groups (Bernstein 2004). In Australia, women and those who do not have post‑secondary educational qualification are over‑represented in low‑paid jobs (SCARC 2004).

Income is a well‑established determinant of wellbeing, both of parents and children. Low‑paid workers experience ongoing financial pressures, spend their money overwhelmingly on necessities, often face debt in order to make ends meet and miss out on participating in wider society (such as recreation and social activities) (SCARC 2004). Low‑paid workers face a constant tension managing time pressures of work and family life on a low wage (Masterman‑Smith, May & Pocock 2006). Financial and time problems could be compounded easily. For example, illness is problematic, especially for casual workers, because it leads to loss of income (Masterman‑Smith, May & Pocock 2006). Additionally, access to education, recreation, health

Page 25: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

13

INTrOduCTION

care and family interactions for the children of these workers are often curtailed (Masterman‑Smith, May & Pocock 2006).

The composition and characteristics of Australian families have changed in the past decades due to social and economic change (Goward et al. 2005). While there has been a decline in two‑parent families with children there has been a rise in lone‑parent families along with same‑sex two‑parent families and step‑blended families (Goward et al. 2005). Over the last two decades, lone‑parent families have increased from 14 per cent of all families with children under the age of 15 years in 1986–88 to 22 per cent in 2004–06 (ABS 2007). This increase may be due to the increase in separation and divorce, as well as the increase in the proportion of births outside of marriage (ABS 2003a). Furthermore, lone‑parent families are much more likely to be headed by mothers. In 2001, 83 per cent of lone parents were mothers (ABS 2003a).

Lone‑parent families tend to be low‑income families (ABS 2007). Forty‑four per cent of lone‑parent families, compared to 14 per cent of all households, have low economic resources (ABS 2007). Compared to parents in two‑parent families, lone parents tend to have lower levels of educational attainment and lower rates of labour force participation (ABS 2007). About one in two lone parents were employed (52 per cent of those employed worked full time) (ABS 2003b) and 61 per cent of lone parents relied on a government pension or allowance as a main source of income in 2003–04 (ABS 2007). Lone‑parent families are more likely to report experiencing financial stress, with 51 per cent of lone parents reporting they ‘could not pay electricity/gas bills on time’ compared to 20 per cent of two‑parent families (ABS 2007). Lone mothers are twice as likely to experience ‘moderate to severe mental disability compared to partnered mothers’ (Crosier, Butterworth & Rodgers 2007). Employment may help reduce lone mothers’ distress by reducing their financial hardship. However, jobs with poor conditions may undermine their wellbeing.

In terms of wellbeing, low‑income families and lone mothers who combine working with providing care for children may face a double jeopardy. Low pay may be compounded by low control, inflexible work times and insecure levels of employment (Bernstein 2004; Fenwick & Tausig 2001; SCARC 2004). Despite this, there is some evidence that children in low‑income families benefit from mothers’ work (Galinsky 1999). Jobs help families stay out of poverty. Jobs that offer lone mothers and low‑paid employed parents’ flexibility, control and security may help optimise wellbeing for parents and their children and avoid compounding pressures of low‑income and poor conditions. Our study aims to assess this possibility.

Children’s health and wellbeing is of vital concern. Children provide the pool of talent and resources for tomorrow’s globally competitive economy and it matters that they have a healthy start and the best life chances. It is well established that income, close family relationships and parents’ wellbeing are resources on which children depend. Consequently, optimising parents’ wellbeing has the potential to flow on to benefit children. A considerable research effort has documented the adverse impacts of poverty and social and economic disadvantage on children. Despite this, virtually no research regards the nature of parents’ work as a key determinant of whether children benefit from their parents’ employment and if so, to what extent.

This study is one of the first to examine associations between work conditions and the wellbeing of parents and their children. It is well established that parent mental health (particularly depression) plays a critical role in children’s social development, mental health and wellbeing. However, to our knowledge, no study has questioned whether parents’ work‑related wellbeing also affects children’s wellbeing. Anecdotal evidence and widespread media coverage would suggest that strains from combining working with raising children are widespread. This study would be one of the first to test whether these strains could have consequences for the wellbeing of Australian children as well as their parents.

1.7 Conceptual framework

The research draws on two conceptual frameworks. The first overarching framework nests children and families within larger social and cultural contexts. The ecological model of child development (Bronfenbrenner 1979) views children’s development and wellbeing as an interplay between biological and

Page 26: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

14 Social Policy research Paper No. 44

other characteristics and children’s environments. The family is the most proximal environment for children, especially in early life, but the importance of environments change as children develop. For example, the place and significance of child care and school, peers and neighbourhood in children’s lives varies as they grow. All these environments are in turn shaped by wider economic, social, political and cultural forces.

The second framework specifies the way parents’ work, which is also a function of the wider economic, social, political and cultural forces, intersects with child developmental stage, parent mental health and wellbeing to influence children’s wellbeing (Strazdins et al. 2006). Furthermore, it draws on the life‑course paradigm, which views family members’ lives as linked and interdependent so that the wellbeing of one family member is shaped by resources in the family environment that may include the wellbeing of other family members. Parents’ wellbeing impacts upon children’s social development, mental health and wellbeing (Connell & Goodman 2002; Downey & Coyne 1990).

The second framework also draws on the established relationship between working conditions and adverse mental health outcomes in employed adults (Stansfeld, North & Marmot 1995). The current study extends this model of the interface between work and family wellbeing to address differences in family types and family socioeconomic position. We propose that elements of parents’ work may influence children’s wellbeing by impacting upon critical resources available in the family environment, especially parent wellbeing.

This study focuses on three elements of parents’ work: work hours, job security and job control. Work hours includes a continuous measure of work hours and whether the parent works part‑time hours (34 hours or less per week), standard full‑time hours (35 to 45 hours per week) or long full‑time hours (46 hours or more per week). Job security encompasses both perceptions of job security and the type of employment tenure (self‑employed, permanent, fixed term, casual) on the basis that casual and fixed‑term employment tends to be less secure than permanent employment (considered to be the most secure) or self‑employment. Job control encompasses work‑related control over tasks in the job and control over hours, in terms of having flexible start and finish times.

Figure 2: Conceptual model of pathways between parent work arrangements and parent and child wellbeing

PARENT WORK

Hours, security, control

PARENT WELLBEING

Psychological distress Work–family strains and gains

CHILD WELLBEING

Social Emotional Academic

FAMILY DIFFERENCES AND DYNAMICS

Child development, sex of parent, family type, family socioeconomic position, earner arrangements

D

B

C

A

Figure 2 illustrates a simplified model of the relationships between parents’ work and parent and child wellbeing. It proposes that parents’ work may alter child wellbeing by optimising (or eroding) parent mental health or by reducing (or increasing) the conflict between work and family (Path A). This will in turn influence the resources available to children within their family environment and thereby influence their wellbeing (Path B) by optimising or eroding their mental health.

Possibly, the elements of work important to child wellbeing could depend on child developmental age (represented by cohort), family socioeconomic position, family type and whether the elements under consideration are part of the mother’s or father’s job characteristics. We do not formally test for interactions based on parents’ sex. We compare cohort differences in mean scores or percentages using analysis of variance (ANOVA) or chi‑square analyses. Paths C and D represent the possibility of interactive effects of family

Page 27: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

15

INTrOduCTION

type (lone or two‑parent families) and socioeconomic position (families in the lowest quartile compared with the families in the remaining quartiles). In the multivariate analyses in Section 6 we test and report on these interactions.

Household income, parent years of education and chronic parent health problems are also critical to family resources and may have implications for parents’ and children’s wellbeing. Both may also confound associations between parents’ work, parents’ mental health and work–family conflict and children’s wellbeing. Therefore we control for household income and parent years of education in all analyses except for analyses stratified by family socioeconomic position.

1.8 research questions

Using samples from three datasets (LSAC B cohort; LSAC K cohort; and HILDA survey households with a child aged 15 to 17 years), the project tests three overarching questions:

1. Do employed parents’ work hours matter to parents’ and children’s wellbeing?

2. Does employed parents’ job security (perceptions and employment tenure) matter to parents’ and children’s wellbeing?

3. Does employed parents’ work‑related control (over hours and tasks) matter to parents’ and children’s wellbeing?

We then ask if associations vary according to child cohort, family type and family socioeconomic position. The study also compares pathways for mothers and for fathers to address potential gender differences in the linkages between parent work and parents’ and children’s wellbeing.

Page 28: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

16 Social Policy research Paper No. 44

Page 29: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

17

SAMPlE ANd MEASurES

2 Sample and measures

2.1 datasets methodology

general description of datasets

This project uses unweighted cross‑sectional data from two studies, the Longitudinal Study of Australian Children (LSAC) and the Household, Income and Labour Dynamics in Australia (HILDA) survey. Detailed information about each, including design and sample, is available in LSAC Technical Paper no. 1 (Soloff, Lawrence & Johnstone 2005) and the HILDA User Manual (eds Goode & Watson 2007).

LSAC is a national, cross‑sequential, longitudinal study of children and their families. The current study used data from Wave 1, which had two cohorts: families with infants (infant cohort: 5,104 infants aged 3 to 19 months) and families with children aged 4 to 5 years old (preschooler cohort: 4,976 children aged between 4 years, 3 months and 5 years, 7 months (AIFS 2006)). It contained data on the study child and his or her parents, but did not gather detailed information about other children in the family. The two cohorts of families were selected from the Medicare enrolment and activity databases held by the Health Insurance Commission. Data were collected via face‑to‑face interviews, mail‑back self‑complete surveys, interviewer observations, direct physical measurements, direct assessments of school readiness and two, 24‑hour time‑use diaries. For more detailed information on the instruments, see the LSAC Data Users Guide (AIFS 2007). The data for the present study were mainly drawn from the interview (mostly completed by the child’s mother) and from self‑complete questionnaires.

In LSAC, data were collected from ‘Parent 1’—the primary parent who had the most knowledge about the child—via an initial interview and a self‑report questionnaire. In 95 per cent of cases, Parent 1 was the study child’s mother. Parent 1 answered the majority of questions about the child’s background (such as their health, education, child care, family details and so on) as well as questions about themselves (such as demographics, lifestyle and couple relationships). Data from ‘Parent 2’ (the residential partner of Parent 1) were collected via self‑report survey only. Parent 2 responded to questions mainly about themselves. Teachers and day‑carers of the study child also filled out self‑report questionnaires where applicable.

In certain sections of the Parent 1 interview questionnaire, Parent 1 answered items on behalf of Parent 2 as well as for themselves. Note this only occurred for families where there were two resident parents. The items for which Parent 1 answered on behalf of Parent 2 included:

�� education

�� income

�� employment status

�� employment type

�� whether absent from work

�� work hours

�� paid leave (recreation, sick, maternity/parental, personal/family and the derived measures of family‑friendly leave; combined sick/recreation leave).

As around 95 per cent of Parent 1s were mothers, it was primarily mothers reporting on the above items for fathers. All other items were self‑report items.

Page 30: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

18 Social Policy research Paper No. 44

The Household, Income and Labour Dynamics in Australia (HILDA) survey is a large national probability sample of Australian households occupying private dwellings. Wave 1 of the HILDA survey commenced in 2001 with a total sample of 7,682 households and 19,914 people. All members of the households providing at least one interview in Wave 1 were pursued in each subsequent wave. The sample has been gradually extended to include any new household members resulting from changes in the composition of the original households. Households were selected from a sampling frame of census collection districts from across Australia. Unlike LSAC, the sampling unit of interest in the HILDA survey is the household rather than the child. HILDA survey data were obtained via self‑complete questionnaires and interviews. Wave 4 was the first year in which data were collected for 15 to 17 year‑old participants, as the HILDA survey does not interview or collect data on children under 15 years of age. The present study uses a sub‑sample of Wave 4 HILDA survey data, households with children aged 15 to 17 years, to derive a sample of families with adolescents. Although the HILDA survey collected data from all household members, the current project examines data from mothers, fathers and adolescents aged 15 to 17 years only.

HILDA survey data were collected via a series of questionnaires. These were: the household form (delivered via interview, collected basic information about the composition of each household); the household questionnaire (delivered via interview, collected household level information about spending, child care and housing); person questionnaires (delivered via interview, collected individual‑level demographic information); and self‑completion questionnaires (self‑completed, collected information about general health and wellbeing, lifestyle, attitudes and values, and so on). The current study uses information from each of these sources.

response rates

From the initial LSAC sample of 9,326 infants and 10,596 children aged 4 to 5 years, the study recruited 5,107 infants and 4,983 children. The overall response rate for the Parent 1 interview was 55 per cent for the infant cohort and 47 per cent for the preschooler cohort (AIFS 2007). The main sources of sample loss from LSAC were: refusals to participate (31.2 per cent in the infant and 35.0 per cent in the preschooler cohorts), non‑contact (10.4 per cent in the infant and 14.2 per cent in the preschooler cohorts) and Health Insurance Commission decision not to contact, due to similarities between a child’s name and a child listed in the ‘fact‑of‑death’ file (3.3 per cent in the infant and 2.4 per cent in the preschool cohorts). The most common reason for refusal to participate was ‘not interested/too busy’ (57.0 per cent) (AIFS 2007).

Of the 11,693 households identified as suitable for the HILDA survey Wave 1 data collection, interviews were completed with all eligible members of 6,872 households and with at least one eligible member of a further 810 households. The household response rate for Wave 1 was therefore 66.0 per cent. Of households who responded in Wave 1, 78.6 per cent were retained for Wave 4 and were therefore relevant to the current study (eds Goode & Watson 2007).

Psychometric properties

Measures used in both LSAC and the HILDA survey were selected on the basis of sound psychometrics properties, brevity, valid benchmarking and use in other longitudinal surveys. In addition, we performed preliminary psychometric testing of multiple item scales to check scale normality and inter‑item reliability for this study. In LSAC, this testing was performed on the combined sample of B and K cohorts (or the B or K cohorts separately, where measures differed between cohorts). In the HILDA survey, we performed psychometric testing using the sub‑sample of households with adolescents aged 15 to 17 years. Results of normality and inter‑item reliability have been reported under the relevant variable name headings.

Page 31: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

19

SAMPlE ANd MEASurES

2.2 Methodology for current study

Participant inclusion criteria

The focus of this study is on families where at least one parent is employed. Therefore, we retained two‑parent families if one parent was unemployed (or not in the labour force), but excluded them if both parents were unemployed and/or not in the labour force.

There was a small number of two‑parent families where the employment of one or both parents was unknown. We excluded these families from the sample because their earner arrangements could not be classified as either breadwinner or dual earner without knowing the employment status of both parents. Similarly, we excluded lone‑parent families if the parent was unemployed or their employment status was unknown. In addition, because of the very small number of lone‑father families in which the father was employed and the small number of same‑sex couples, we excluded these families from analysis.

In the infant and preschooler samples, the sample of parents and their children included in the study satisfied the following conditions:

�� parent(s) were employed in the last month

�� sample had complete data on all of the following variables:

�– work arrangements (work hours, job security, job control, flexible hours)

�– parent wellbeing (parent psychological distress)

�– the infant temperament scale

�– the Strengths and Difficulties Questionnaire (SDQ) in the preschooler sample.

In the adolescent sample, the sample of parents and children with complete data satisfied the following conditions:

�� parent(s) were in paid work

�� parent(s) had parenting responsibilities for resident children aged 17 years or less

�� randomly selected for one study child (to avoid clustered data in multiple‑child families)

�� had complete data on the following variables:

�– work arrangements (work hours, job security, job control, flexible hours)

�– parent wellbeing (SF‑36 mental health)

�– adolescent wellbeing, including the mental health subscales of the SF‑36, school retention and self‑rated education performance.

Flow charts showing selection steps for the infant, preschooler and adolescent samples are presented in Figures 3, 4 and 5 below.

Page 32: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

20 Social Policy research Paper No. 44

Figure 3: Families with infants, LSAC B cohort (n=3,266): flow chart of parents in each family employment category with complete data

All families 5,107

Families with an employed parent 4,529

No employed parent/unknown 578

Couple(a)

4,412 (4,417)Lone(b)

112

Unknown 13

Breadwinner 2,075

Dual earner 2,324

Complete data 74

Incomplete data 38

Complete data 1,307

Incomplete data 768

Complete data 1,885

Incomplete data 439

(a) Five same‑sex couples were excluded from further analysis.(b) All lone employed parents were mothers.

Figure 4: Families with preschoolers, LSAC K cohort (n=3,206): flow chart of parents in each family employment category with complete data

All families 4,983

Families with an employed parent 4,401

No employed parent/unknown 582

Couple 4,102(a) (4,107)

Lone 280(b) (294)

Unknown 13

Breadwinner 1,628

Dual earner 2,461

Complete data 198

Incomplete data 82

Complete data 1,032

Incomplete data 596

Complete data 1,976

Incomplete data 485

(a) Five same‑sex couples were excluded from further analysis.(b) Fourteen of the lone employed parents were lone fathers. These cases were excluded from further analysis.

Page 33: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

21

SAMPlE ANd MEASurES

Figure 5: Families with adolescents, HILDA survey (n=360): flow chart of parents in each family employment category with complete data

All families 508

Families with an employed parent 451

No employed parent/unknown 57

Couple 377

Lone(a)

60 (74)

Unknown 23

Breadwinner 84

Dual earner 270

Complete data 45

Incomplete data 15

Complete data 62

Incomplete data 22

Complete data 253

Incomplete data 17

(a) Fourteen of the lone employed parents were fathers. These cases were excluded from further analysis.

Missing data

To avoid inconsistencies in samples across analyses, the samples for all three cohorts were restricted to families who had complete data (that is, no missing data) on all variables relevant to the current study. To determine the impact of this on the sociodemographic composition of the retained sample compared to the original sample, we conducted a missing data analysis. Full details of this analysis are presented in Appendix A.

For both LSAC and the HILDA survey we computed an indicator of whether a case had complete data on all variables or had missing data on one or more of the variables. We then conducted independent t‑tests to check for differences between employed parents with complete versus incomplete data on age, years of education, combined yearly income and mental health. In the LSAC sample (infant and preschooler cohorts), the differences between parents with complete and incomplete data were very small. There were no differences observed in the HILDA survey adolescent sample.

Significant differences in the LSAC sample were as follows:

�� In the infant cohort, employed mothers with complete data were younger, had more years of education and had lower psychological distress scores.

�� In the infant cohort, employed fathers with complete data were younger, had more years of education and had higher levels of combined yearly income.

�� In the preschooler cohort, employed mothers with complete data had more years of education and had combined yearly incomes.

�� In the preschooler cohort, employed fathers with complete data had more years of education, higher combined yearly incomes and higher levels of psychological distress.

Note that these differences were very small and were therefore unlikely to have made a large difference to the outcomes of this report.

Page 34: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

22 Social Policy research Paper No. 44

representativeness of sample

In selecting parents with complete data only, it is possible that the sample under‑represented or over‑represented some groups of participants. To investigate this possibility, we performed chi‑square tests comparing the prevalence of employed mothers and fathers with and without complete data for family type, earner arrangements, employment tenure and part or full‑time employment status.

The analyses show that the differences are in employment tenure. Compared to the samples of all employed mothers and fathers, the complete data samples were as follows:

�� In the infant cohort, lone mothers and mothers on fixed‑term tenure are under‑represented and couple mothers and mothers on permanent/ongoing tenure are over‑represented.

�� In the infant cohort, self‑employed fathers and fathers on casual tenure are under‑represented while fathers on permanent tenure are over‑represented.

�� In the preschooler cohort, self‑employed mothers and mothers on casual tenure are under‑represented while mothers on permanent tenure are over‑represented.

�� In the preschooler cohort, self‑employed fathers, fathers on casual tenure and fathers working part time are under‑represented, while fathers working full time are over‑represented.

�� In the adolescent cohort, lone mothers are under‑represented and mothers in two‑parent families are over‑represented.

Comparison of sample against AbS data

Both LSAC and the HILDA survey slightly under‑represent some families, especially those with low socioeconomic position. (See AIFS 2006 and eds Goode & Watson 2007 for detailed discussions of sample representativeness for LSAC and the HILDA survey respectively.)

We used census data on the employment status of parents by family type and age of child (see ABS 2003b, 2004) to assess the representativeness of the sub‑samples of lone‑mother families, breadwinner two‑parent families and dual‑earner families. The results are reported in Table 1. We did not assess the representativeness of low‑SEP and medium/high‑SEP families in these comparisons because our measure of household socioeconomic position is different from the census definition. However, the sub‑samples likely under‑represent low‑SEP families, given under‑representation in the complete LSAC and HILDA survey samples.

Page 35: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

23

SAMPlE ANd MEASurES

Tabl

e 1:

Co

mpa

rison

of e

mpl

oyed

mot

her a

nd fa

ther

sub

‑sam

ples

aga

inst

AB

S da

ta

LSAC

Infa

nts

(3–1

9 m

onth

s)

ABS(a

)

Fam

ilies

with

ch

ildre

n 0–

2 ye

ars

LSAC

Pres

choo

lers

(4–5

yea

rs)

ABS(a

)

Fam

ilies

with

ch

ildre

n 3–

4 ye

ars

HIL

DA

Adol

esce

nts

(15–

17 y

ears

)

ABS(a

)

Fam

ilies

with

de

pend

ent s

tude

nts

15–2

4 ye

ars

Mot

hers

All

n1,

867

309,

100

2,08

324

6,90

027

771

3,40

0

Lone

par

ent

% o

f em

ploy

ed

mot

hers

4.0

8.5

9.5

10.8

16.3

14.4

Bre

adw

inne

r(t

wo‑

pare

nt fa

mili

es)

% o

f em

ploy

ed

mot

hers

3.4

4.3

3.1

4.9

2.8

6.3

Dua

l ear

ner

(tw

o‑pa

rent

fam

ilies

)%

of e

mpl

oyed

m

othe

rs92

.787

.287

.484

.280

.979

.3

Fath

ers

All

n2,

832

576,

400(b

)2,

633

373,

200(b

)27

073

9,60

0(b)

Bre

adw

inne

r(t

wo‑

pare

nt fa

mili

es)

% o

f em

ploy

ed

fath

ers

43.9

53.2

36.8

44.3

20.0

23.5

Dua

l ear

ner

(tw

o‑pa

rent

fam

ilies

)%

of e

mpl

oyed

fa

ther

s56

.146

.863

.255

.780

.076

.5

(a)

AB

S da

ta fo

r bot

h ta

bles

are

take

n fr

om T

able

10:

All

child

ren:

em

ploy

men

t sta

tus

of p

aren

ts b

y fa

mily

type

by

age

of c

hild

(see

AB

S 20

03b,

p. 2

6).

(b)

This

figu

re re

pres

ents

em

ploy

ed fa

ther

s in

two‑

pare

nt fa

mili

es o

nly

(tha

t is,

exc

lude

s lo

ne‑f

athe

r fam

ilies

).N

ote:

D

ue to

roun

ding

, per

cent

ages

may

not

add

to 1

00 p

er c

ent.

Page 36: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

24 Social Policy research Paper No. 44

The LSAC infant and preschooler samples may slightly under‑represent employed lone mothers and couple mothers who are breadwinners. The adolescent sample of employed mothers in the HILDA survey also slightly under‑represents mothers in two‑parent families who are breadwinners, but slightly over‑represents employed lone mothers. In addition, both the LSAC infant and preschooler samples and the HILDA survey adolescent samples slightly under‑represent fathers who are breadwinners, and over‑represent fathers in dual‑earner arrangements.

Overview of analysis approach

In this study, we analyse three cohorts of children: infants (3 to 19 months), preschoolers (4 and 5 year olds) and adolescents (15 to 17 year olds). Then we present analyses pertaining to these children and their parents. Cases are defined by each child, and we report results separately for sub‑samples of employed mothers and employed fathers. This means that in two‑parent families, both parents are included in the study. It also means that as we present results of children’s wellbeing separately for samples of employed mothers and employed fathers, some children are included in both the mother and father sub‑samples. Given this overlap, we do not make comparisons between the sub‑samples of employed mothers and employed fathers. Information is presented separately for each cohort to allow comparisons across the life course.

Our intention was to focus on life‑course variations in work and family arrangements, in terms of types of work arrangements and associations with parents’ and their children’s wellbeing. Most often, family life course is assessed in relation to the age of the youngest child. However, the LSAC and HILDA survey cohorts are based on the age of the study child, not the age of the youngest child. Thus, for example, some families in the preschooler or adolescent cohort also had infants, and vice versa. Therefore the cohorts are not equivalent to a life‑course classification. We decided to conduct analyses using cohorts to maximise the data on children’s outcomes (there is no sibling data in LSAC), but we also adjusted for age of youngest child in the multivariate models. We note that the mixture of sibling ages in the cohorts may reduce our capacity to detect family life‑course effects.

The analyses in this report take one of three forms. First, some analyses are descriptive analyses, meaning they describe the distribution of individual variables. Second are bivariate analyses, which explore the relationship between two variables. Third are multivariate analyses, which explore relationships between three or more variables. Multivariate analyses are often conducted in order to refine bivariate analyses, taking into account the effect of a third variable. In a multivariate analysis it would be possible to examine the effect of sex on psychological distress while removing the effect of income. Multivariate analyses can also be used to explore relationships between three or more variables of primary interest. For example, they could be used to determine whether the relationship between sex and psychological distress is similar for people of different ages. This report conducts multivariate analyses in both these ways.

When interpreting the analyses in this report it is important to note that the data are cross‑sectional, meaning they were collected at only one point in time. An implication of this is that while the data can identify associations between variables, it cannot be used to infer causation. For example, although we may observe an association between long work hours and psychological distress, we cannot determine that long work hours cause psychological distress because we do not know whether long work hours preceded psychological distress or vice versa.

We use a more stringent level of statistical significance than is standard (p<0.01 rather than p<0.05) in some sections of this report. This is because there are some homogeneity of variance violations and a large number of similar analyses in some sections. Using a more stringent significance level is a conservative approach designed to minimise the likelihood of false findings. A consequence of this, however, is that we may miss some findings that would otherwise have been statistically significant. With this in mind, we sometimes comment on non‑significant trends in the data, understanding that these trends may have reached significance had we used a more lenient approach. These trends should be interpreted with caution.

Each section includes a small method section outlining specific statistical analyses used and applicable methodological issues.

Page 37: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

25

SAMPlE ANd MEASurES

2.3 Caveats

Sampling design and sample weights

LSAC used a complex sampling design that was stratified by state and clustered by postcode. This sampling design violates the assumption of the independence of observations and may inflate the standard errors of estimates. Compared with 2001 census data, LSAC also oversampled and undersampled some population groups and so sample weights have been developed to address this problem.

To check the impact that stratifying, clustering and use of weighted data had on our results, we reran selected key findings after adjusting for these effects. We found that adjusting for clustering and/or using sample weighting has minimal impact on the point estimates or significance of main effects estimates in the multiple regressions, except where estimates are already marginal or borderline. Please see Appendix B for details on selected analyses of strata, clustering and weighting.

More substantial differences in the significance tests for interaction effects were observed, and this is relevant to the analyses reported in Section 6. The point estimates and standard errors changed little after adjustment for the design effect. After also adjusting for the sample weights, we found that there were only moderate changes in the standard errors, but the point estimates and test statistics for the interactions were more variable. This is expected, as the interaction term is a function of changes in four point estimates. We therefore recommend caution in interpreting these interactions.

Details of analyses adjusting for sample design and weights are available upon request.

restrictions on analyses due to small sample sizes

A key problem faced by the study was the small numbers of employed parents in some subgroups of interest. Table 2 presents descriptive statistics for sub‑sample sizes for employed mothers and employed fathers across the three cohorts, and the proportion of all employed mothers/fathers that they represent. There are very small numbers of breadwinner mothers in all three cohorts and a small number of parents in all adolescent subgroups. These groups are too small to produce meaningful results in high‑level analyses. Even very large differences between sub‑samples are unlikely to be statistically different with samples of this size and estimates will be unreliable. As such, these groups are not included in the multivariate analyses.

Page 38: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

26 Social Policy research Paper No. 44

Table 2: Employed mothers: sub‑samples in cohort analyses

Infants Preschoolers Adolescents

Mothers

All n 1,867 2,083 277

Low socioeconomic position

n(%)

283(15.2)

365(17.5)

63(22.7)

Med/high socioeconomic position

n(%)

1,584(84.8)

1,718(82.5)

214(77.3)

Lone parent n(%)

74(4.0)

198(9.5)

45(16.3)

Couple parent n(%)

1,793(96.0)

1,885(90.5)

232(83.7)

Breadwinner n(Employed mothers %)

(Couple mothers %)

63(3.4)(3.5)

64(3.1)(3.4)

8(2.8)(3.4)

Dual earner n(Employed mothers %)

(Couple mothers %)

1,730(92.7)(96.5)

1,821(87.4)(96.6)

224(80.9)(96.6)

Fathers

All n 2,832 2,633 270

Low socioeconomic position

n(%)

554(19.6)

495(18.8)

57(a)

(21.1)

Med/high socioeconomic position

n(%)

2,278(80.4)

2,138(81.2)

210(77.7)

Breadwinner n(%)

1,244(43.9)

968(36.8)

54(20.0)

Dual earner n(%)

1,588(56.1)

1,665(63.2)

216(80.0)

(a) Socioeconomic position could not be calculated for 3 employed fathers (1.1 per cent) in the HILDA survey owing to missing data on occupational prestige. To maximise power we have retained these cases in analysis.

Notes: Low‑SEP families scored in the bottom quartile on the measure of household socioeconomic position and medium/high‑SEP families scored in the top three quartiles on this measure. (For more detailed information on the derivation of household socioeconomic position, see Blakemore, Strazdins and Gibbings 2009). Breadwinner families have one parent employed and the other parent either unemployed or not in the labour force. Dual‑earner families have both the mother and father employed.

Page 39: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

27

SAMPlE ANd MEASurES

Employed mothers on leave

In the infant and preschooler samples, a number of the mothers who indicated they were employed in the last month were on paid/unpaid maternity leave when the data were collected (19.2 per cent in the infant sample and 2.2 per cent in the preschooler sample). However, the data do not record the timing or duration of this leave, thus the potential impact of these cases on our analyses is unclear. To investigate whether these cases had the potential to confound our findings, we conducted preliminary bivariate analyses comparing employed mothers on maternity leave with those not on leave on the work arrangements and parent outcome variables.

In the preschooler sample, employed mothers on leave differed from those not on leave in type of employment tenure only. In the infant sample, however, employed mothers on leave had less job security, more job control, more flexible hours, worked more hours and had more work–family conflict than those not on leave. To maximise power, we retained these cases in the sample and adjusted for whether employed mothers are on leave or not in the multivariate analyses. An implication of this are that some families classified as dual earner may in fact be comprised of an employed father and an employed mother who is on paid/unpaid maternity leave.

Details of these analyses are available in Appendix C.

Flexible hours

In LSAC, flexibility in work hours was assessed by the question ‘If you sometimes need to change the time when you start or finish your workday, is this possible?’. In the HILDA survey, respondents were asked to indicate whether flexible hours were available personally or for other employees working at a similar level to them in their workplace. The difference in these questions is problematic for comparison in this study as it is unclear from the HILDA survey whether respondents were personally entitled to flexible hours. Thus, comparisons of flexible hours in the LSAC and HILDA survey studies should be treated with caution.

Self‑report measures

Note that all measures in this study were self‑reported (or parent reports of child outcomes) and therefore subject to potential bias (social desirability, personality and mood).

2.4 Parent and child wellbeing measures

Parent mental health

LSAC infant and preschooler cohorts

Parental mental health was measured using the six‑item self‑report Kessler (K6) Scale (Kessler et al. 2003), which provides behavioural, emotional and cognitive indicators of non‑specific psychological distress. The scale consists of six questions that ask respondents how often they experienced symptoms of psychological distress during the past four weeks. Respondents use a five‑point scale ranging from ‘all of the time’ to ‘none of the time’. For a list of items, see Appendix D.

Two scores were derived from the K6 items: a sum score and a case threshold. The sum score is used to distinguish the severity of cases and the case threshold indicates the proportion of the sample screening positive for significant psychological distress.

To calculate the sum score, the original six items were first reverse scored and recoded so that scores ranged from 0 to 4 and higher values referred to higher levels of distress. These items were then summed to create a final mean score ranging from 0 to 24, with higher scores indicating higher levels of psychological distress.

Classification scores for the 10‑item Kessler Scale published in the National Mental Health Survey (ABS 2001c) were used to derive the case threshold for distress. In this survey, scores were grouped as follows: low (10–15);

Page 40: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

28 Social Policy research Paper No. 44

moderate (16–21); high (22–29); and very high (30–50). To derive the equivalent bandings for the K6, we multiplied scores by 0.66 (as the range of scores for the K6 was 0.66 that for the K10, K6 scores were grouped as follows: low (0–3); moderate (4–7); high (8–12); and very high (13–24)). Those who scored in the high banding or above were classified as above the threshold for psychological distress. Those who scored in the low or moderate bandings were classified as below the threshold for psychological distress.

Both mothers’ and fathers’ scale scores on the K6 had significant kurtosis (mothers: kurtosis=2.84; se kurtosis=0.05; fathers: kurtosis=4.45; se kurtosis=0.06). This was due to a high concentration of individuals with low scores on psychological distress. The scales demonstrated high inter‑item reliability for both mothers and fathers (mothers =0.84; fathers =0.81).

hIldA survey adolescent cohort

For the adolescent cohort, parent mental health was measured using the five‑item mental health subscale from the SF‑36 Health Survey (SF‑36) (Ware et al. 2000). Respondents indicated how frequently in the past four weeks they had felt feelings such as: ‘nervousness’ or ‘calm and peaceful’ on a scale ranging from 1 (‘all of the time’) to 6 (‘none of the time’). A full list of items is presented in Appendix D.

A total score for the SF‑36 mental health subscale was produced by summing raw scores and transforming these to a 0 to 100 standardised scale (for details on the transformation refer to Ware et al. 2000), where high scores represent better wellbeing. If valid responses were made on greater than or equal to half the items in the subscale, a person‑specific raw score was imputed for missing data based on the average score for the subscale.

We derived an indicator of whether respondents were below or above the clinical cut‑off on the mental health subscale using the cut‑off score developed by Newnham, Harwood and Page (2007). This cut‑off score was calculated from normal population data from the 1997 ABS National Health Survey and clinical patient data from a Perth clinic (eds Goode & Watson 2007). The cut‑off for mental health problems was 55, with those scoring at or below 55 considered dysfunctional and those scoring above 55 considered functional.

Both mothers’ and fathers’ transformed scores on the SF‑36 were normally distributed and demonstrated strong inter‑item reliability (mothers =0.83; fathers =0.80).

Infant wellbeing

Infant temperament was measured using items adapted from the Australian Temperament Project’s (ATP) Short Temperament Scale for Infants (STSI) (Sanson, Prior, Oberklaid & Northam 1987). The STSI is a 12‑item parent‑report scale that assesses temperament in children younger than 12 months old. The scale consists of three four‑item subscales: approachability, irritability and cooperativeness. The current study uses the irritability and approachability subscales only (for a list of items see Appendix D). The irritability subscale assesses how difficult an infant is to soothe and the approachability subscale assesses infants’ reactions to new situations and people. Responses were made on a six‑point scale ranging from 1 (‘almost never’) to 6 (‘almost always’). A total subscale score was calculated for each subscale by reverse scoring the items where appropriate and averaging the four items. Final subscale scores ranged from 1 to 6 and higher scores indicated more irritable or approachable behaviour. See Appendix D for the full items in each subscale.

The subscale scores on the STSI were normally distributed. The approachability subscale demonstrated very strong inter‑item reliability (=0.98) while the irritability subscale had poorer reliability (=0.57).

Preschooler wellbeing

Children’s behavioural and emotional problems were measured using a modified SDQ (Goodman 1997). The SDQ is a parent and teacher report inventory used to assess children on three types of problem behaviours: conduct/oppositional, hyperactivity/inattention and anxiety/depressive. Only parent‑rated SDQ scores are used in the current study (teacher‑rated SDQ is also collected in LSAC). The modified version used in LSAC consists of five subscales (25 statements in total) that assess aspects of the child’s behaviour over

Page 41: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

29

SAMPlE ANd MEASurES

the past six months (for the full list of items, see Appendix D). There are five items addressing emotional symptoms (for example, ‘often unhappy, downhearted or tearful’); five items for conduct/oppositional problems (for example, ‘often has temper tantrums or hot tempers’); five items for hyperactivity/inattention (for example, ‘restless, overactive, cannot stay still for long’)’ five items for peer relationship problems (for example, ‘gets on better with adults than with other children’) and five items for pro‑social behaviour (for example, ‘considerate of other people’s feelings’). Parents rated how true/typical the statement is of the child’s behaviour using a three‑point scale (1=‘not true’, 2=‘somewhat true’, 3=‘certainly true’).

Total SDQ sum scores were derived by summing the scores of the four ‘problem behaviour’ subscales (emotional, conduct, hyperactivity and peer problems). Thus, scores ranged from 0 to 38, with scores from 0–13 indicating normal behaviour, 14–16 indicating borderline behaviour and 17–40 indicating abnormal behaviour.

The parent‑rated SDQ scores were normally distributed and demonstrated acceptable levels of inter‑item reliability (=0.64).

Adolescent wellbeing

Adolescent mental health was measured using the same mental health subscale of the SF‑36 as used for parents, and with the same clinical cut‑off. Refer to the parent mental health section (HILDA survey adolescent cohort) for details. The values on the transformed adolescent mental health subscale (SF‑36) were normally distributed and the scale had good internal reliability (=0.79).

2.5 Measures of employment arrangements and parent work conditions

Employment definitions

All parent employment variables used in LSAC and the HILDA survey adhered to ABS labour statistic definitions. Employed refers to those parents working in paid employment for one hour or more per week and included parents who have a job that they are currently absent from for reasons such as maternity/parental leave, sick leave or strike. Unemployed refers to parents looking for work and ready to start a job. Not in the labour force refers to parents not in paid employment and not looking for employment, including those who are permanently unable to work, or permanently not intending to work (for example, in the case of retired individuals aged 65 years and over).

Family employment type (dual earner/breadwinner)

LSAC

Family employment type was derived from items assessing couple or lone parent status (‘does the study child have 2 parents in the home?’) and parent labour force status (if parents are employed, unemployed or not in the labour force). Two‑parent families were coded as either (1) dual earner (where both parents were employed) or (2) breadwinner (where only one parent was employed). Lone‑parent families were coded as (1) lone‑parent employed. (Note that our sample only includes families with one employed parent.)

HILDA survey

Family employment type was derived from the family‑type variable (see ‘Sex of parent and family type’ in Section 2.6) and the broad labour force status variables for the mothers and fathers (employed, unemployed or not in the labour force). Two‑parent families were coded as either (1) dual earner (where both parents were employed) or (2) breadwinner (where mother or father was employed). Lone‑parent families were coded as (1) lone parent employed.

Page 42: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

30 Social Policy research Paper No. 44

Parent work hours per week (continuous)

LSAC and the HILDA survey

In both LSAC and the HILDA survey, the number of work hours per week were measured using an item taken from the Labour Statistics Survey (ABS 2005), which assessed usual hours worked per week (including paid or unpaid overtime) in all jobs, but not including on‑call hours. Those who worked irregular hours were asked to provide an average over the last four weeks.

Categorical work hours

LSAC and the HILDA survey

In both studies, definitions of part time, standard full time, long and very long work hours were derived from Labour Statistics Survey (ABS 2005). Responses were coded as part time (34 hours or less per week); standard full time (35 or more per week); long work hours (46–54 hours per week); and very long work hours (55+ hours per week). These definitions of part and full‑time hours are consistent with the ABS definitions as outlined in Labour statistics: concepts, sources and methods (ABS 2005).

Those who worked irregular hours were asked to indicate the average hours per week over the past four weeks. Responses were coded as 1=works part time (1–34 hours) and 2=works full time (35 hours or more).

Total household work hours (all families)

LSAC and the HILDA survey

A measure of total household work hours for all families was derived from family employment type, indicators of whether both parents were employed and parents’ usual work hours per week. For breadwinner and lone (employed) parent households, the value was the number of work hours of the employed parent. For dual‑earner households, the value was the sum of the mother’s and father’s work hours. If the employment status or number of work hours was unknown for either parent, or if both parents were unemployed/not in the workforce, the value on the derived variable was treated as missing.

Job security

LSAC

Perceived job security was measured using one item from the PATH Through Life Project (CMHR 2005) job insecurity scale, which assessed overall feeling of security in current job. The item for LSAC was ‘How secure do you feel about your job or career in your current workplace?’ (1=very insecure to 4=very secure). Job security was dichotomised with the bottom two responses coded as low security and the top two responses coded as high security.

HILDA survey

Perceived job security was measured using an item adapted from the British Household Panel Survey questions on job satisfaction (see eds Goode & Watson 2007). The item was ‘I have a secure future in my job’ (1=strongly disagree to 7=strongly agree). The item was dichotomised such that those with values of 1 to 2 were coded as having low security and those with values from 3 to 7 coded as having high security on each item.

Page 43: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

31

SAMPlE ANd MEASurES

Employment tenure

LSAC

Employment tenure was derived from two items adapted from the Labour Statistics Survey (ABS 2005): one assessing self‑employment and one assessing job tenure. The self‑employment item asked: ‘Does the parent work for an employer or in the parent’s own business?’ (1=employer; 2=own business, 3=other/uncertain). Those who worked for their own business were coded as self‑employed; those who worked for an employer were coded as not self‑employed. All other responses were coded as missing.

The job tenure of employees was coded into four categories: (1) permanent or ongoing position; (2) fixed term tenure; (3) casual basis; and (4) on some other basis. A final measure of employment tenure was derived from the responses to these two variables so that self‑employed individuals were coded as (1), and those who were not self‑employed were further classified into categories based on job tenure (2=permanent; 3=fixed; and 4=casual). Those employed ‘on some other basis’ were coded as missing.

HILDA survey

Employment tenure was derived from two HILDA survey derived variables from the labour mobility module of the February 2000 ABS Monthly Population Survey: one assessing self‑employment and one assessing type of employment tenure (see eds Goode & Watson 2007).

The self‑employment variable classified respondents into four categories (1=employee; 2=employee of own business; 3=employer/self‑employed; and 4=unpaid family worker). We then recoded this variable so that employees and unpaid family workers were coded as 0=not self‑employed. Those who indicated they were employees of their own business or employer/self‑employed were coded as 1=self‑employed.

The employment tenure variable classified respondents into four categories: 1=fixed‑term tenure; 2=employed on a casual basis; 3=employed on an ongoing basis; and 4=other. The final employment tenure variable was derived from the responses to these two variables so that self‑employed individuals were coded as (1), and those who were not self‑employed were further classified into categories based on job tenure (2=permanent; 3=fixed; and 4=casual). Those employed on another basis were coded as missing.

Job control

LSAC and the HILDA survey

Job control was measured using one item: ‘I have a lot of freedom to decide how I do my job’. Responses in LSAC were measured on a five‑point scale ranging from 1 (strongly disagree) to 5 (strongly agree). The HILDA item was answered on a seven‑point scale (ranging from ‘strongly disagree’ to ‘strongly agree’). These items were then dichotomised by assigning those with values of 1 to 2 a score of 0=low control and those with values from 3 to 7 a score of 1=high control.

Flexible hours

LSAC

Flexibility in work hours was measured using one item from the National Study of the Changing Workforce (Bond et al. 2002). This item assessed the degree of flexibility participants had in changing the finish or start time of the workday. The item was ‘If you sometimes need to change the time when you start or finish your workday, is this possible?’ (1=yes, I am able to work flexible hours to 4=no, definitely not). Responses were reverse scored so that higher scores indicated greater flexibility; recoded into three categories so that 3=flexible hours, 2=some flexibility and 1=non‑flexible hours; and then dichotomised so that 0=no, 1=yes.

Page 44: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

32 Social Policy research Paper No. 44

HILDA survey

Flexible work hours were assessed by one item adapted from the employment benefits module of the HILDA survey (see eds Goode & Watson 2007) asking whether respondents were entitled to flexible start and finish times.

Note that the question wording for this item was problematic because respondents were asked to indicate ‘if these entitlements were available personally or for other employees working at a similar level to them in their workplace’. Thus, it is unclear whether the respondents themselves were entitled to flexible hours.

2.6 Family characteristics and sociodemographic measures

Sex of parent and family type

LSAC and the HILDA survey

Sex of parent was measured by a self‑report item with dichotomous response format (male/female). This item was then used to classify families into five types: (1) one male parent; (2) one female parent; (3) one male and female parent; (4) two male parents; and (5) two female parents. As there were no families with two male parents and very few with either one male or two female parents, we excluded these cases from subsequent analyses. Therefore, the final measure of family type categorised families as either (1) lone mother or (2) two parent.

Number of resident children in household

LSAC

The number of children in the household is derived from demographic items taken from the family characteristics module of the Australian Bureau of Statistics 2001 Census (ABS 2001b).

HILDA survey

For the HILDA survey, the number of resident children in the household was derived from two demographic items on resident children’s age and intra‑household relationships based on items from the British Household Panel Survey and the US Panel Study of Income Dynamics (see eds Goode & Watson 2007). An intra‑household relationship grid showed how each person is related to the reference person in the household. Note that in the HILDA survey, step or foster children are excluded from this variable.

Age of youngest child in household

LSAC

The age of the youngest child in the household was derived from demographic items detailing the study child age in months, and the age in years of each of the siblings in the households based on items adapted from the family characteristics module of the Australian Bureau of Statistics 2001 Census (ABS 2001b). To derive a measure of the age of the youngest resident child, we first converted the age of study child in months to years. We then derived a measure of the age of youngest resident child by computing the minimum age out of the age of study child in years and the age in years of each sibling in the household.

HILDA survey

This variable was derived by the HILDA survey based on responses to two demographic items on resident children’s age and intra‑household relationships based on items from the British Household Panel Survey (see eds Goode & Watson 2007).

Page 45: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

33

SAMPlE ANd MEASurES

Parent age

LSAC

Parent age was measured by the item ‘What was the parent’s age last birthday?’ based on age at the time the survey was undertaken. Age was recorded in years.

HILDA survey

Parent age was measured by one item asking the parent’s age last birthday as at 30 June 2004. Age was given in years. Participants aged 90 years and over were coded as aged 90 years to protect respondent confidentiality.

Child care

LSAC

Regularity of the use of child care was measured by one item developed for LSAC: ‘Study child has been looked after regularly by others’. Response choices were ‘yes’ or ‘no’.

HILDA survey

Use of child care was assessed by one item adapted closely from the Negotiating the Life Course Survey (see eds Goode & Watson 2007): ‘Have you used or thought about using child care (see list of types of child care below) so you can undertake paid work in last 12 months?’ Responses were coded as: 1=used or thought about using child care in last 12 months and 2=hasn’t used child care in last 12 months.

Financial hardship

LSAC and the HILDA survey

Financial stress and hardship were measured using six parent interview items adapted from the Australian Bureau of Statistics Household Expenditure Survey (Bray 2001). These items assessed various aspects of financial hardship experienced over the past 12 months, such as ability to pay rent/mortgage, pay bills on time, provide heating or cooling, provide meals, as well as needing to pawn or sell something, or seek welfare assistance.

Items were answered using a yes/no response format. The scale was scored by summing the number of ‘yes’ responses so that scores ranged from 0=no hardship to 6=significant hardship. As the scale scores were significantly skewed with the majority of scores in the no hardship end, we recoded the hardship scale score to be a dichotomous score. Those scoring from 0 to 2 were recoded as having no hardship and those scoring from 3 to 6 were recorded as having hardship. The financial hardship scale had acceptable inter‑item reliability for LSAC (=0.67) and the HILDA survey (mothers’ version =0.73; fathers’ version =0.71).

Parent years of education

In LSAC, a measure of parent years of education was derived from the following three variables:

�� highest level of primary/secondary schooling attained

�� whether the respondent had achieved a trade certificate or other education qualification

�� highest post‑secondary qualification level attained.

In the HILDA survey, a measure of parent years of education was derived from the following three derived variables:

�� highest level of school

Page 46: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

34 Social Policy research Paper No. 44

�� age left school

�� highest post‑secondary qualification level.

In both datasets, parent years of education were top‑coded so that a parent with a postgraduate degree was allocated the highest value (20 years of education) regardless of their actual years of education. This approach is based on the assumption that one individual with a postgraduate degree has relatively equal levels of employment opportunities as other postgraduate degree holders regardless of their previous years of education. Therefore, the years of education allocated for a particular qualification are based on the implications of possessing that qualification for level of workforce participation. Higher scores therefore indicate higher educational attainment but not necessarily more years of education. For more information on top‑coding parents’ years of education, see Blakemore, Gibbings and Strazdins (2006).

Combined parent annual income

LSAC

Combined parent annual income was measured using one item modified from the National Longitudinal Survey of Children and Youth (NLSCY) 1994–1995 (Statistics Canada 1995). This item assesses present yearly combined income for both parents and includes pensions and allowances before tax, superannuation or health insurance. Participants self‑reported the appropriate gross income range from a choice of 15 or selected the nil or negative income responses.

Total household income

HILDA survey

All income items in the HILDA survey are adapted from the Australian Bureau of Statistics Survey of Income and Housing Costs 1999–2000 (eds Goode & Watson 2007).

We derived a measure of total household income by summing the values from three HILDA survey derived income variables: (1) positive household income (imputed); (2) negative household income (imputed); and (3) household windfall (irregular) income. The positive and negative income variables included income from all private (market and private transfers) and public (pension and benefit) sources, summed to the household level. Note these two variables included Family Tax Benefits and Child Care Benefits, but excluded windfall (irregular) income sources. The windfall income variable included irregular sources of income such as inheritances, redundancies and payments from parents but did not include resident parent transfers. For further details on the derivation of these variables, see the Financial Year Income Model diagrams in the HILDA User Manual (eds Goode & Watson 2007).

For all three income variables, missing data were imputed using the longitudinal imputation method developed by Little and Su (1989). Income data were imputed based on trend (wave) and individual (person) level information. For new entrants interviewed in Wave 4 who did not respond to some of the income questions, the nearest neighbour regression imputation method was used. Also prior to imputation, household population weights were applied to income variables to adjust for potential selection bias. For a more detailed discussion of the use of weights in the HILDA survey, see HILDA Project Technical Paper Series No. 1/04: Wave 2 Survey Methodology (Watson & Wooden 2004).

Family socioeconomic position

LSAC and the HILDA survey

Family socioeconomic position was derived for both LSAC and the HILDA survey datasets using the method developed by Blakemore, Strazdins and Gibbings (2009). The term ‘socioeconomic position’ refers to the relative position of families in terms of their social and economic resources available to family members, and is

Page 47: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

35

SAMPlE ANd MEASurES

often used interchangeably with terms such as socioeconomic status, social class, social stratification, social inequity and social status (Blakemore, Strazdins & Gibbings 2009). We received permission from the authors to use the final socioeconomic position scores that they generated for LSAC and the HILDA survey.

Blakemore and colleagues’ method constructs socioeconomic position from standardised scores for three variables (as described above):

�� household income (total household income for the HILDA survey, combined yearly income for LSAC; see above). Note we performed a natural log transformation on the income sources to correct for differences in distributions

�� mothers’ and fathers’ years of education

�� occupational status (as determined by main occupation and occupation status).

A continuous measure of socioeconomic position was derived from standardised z scores for the total household income, mothers’ and fathers’ years of education, and occupational prestige.

For LSAC, we used the socioeconomic position scores calculated for all families (that is, with lone and two‑parent families in the same distribution). In the HILDA survey, socioeconomic position was derived for all parents with valid data on all the component variables, by calculating the mean of the standardised measures of mother and father years of education, mother and father occupational prestige, and the log of total household income. If either parent had missing data on any of the component variables, the socioeconomic position variable was coded as missing.

In both LSAC and the HILDA survey, families whose socioeconomic position score was in the bottom quartile were termed ‘low SEP’. The remaining families whose socioeconomic position score was in the top three quartiles were termed ‘medium to high SEP’.

Page 48: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

36 Social Policy research Paper No. 44

Page 49: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

37

FAMIlIES ACrOSS ThE COhOrTS: SOCIOdEMOgrAPhIC ChArACTErISTICS ANd WEllbEINg

3 Families across the cohorts: sociodemographic characteristics and wellbeing

3.1 Overview and aims

This section describes the key similarities and differences in the sociodemographic characteristics and wellbeing of families. It has two key aims. The first is to assess whether there are clear differences across the cohorts in hardship, education or household composition. These sociodemographic differences form the backdrop to our research questions on work, families and wellbeing. The second aim is to investigate employed parents’ and their children’s wellbeing across and within each cohort. Does mothers’ and fathers’ wellbeing vary by earner arrangements, socioeconomic position or between lone or two‑parent families? Does infant, preschool or adolescent child wellbeing differ between breadwinner and dual‑earner families? Is child wellbeing different in lone‑parent families with an employed mother compared to two‑parent families with an employed mother? Is there evidence of wellbeing disparities linked to socioeconomic position?

3.2 Methods

We conceptualised wellbeing as a multidimensional construct that, for parents, was indicated by mental health‑related measures (following the 1948 World Health Organization’s definition (WHO 2007)).

For children and adolescents, wellbeing was measured by social and emotional functioning. Wellbeing is often considered to have an economic dimension (family income and standard of living). However, in this study we treated economic factors as a driver of wellbeing, comparing differences in parent and child wellbeing in socioeconomically advantaged and disadvantaged families. Reciprocal relationships probably also exist, whereby poorer wellbeing limits parents’ capacity to earn income, but testing this would require multi‑wave data.

For employed parents, we explore psychological distress. Psychological distress provides an indication of parents’ overall mental health and may indicate susceptibility to anxiety and depression. High levels of psychological distress indicate clinically significant mental health symptoms and likely impairment in functioning.

For children and adolescents, social and emotional functioning is our core indicator of wellbeing. Social and emotional functioning is important in its own right, but also can shape children’s later life chances and social participation (Keating & Hertzmann 1999). In this study, we examine infants’ temperament as a proxy for social or emotional function. It should be noted that temperament involves relatively innate aspects of children’s functioning and it is not a measure of mental health. However, we use validated scales to assess preschoolers’ social and behavioural problems and adolescent mental health scores (see Section 2 for details on each wellbeing measure).

Note that we were unable to completely harmonise wellbeing measures across cohorts. Parent psychological distress was measured differently in the HILDA survey than in LSAC, so we are unable to statistically compare the adolescent cohort with the infant and preschooler cohorts. Similarly, because the infant, preschooler and adolescent outcomes are measured using different variables with different scales, we were unable to statistically test differences in child wellbeing across cohorts.

Page 50: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

38 Social Policy research Paper No. 44

This section is presented in three parts. The first section describes the sociodemographic characteristics of three cohorts and then examines any differences in demographics based on family socioeconomic position and lone parenthood (mothers only, the sample of single parent fathers is too small to analyse). The second section describes parent wellbeing for the whole sample and by socioeconomic position, lone parenthood, and family earner arrangements (dual earner compared with breadwinner). The third and final section examines children’s wellbeing via these same categories.

To test for statistically significant differences across the cohorts, we use independent sample t‑tests for continuous variables and chi‑square tests for categorical variables. Due to unequal variances and the number of comparisons, we use a conservative (p<0.01) level for reporting significant differences. Note that we found few differences in comparisons involving the adolescent cohort, possibly linked to small sample sizes and hence less reliable estimates. Note also that these analyses are bivariate and therefore do not adjust for potential confounders; see Section 2 for a discussion of significance levels and bivariate versus multivariate analyses.

3.3 Family characteristics and socioeconomic demographics

This section examines family sociodemographic characteristics for employed mothers’ and fathers’ in each cohort. Table 3 displays these results. Not surprisingly, parents’ age, age of youngest child and family size (number of children in the family) increased across the three cohorts, while the likelihood of having an infant or using child care decreased. On average, parents in the infant cohort were in their early thirties, parents in the preschooler cohort were in their mid‑thirties and parents in the adolescent cohort were in their mid‑forties. For families in the infant cohort, the average age of the youngest child was around 8 months, around 90 per cent of families had an infant aged 12 months or less, and the majority of families had only one child. Most families in the preschooler cohort had a toddler as the youngest child and had two children. Very few families in the adolescent cohort had an infant, the average age of the youngest child was around 12 years, and the majority of families had three or more children. While child care use was 40 to 50 per cent in the infant and preschooler cohorts, less than 20 per cent of families with an adolescent used child care.

Table 3 also shows socioeconomic characteristics of mothers and fathers for each cohort. The average income and proportion of families experiencing financial hardship were similar for mothers and fathers in all cohorts, with approximately 20 per cent of families reporting financial hardship. It is difficult to directly compare family income across cohorts as different measures were used in LSAC and the HILDA survey (LSAC measured combined parent income and included both salary and other sources of income, whereas the HILDA survey used household income that included income from all family members, not just parents). However, estimates based on combined parent salary in the adolescent cohort indicate that there is little, if any, change across the cohorts, with all showing a median parent income of between $52,000 and $77,999.

Mothers’ and fathers’ had an average of about 14 or 15 years of education in all cohorts, which is approximately equivalent to bachelor degree or similar qualification. Mothers’ and fathers’ in the adolescent cohort tended to have slightly fewer years of education compared with their counterparts in the younger cohorts.

Page 51: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

39

FAMIlIES ACrOSS ThE COhOrTS: SOCIOdEMOgrAPhIC ChArACTErISTICS ANd WEllbEINg

Tabl

e 3:

Fa

mily

cha

ract

eris

tics

of e

mpl

oyed

mot

hers

and

fath

ers

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Mot

her

n=1,

867

Fath

ern=

2,83

2M

othe

rn=

2,08

3Fa

ther

n=2,

633

Mot

her

n=27

7Fa

ther

n=27

0

Youn

gest

chi

ld (y

ears

/mon

ths)

Mea

n8

m8

m3

y, 5

m3

y, 1

m12

y, 6

m11

y, 8

m

Mot

her a

ge (y

ears

)M

ean

31.8

031

.56

35.2

935

.32

44.1

343

.52

Fath

er a

ge (y

ears

)M

ean

33.9

333

.69

37.7

237

.53

45.9

745

.61

Infa

nt in

fam

ily%

91.8

093

.40

16.6

022

.50

0.70

2.60

No.

of c

hild

ren

Mea

n1.

771.

902.

292.

454.

385.

04

1%

46.0

040

.00

11.9

07.

906.

503.

00

2%

36.4

037

.90

56.0

054

.00

19.9

010

.70

3–4

%16

.50

20.5

030

.80

35.2

037

.90

40.0

0

5+%

1.10

1.70

1.30

2.80

35.7

046

.30

Use

s ch

ild c

are(a

)%

56.8

038

.10

55.1

040

.60

17.0

015

.60

Hou

seho

ld in

com

e (p

.a.)

Med

ian

$52,

000–

$77,

999

$52,

000–

$77,

999

$52,

000–

$77,

999

$52,

000–

$77,

999

$92,

270

($75

,600

)(b)

$96,

150

($77

,712

)(b)

Fina

ncia

l har

dshi

p%

22.4

023

.30

23.2

021

.40

23.5

021

.10

Pare

nt e

duca

tion

(yea

rs)

Mea

n15

.58

14.9

815

.09

14.9

814

.08

14.7

6

Inco

mpl

ete

seco

ndar

y%

6.40

10.1

013

.70

12.5

032

.90

20.4

0

Com

plet

e se

cond

ary

%13

.40

10.2

013

.30

9.00

11.2

04.

40

Cert

ifica

te/D

iplo

ma

%34

.90

47.6

035

.30

45.6

026

.40

42.2

0

Bac

helo

r or h

ighe

r%

45.3

032

.10

37.7

032

.90

29.6

033

.00

(a)

In th

e ad

oles

cent

sam

ple,

onl

y fa

mili

es w

ith

child

ren

unde

r 14

year

s w

ere

aske

d ab

out c

hild

car

e us

e. T

hus,

fam

ilies

wit

hout

chi

ldre

n yo

unge

r tha

n th

e st

udy

child

(a

ged

15 to

17

year

s) w

ere

not a

sked

. We

repo

rt th

e pe

rcen

tage

of t

he e

ntir

e ad

oles

cent

sam

ple

usin

g ch

ild c

are.

(b)

Hou

seho

ld in

com

e in

the

adol

esce

nt s

ampl

e in

clud

es in

com

e fr

om a

ll fa

mily

mem

bers

. Com

bine

d pa

rent

ann

ual s

alar

y is

not

ed in

par

enth

eses

.N

otes

: D

ue to

roun

ding

, per

cent

ages

may

not

add

up

to 1

00 p

er c

ent.

m re

fers

to m

onth

s, y

to y

ears

.

Page 52: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

40 Social Policy research Paper No. 44

3.4 Sociodemographic differences by disadvantage

The first disadvantage we examine is family socioeconomic position. Socioeconomic position represents a relative rank combining measure of parent education, occupational prestige and family income (see Section 2.6 for details of how socioeconomic position was coded). Families falling in the bottom quartile are considered to have low socioeconomic position.

Parents from low‑SEP families were younger than those from the more advantaged, medium/high‑SEP families in all three cohorts (see Table 4). Family size was associated with socioeconomic position in different ways across cohorts. In the preschooler cohort, there were slightly more children per family in low‑SEP compared with medium/high‑SEP families (this difference was significant in the employed father sample), while in the adolescent cohort there were slightly more children in the medium/high‑SEP than in low‑SEP families (significant in the employed mother comparisons).

Our measure of child care use covered both formal and informal care. Low‑SEP families in the infant cohort were less likely to use child care compared with medium/high‑SEP families, and this difference was apparent for both the employed mother and father sub‑samples. In the preschooler cohort, low‑SEP families with an employed father were less likely to use child care.

Not surprisingly, low‑SEP families reported lower median income compared with more advantaged families in each cohort (see Table 5). There were also clear differences between the extent to which low‑SEP families experienced financial hardship compared with the medium/higher‑SEP families. In nearly all comparisons, low‑SEP families were approximately twice as likely to experience financial hardship.

Parents from medium/high‑SEP families had more years of education than those from low‑SEP families in all three cohorts. These results are as we would expect, and indicate that the socioeconomic position measure correctly classifies families in each of the three cohorts. Note, however, the lower income for medium/high‑SEP families in the infant cohort compared with medium/high‑SEP families in the preschooler and adolescent cohorts. This may be due to the proportion of mothers with infants on unpaid leave (still in the labour market but not contributing income).

Page 53: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

41

FAMIlIES ACrOSS ThE COhOrTS: SOCIOdEMOgrAPhIC ChArACTErISTICS ANd WEllbEINg

Tabl

e 4:

Em

ploy

ed m

othe

rs’ a

nd fa

ther

s’ fa

mily

cha

ract

eris

tics

by s

ocio

econ

omic

pos

ition

: with

in‑c

ohor

t com

paris

ons

Mot

hers

Fath

ers

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Low

Med

/hig

hLo

wM

ed/h

igh

Low

Med

/hig

hLo

wM

ed/h

igh

Low

Med

/hig

hLo

wM

ed/h

igh

n=28

3n=

1,58

4n=

365

n=1,

718

n=63

n=21

4n=

554

n=2,

278

n=49

5n=

2,13

8n=

57n=

210

Youn

gest

chi

ld

(yea

rs/m

onth

s)M

ean

8 m

9 m

3 y,

7 m

3 y,

4 m

*12

y, 9

m12

y, 6

m8

m9

m3

y, 2

m3

y, 2

m11

y, 6

m11

y, 9

m

Age

(par

ent)

(yea

rs)

Mea

n29

.56

32.2

0*33

.59

35.6

5*42

.88

44.5

032

.07

34.0

8*36

.29

37.8

2*44

.08

45.9

3*

Infa

nt in

fam

ily%

92.6

091

.70

14.8

017

.10

0.00

0.90

94.9

093

.10

22.4

022

.10

3.50

2.40

No.

of c

hild

ren

Mea

n1.

741.

722.

252.

213.

504.

63*

1.89

1.82

2.42

2.31

*4.

795.

14

1%

45.0

046

.00

12.6

011

.80

15.9

03.

70*

38.1

040

.40

8.30

7.90

5.30

2.10

2%

35.0

036

.70

52.1

056

.80

28.6

017

.30*

37.2

038

.00

46.5

055

.70*

12.3

09.

50

3–4

%17

.70

16.3

033

.40

30.2

031

.70

39.7

0*22

.00

20.1

040

.40

34.1

0*40

.40

40.0

0

5+%

1.40

1.00

1.90

1.20

23.8

039

.30*

2.70

1.40

4.80

2.40

*42

.10

48.2

0

Use

s ch

ild c

are(a

)%

50.2

058

.00*

55.6

054

.90

12.7

018

.20

23.6

041

.60*

34.5

042

.00*

8.80

17.6

0

(a)

In th

e ad

oles

cent

sam

ple,

onl

y fa

mili

es w

ith

child

ren

unde

r 14

year

s w

ere

aske

d ab

out c

hild

car

e us

e. W

e re

port

the

perc

enta

ge o

f the

ent

ire

adol

esce

nt s

ampl

e us

ing

child

car

e.N

otes

: Si

gnifi

canc

e te

sts

com

pare

low

and

med

ium

/hig

h‑S

EP fa

mili

es w

ithi

n ea

ch c

ohor

t. D

ue to

roun

ding

, per

cent

ages

may

not

add

to 1

00 p

er c

ent.

m re

fers

to m

onth

s, y

to y

ears

. *p

<0.

01.

Page 54: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

42 Social Policy research Paper No. 44

Tabl

e 5:

Em

ploy

ed m

othe

rs’ a

nd fa

ther

s’ s

ocio

econ

omic

cha

ract

eris

tics

by s

ocio

econ

omic

pos

ition

: with

in‑c

ohor

t com

paris

ons

Mot

hers

Fath

ers

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Low

Med

/hig

hLo

wM

ed/h

igh

Low

Med

/hig

hLo

wM

ed/h

igh

Low

Med

/hig

hLo

wM

ed/h

igh

n=28

3n=

1,58

4n=

365

n=1,

718

n=63

n=21

4n=

554

n=2,

278

n=49

5n=

2,13

8n=

57n=

210

Hou

seho

ld

inco

me

(p.a

.)M

edia

n$4

1,60

0–$5

1,99

9$5

2,00

0–$7

7,99

9$4

1,60

0–$5

1,99

9$7

8,00

0–$1

03,9

99$6

0,36

6($

28,3

47)(a

)$1

18,2

58($

89,4

05)(a

)$4

1,60

0–$5

1,99

9$5

2,00

0–$7

7,99

9$4

1,60

0–$5

1,99

9$7

8,00

0–$1

03,9

99$6

4,23

3($

52,0

00)(a

)$1

04,8

18($

85,6

22)(a

)

Fina

ncia

l ha

rdsh

ip%

37.8

019

.70*

38.4

020

.00*

36.5

019

.60*

41.0

019

.00*

39.6

017

.20*

29.8

018

.60

Educ

atio

n (y

ears

)M

ean

13.1

216

.02*

12.5

515

.63*

11.9

214

.71*

12.8

815

.49*

12.6

315

.52*

11.9

615

.56*

Inco

mpl

ete

seco

ndar

y%

26.5

02.

80*

47.4

06.

50*

65.1

023

.40*

*34

.10

4.20

*42

.40

5.50

*63

.20

8.10

*

Com

plet

e se

cond

ary

%27

.90

10.9

0*20

.50

11.8

0*11

.10

11.2

016

.60

8.70

*10

.10

8.80

*5.

304.

30

Cert

ifica

te/

Dip

lom

a%

43.8

033

.30*

31.0

036

.30*

23.8

027

.10

48.2

047

.50*

46.9

045

.30*

29.8

045

.70*

Bac

helo

r or

high

er%

1.80

53.0

0*1.

1045

.50*

0.00

38.3

0*1.

1039

.60*

0.60

40.4

0*1.

8041

.90*

(a)

Hou

seho

ld in

com

e in

the

adol

esce

nt s

ampl

e in

clud

ed in

com

e fr

om a

ll fa

mily

mem

bers

. Com

bine

d pa

rent

ann

ual s

alar

y (b

ut n

ot in

clud

ing

bene

fits

or o

ther

inco

me)

is n

oted

in

pare

nthe

ses.

Not

es:

Sign

ifica

nce

test

s co

mpa

re lo

w a

nd m

ediu

m/h

igh‑

SEP

fam

ilies

wit

hin

each

coh

ort.

Due

to ro

undi

ng, p

erce

ntag

es m

ay n

ot a

dd to

100

per

cen

t. *

p<0.

01; *

*p<

0.00

1.

Page 55: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

43

FAMIlIES ACrOSS ThE COhOrTS: SOCIOdEMOgrAPhIC ChArACTErISTICS ANd WEllbEINg

We now examine the sociodemographics of employed mothers in lone and two‑parent families (see Table 6), as lone parenthood is often associated with more disadvantaged circumstances. Employed lone mothers were, on average, a year younger than employed mothers from two‑parent families in the preschooler cohort, and three years older than coupled mothers in the adolescent cohort.

In the preschooler and adolescent cohorts, lone mothers had fewer dependent children than coupled mothers (this difference was particularly marked in the adolescent cohort). In the adolescent cohort, the average number of children in lone‑mother families was two, compared with four children in two‑parent families. In the preschooler cohort, lone‑mother families were substantially less likely to have an infant in the household than two‑parent families.

Employed lone mothers were also much more likely than mothers in two‑parent families to use child care in the infant and preschooler cohorts. This suggests that lone mothers are more dependent on child care to manage work and care responsibilities. Our study does not distinguish between types of child care used. However, other research indicates that lone mothers tend to be more reliant on informal care than two‑parent families (Craig 2004).

In all three cohorts, lone‑mother families had substantially less household income and experience more financial hardship than two‑parent families where the mother was employed. Lone mothers were more than twice as likely to report experiencing financial hardship as mothers in two‑parent families in every cohort. Lone mothers with preschoolers tended to have fewer years of education than mothers in two‑parent families.

Taken together, these results indicate that even when the mother is employed, many lone‑mother families remain socioeconomically disadvantaged.

Page 56: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

44 Social Policy research Paper No. 44

Tabl

e 6:

Em

ploy

ed m

othe

rs’ f

amily

cha

ract

eris

tics

by fa

mily

type

: with

in‑c

ohor

t com

paris

ons

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Lone

n=74

Coup

len=

1,79

3Lo

nen=

198

Coup

len=

1,88

5Lo

nen=

45Co

uple

n=23

2

Youn

gest

chi

ld (y

ears

/mon

ths)

Mea

n9

m8.

5 m

4 y,

3 m

3 y,

4 m

*13

y, 6

m12

y, 4

m

Mot

her a

ge (y

ears

)M

ean

29.8

731

.88

34.0

735

.42*

46.2

043

.73*

Infa

nt in

fam

ily%

90.5

091

.90

4.00

17.9

0**

0.00

0.90

No.

of c

hild

ren

Mea

n1.

621.

782.

002.

32**

2.13

4.81

**

1%

56.8

045

.60

29.8

010

.00*

*24

.40

3.00

2%

27.0

036

.80

48.5

056

.80

46.7

014

.70

3–4

%16

.20

16.5

020

.80

31.8

028

.90

39.7

0

5+%

0.00

1.10

1.00

1.40

0.00

42.7

0

Child

car

e(a)

%73

.00

56.1

0*73

.70

46.9

0*17

.80

16.8

0

Hou

seho

ld in

com

e (p

.a.)

Med

ian

$26,

000–

$31,

199

$52,

000–

$77,

999

$31,

200–

$36,

399

$52,

000–

$77,

999

$51,

590

($27

,000

)(b)

$99,

556

($81

,547

)(b)

Fina

ncia

l har

dshi

p%

52.7

021

.20*

*49

.50

20.5

0**

44.4

019

.40*

Mot

her e

duca

tion

(yea

rs)

Mea

n15

.02

15.6

014

.25

15.1

7**

14.0

214

.09

Inco

mpl

ete

seco

ndar

y%

4.00

6.50

*21

.20

12.9

0**

31.1

033

.20

Com

plet

e se

cond

ary

%14

.90

13.4

0*14

.60

13.2

0**

11.1

011

.20

Cert

ifica

te/D

iplo

ma

%55

.40

34.0

0*42

.40

34.6

0**

33.3

025

.00

Bac

helo

r or h

ighe

r%

25.7

046

.10*

21.7

039

.40*

*24

.40

30.6

0

(a)

In th

e ad

oles

cent

sam

ple,

onl

y fa

mili

es w

ith

child

ren

unde

r 14

year

s w

ere

aske

d ab

out c

hild

car

e us

e. W

e re

port

the

perc

enta

ge o

f the

ent

ire

adol

esce

nt s

ampl

e us

ing

child

car

e.(b

) H

ouse

hold

inco

me

in th

e ad

oles

cent

sam

ple

incl

udes

inco

me

from

all

fam

ily m

embe

rs. C

ombi

ned

pare

nt a

nnua

l sal

ary

is n

oted

in p

aren

thes

es.

Not

es:

Sign

ifica

nce

test

s co

mpa

re lo

ne a

nd tw

o‑pa

rent

fam

ilies

wit

hin

each

coh

ort.

Due

to ro

undi

ng, p

erce

ntag

es m

ay n

ot a

dd to

100

per

cen

t. m

refe

rs to

mon

ths,

y to

yea

rs.

*p<

0.01

; **p

<0.

001.

Page 57: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

45

FAMIlIES ACrOSS ThE COhOrTS: SOCIOdEMOgrAPhIC ChArACTErISTICS ANd WEllbEINg

3.5 Employed parents’ wellbeing

Table 7 presents descriptive statistics of parent wellbeing in the three cohorts for the whole sample and then split by family socioeconomic position, earner arrangement and family structure (mother only). Note that across‑cohort comparisons could not be made with the adolescent cohort as the measures of wellbeing are different from those used in the younger two cohorts.

Across the cohorts, approximately 10 per cent of mothers reported experiencing elevated levels of psychological distress, with slightly fewer fathers reaching the threshold. Compared with those in the infant cohort, employed mothers from the preschooler cohort experienced more psychological distress and were more likely to score above the clinical cut‑off on the K6. Employed mothers in the adolescent cohort appeared to have a lower prevalence of psychological distress compared with mothers in the preschooler or infant cohort, although we could not test this difference statistically.

Family socioeconomic position did not seem to be associated with employed mothers’ psychological distress. Mothers from medium/high‑SEP and low‑SEP families experienced similar levels of psychological distress in all three cohorts. Across cohorts, there appeared to be a higher proportion of mothers above the clinical cut‑off for distress in the preschooler cohort than in the other two cohorts, although the comparison with the infant cohort did not reach statistical significance.

In preschooler cohort, more fathers from low‑SEP families than medium/high SEP families were above the K6 clinical cut‑off. However, employed fathers in the infant and preschooler cohorts showed no statistical differences in average level of psychological distress as a function of family socioeconomic position. Similarly, in the adolescent cohort, fathers from more advantaged families had slightly better mental health than fathers from low‑SEP families, as evidenced by higher scores on the mental health subscale of the SF‑36. Note that for both employed mothers and fathers, relatively large standard deviations indicate that mean scores varied widely within sub‑samples.

Lone mothers in the infant and preschooler cohorts showed higher levels of psychological distress, and were much more likely to be above the clinical cut‑off on the K6, compared with mothers from two‑parent families. In the infant cohort, the proportion of lone mothers scoring above the cut‑off was more than twice the proportion of partnered mothers. A similar trend of elevated distress among lone mothers was observed for mothers’ scores on the mental health subscale in the adolescent cohort; however, the differences were smaller and not statistically significant.

Breadwinner (where only one parent is employed) and dual‑earner (where both parents are employed) mothers did not show statistically significantly differences in levels of psychological distress. Note that all estimates for breadwinner mothers are likely to be imprecise and the statistical tests for differences underpowered, given the very small samples. Breadwinner and dual‑earner fathers also showed no differences in average psychological distress. In the preschooler cohort proportionally more breadwinner fathers were above the K6 clinical cut‑off than fathers from dual‑earner families.

Page 58: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

46 Social Policy research Paper No. 44

Table 7: Employed mothers’ and fathers’ wellbeing

Infants Preschoolers Adolescents

Whole sample

Mother Father Mother Father Mother Father

n=1,867 n=2,832 n=2,083 n=2,633 n=277 n=270

Mental health(a) MeanSD

3.243.15

3.303.12

3.72*3.31

3.333.20

75.0313.91

75.3614.81

Above threshold % 10.00 9.20 12.60 9.80 8.30 10.20

Socioeconomic position

Low Med/high Low Med/high Low Med/high

Mothers

n=283 n=1,584 n=365 n=1,718 n=63 n=214

Mental health(a) MeanSD

3.373.56

3.223.07

3.973.36

3.673.30

73.4415.41

75.5013.45

Above threshold % 11.40 9.80 14.10 12.40 9.50 7.90

Fathers

n=554 n=2,278 n=495 n=2,138 n=57 n=210

Mental health(a) MeanSD

3.243.66

3.312.97

3.513.84

3.293.02

73.3114.62

75.5814.87

Above threshold % 11.30 8.80 14.20 8.90* 14.00 10.00

Mother family structure

Lonen=74

Couplen=1,793

Lonen=198

Couplen=1,885

Lonen=45

Couplen=232

Mental health(a) MeanSD

4.723.95

3.18**3.10

4.613.94

3.63*3.22

72.6013.82

75.5013.91

Above threshold % 21.60 9.60* 19.60 12.00* 6.70 8.60

Earner arrangement

Breadwinner Dual earner Breadwinner Dual earner Breadwinner Dual earner

Mothers

n=63 n=1,730 n=64 n=1,821 n=8 n=224

Mental health(a) MeanSD

2.652.43

3.202.65

4.253.95

3.613.19

69.5012.45

75.7013.94

Above threshold % 7.90 9.60 17.20 11.80 0.00 8.90

Fathers

n=1,244 n=1,588 n=968 n=1,665 n=54 n=216

Mental health(a) MeanSD

3.373.17

3.243.07

3.503.40

3.233.06

76.7413.14

75.0215.20

Above cut‑off % 10.20 8.60 11.90 8.70* 7.40 11.60

(a) Parent mental health is measured by the K6 in the infant and preschooler cohort (a measure of distress) and by the SF‑36 in the adolescent cohort (a measure of wellbeing).The scales range from 1–7 in adolescent samples and 1–5 in infant and preschooler cohorts.

Notes: Significance tests compared wellbeing across the infant and preschool cohorts only. SD=standard deviation. *p<0.01; **p<0.001.

Page 59: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

47

FAMIlIES ACrOSS ThE COhOrTS: SOCIOdEMOgrAPhIC ChArACTErISTICS ANd WEllbEINg

3.6 Children’s wellbeing

The following section describes patterns in children’s socio‑emotional wellbeing across the three cohorts. First, wellbeing is examined for the sample of children with employed parents and then split by family socioeconomic position, earner arrangement and single parenthood (mothers only). Given that different measures were used to assess children’s wellbeing at each age, we were unable to make direct comparisons between infants, preschoolers and adolescents in terms of their wellbeing. However, using information published by the developers of the scales and normative data where available, we have attempted to describe what the wellbeing scores represent in terms of behaviour and whether these scores are consistent with, or deviate from, population norms.

Table 8 presents average wellbeing scores of children in each cohort, for the mother and father samples, as well as the proportion of preschoolers and adolescents above the thresholds for distress. On average, infants were highly approachable, with average scores of almost 5 out of a maximum score of 6. In addition, infants scored an average of 2.5 on irritability out of a maximum score of 6, suggesting relatively low levels of irritability. Infants with highly approachable temperaments tend to be friendly towards new people and to adjust to new environments quickly (Sanson, Prior, Oberklaid & Northam 1987). Infants with low irritability are usually calm and can be easily distracted from crying by a game or song. In comparison with Australian population norms (Fisher, Rowe & Feekery 2004), infants in our study have an approximately average level of approachability and a slightly higher than average level of irritability.

Preschoolers scored an average of 8 to 9 on the SDQ (out of a maximum score of 38). Higher scores indicate more behavioural and emotional problems such as anxiety, conduct problems, hyperactivity and peer problems, which suggests that preschoolers in this sample have relatively few behavioural and emotional problems. The average scores obtained in our sample are fairly consistent with Australian population norms that indicate that the average score on the SDQ for an Australian child aged 4 to 6 years is 9.05 for boys, and 7.53 for girls (Hawes & Dadds 2004). In addition, only a small proportion of our preschooler sample (around 7 per cent) scored in the clinically significant band of the SDQ, suggesting low rates of mental health disorders. The prevalence of clinically significant behaviour obtained in our sample is consistent with community samples in which approximately 10 per cent of community samples score within the clinical banding.

On average, adolescents scored between 73 and 76 on the SF‑36 mental health subscale on average (see Table 8). This scale assesses feelings of nervousness and depression, with higher scores indicating less frequent nervous or depressed feelings. Population norms for individuals aged 18 to 24 years published in the ABS National Health Survey (ABS 1997) reported 75.2 as the mean score on the mental health subscale for this age group. Therefore, the scores in our adolescent sub‑sample are fairly consistent with population norms. Adopting the clinical cut‑off of 55 (derived by Newnham, Harwood & Page 2007), we found that between 10 and 14 per cent of adolescents score below this cut‑off and are likely to have dysfunctional levels of nervousness and depression.

Page 60: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

48 Social Policy research Paper No. 44

Table 8: Child wellbeing across cohorts

Mother Father

Infants n=1,867 n=2,832

Approachability (+) MeanSD

4.780.80

4.760.81

Irritability (–) MeanSD

2.460.78

2.480.81

Preschoolers n=2,083 n=2,633

SDQ total score (–) MeanSD

8.454.79

8.604.88

SDQ above threshold % 6.80 7.10

Adolescents n=277 n=270

SF‑36 mental health (+) MeanSD

73.6017.01

75.1115.64

SF‑36 above threshold % 13.40 10.40

Notes: (+) High score=positive wellbeing, (–) High score=negative wellbeing. SD=standard deviation.

Table 9 presents infant, preschooler and adolescent wellbeing by family socioeconomic position. There were no significant differences between infants from advantaged (medium/high‑SEP) and disadvantaged (low‑SEP) families on the temperament scales, suggesting that socioeconomic position did not influence their social–emotional functioning. This was also suggested in the adolescent cohort, where there were no significant differences between adolescents from advantaged and disadvantaged families in average scores on the mental health subscale of the SF‑36, nor in the proportions above the clinical threshold on these measures.

In contrast, preschoolers from medium/high‑SEP families demonstrated consistently better social–emotional wellbeing compared with preschoolers from low‑SEP families. Specifically, preschoolers from the more socioeconomically advantaged families had fewer problem behaviours on average, and were less likely to score above the clinical threshold on the SDQ.

Table 9: Child wellbeing: within‑cohort comparisons by socioeconomic position of employed mothers and fathers

Mother Father

Low Med/high Low Med/high

Infants n=283 n=1,584 n=554 n=2,278

Approachability (+) MeanSD

4.790.77

4.780.81

4.690.83

4.780.81

Irritability (–) MeanSD

2.430.77

2.470.78

2.462.46

2.490.80

Preschoolers n=365 n=1,718 n=495 n=2,138

SDQ total score (–) MeanSD

10.025.34

8.11**4.60

10.515.17

8.16**4.70

SDQ above threshold % 12.90 5.50** 12.90 5.80**

Adolescents n=63 n=214 n=57 n=210

SF‑36 mental health (+) MeanSD

71.3618.43

74.2616.56

76.1415.26

74.7815.84

SF‑36 above threshold % 15.90 12.60 8.80 11.00

Notes: Significance tests were not conducted for child wellbeing across cohorts. (+) High score=positive wellbeing, (–) High score=negative wellbeing. SD=standard deviation. **p<0.001.

Page 61: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

49

FAMIlIES ACrOSS ThE COhOrTS: SOCIOdEMOgrAPhIC ChArACTErISTICS ANd WEllbEINg

Both infants and adolescents had similar levels of social and emotional functioning irrespective of whether they were in a lone or couple family (see Table 10). However, preschoolers with employed lone mothers had significantly more problem behaviours on the parent‑rated SDQ, and proportionally more were above the clinical threshold than preschoolers from couple families. This suggests that, overall, preschoolers from employed lone‑mother families have poorer social–emotional functioning than preschoolers from two‑parent families. (Note that earlier findings indicate a relationship between socioeconomic position and child wellbeing that might explain this association.)

Table 10: Child wellbeing by mothers’ family type: within‑cohort comparison

Mothers

Lone Couple

Infants n=74 n=1,793

Approachability (+) MeanSD

4.770.72

4.780.81

Irritability (–) MeanSD

2.630.79

2.450.78

Preschoolers n=198 n=1,885

SDQ total score (–) MeanSD

9.765.30

8.31*4.71

SDQ above threshold % 11.10 6.40*

Adolescents n=45 n=232

SF‑36 mental health (+) MeanSD

72.0016.90

73.9117.05

SF‑36 above threshold % 15.60 12.90

Notes: Significance tests compare lone and two‑parent families within the cohort. (+) High score=positive wellbeing, (–) High score=negative wellbeing. SD=standard deviation. *p<0.01.

Table 11 displays child wellbeing for breadwinner and dual‑earner families across the three cohorts. Infants, preschoolers and adolescents from families with a breadwinner mother showed similar outcomes in social–emotional functioning and learning as their counterparts from dual‑earner families. Note that comparisons involving breadwinner mother families are tentative due to small sample sizes.

However, infants and preschoolers from families with a breadwinner father showed poorer wellbeing than infants and preschoolers from dual‑earner families. Infants from families with a breadwinner father were more irritable, and proportionally more preschoolers from families with a breadwinner father had scores above the clinical cut‑off on the SDQ, indicating problem behaviour.

There were no clear patterns of wellbeing based on family earner arrangements apparent in the adolescent cohort.

Page 62: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

50 Social Policy research Paper No. 44

Table 11: Infants’ and preschoolers’ wellbeing by mothers’ earner arrangement: within‑cohort comparison

Mothers Fathers

Breadwinner Dual earner Breadwinner Dual earner

Infants n=63 n=1,730 n=1,244 n=1,588

Approachability (+) MeanSD

4.750.96

4.780.80

4.720.83

4.790.80

Irritability (–) MeanSD

2.370.75

2.460.78

2.550.85

2.44**0.77

Preschoolers n=64 n=1,821 n=968 n=1,665

SDQ total score (–) MeanSD

8.815.50

8.294.68

9.205.21

8.25**4.64

SDQ above threshold % 12.50 6.20 9.50 5.70**

Adolescents n=8 n=2,241 n=54 n=216

SF‑36 mental health (+) MeanSD

80.508.12

73.6817.25

76.1613.64

74.8416.12

SF‑36 above threshold % 0.00 13.40 7.40 11.10

Notes: Significance tests compare families within each cohort. (+) High score=positive effect, (–) High score=negative effect. SD=standard deviation. **p<0.001.

3.7 Summary

The section assessed whether there were differences in family sociodemographics, parent wellbeing and child wellbeing across the infant, preschooler and adolescent cohorts. We also investigated differences in sociodemographics and wellbeing by family socioeconomic position, family type and earner arrangements. Our analyses indicate the following:

�� Median household income and parent level of education were similar across all three cohorts, as was the proportion of families experiencing financial hardship.

�� The experience of financial hardship in low‑SEP families was nearly double that reported by medium/high‑SEP families. This hardship gap was consistently observed across the infant, preschool and adolescent cohort.

�� Lone employed mothers show a much greater risk of experiencing financial hardship than employed mothers in two‑parent families. This difference was observed in all three child‑age cohorts.

�� Mothers with preschoolers reported poorer wellbeing than mothers of infants. However, fathers’ wellbeing did not vary across cohorts.

�� Proportionally more employed fathers in low‑SEP families reported psychological distress. This wellbeing gap, although slight, was observed in both the infant and preschooler cohorts, with a similar trend observed in the adolescent cohort.

�� Employed lone mothers in the infant and preschooler cohorts showed nearly double the rate of distress compared with employed mothers in two‑parent households. Differences in distress between lone and partnered mothers with adolescents were less marked and not statistically significant.

�� Breadwinner and dual‑earner mothers had similar levels of psychological distress (but note small sample sizes). Proportionally more breadwinner fathers were above the K6 threshold for distress compared with fathers from dual‑earner families in the preschooler cohort.

Page 63: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

51

FAMIlIES ACrOSS ThE COhOrTS: SOCIOdEMOgrAPhIC ChArACTErISTICS ANd WEllbEINg

�� The socio‑emotional functioning of infants, preschoolers and adolescents in this study was comparable to the norms and prevalence rates reported in other studies.

�� Preschool children showed the clearest variation in mental health‑related wellbeing according to the family socioeconomic position. Preschoolers from disadvantaged families showed more social or emotional problem behaviours than preschool children from more advantaged families.

�� Preschool‑aged children of employed lone mothers showed nearly double the rate of social and emotional problems compared with children of employed mothers in two‑parent families.

�� We found little evidence that infants’ or preschoolers’ wellbeing differed according to whether their mothers were dual earners or breadwinners. However children’s wellbeing tended to be poorer in families where fathers were breadwinners rather than dual earners in both of these younger age cohorts.

Page 64: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

52 Social Policy research Paper No. 44

Page 65: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

53

MOThErS’ ANd FAThErS’ ExPErIENCE OF WOrk

4 Mothers’ and fathers’ experience of work

4.1 Overview and aims

The focus of this section is mothers’ and fathers’ work arrangements and conditions. Despite the unprecedented movement of women into the workforce and the efforts of policy to improve gender equity, there is evidence of clear differences between mothers’ and fathers’ experience of work. While these differences are well documented, few studies have had the opportunity to compare the work experiences of mothers and fathers in families with infants, preschoolers and adolescents.

There has been considerable interest in parents’ work hours, but there is much less evidence on the quality and conditions of the jobs mothers and fathers can access. The first set of analyses in this section address this issue. Do mothers and fathers work in jobs with similar conditions? Are mothers’ and fathers’ jobs equally flexible? Do mothers’ job conditions get better as their children become older and mothers work longer hours?

A second set of analyses investigates how two‑parent families share earning income. How different are mothers’ and fathers’ earning arrangements and does this gender‑based distinction persist across all cohorts? Does the experience of being a dual‑earner parent or breadwinner differ between mothers and fathers? Do fathers change their work hours when their children get older and, if so, by how much?

Thus this section has two aims:

�� First, to describe and compare the work experiences of mothers and fathers, in terms of hours, conditions and work arrangements. These comparisons look within cohorts to give a snapshot of employed mothers and fathers with infants, preschoolers or adolescents.

�� Second, to describe similarities and differences across cohorts in earner arrangements, hours and conditions for mothers and fathers. Any patterns we discover are suggestive of how child age may modulate parents’ work, although this needs to be confirmed in additional studies using longitudinal data.

Evidence from this section could inform several current policy areas. First, the section is relevant for social inclusion agendas focused on gender equity and workforce participation. It provides descriptive evidence on whether Australian mothers work in poorer quality jobs than fathers and if this situation is relatively transient (observed only in the infant cohort) or persistent (observed across all cohorts). It also provides evidence on mothers’ capacity to participate and succeed in the labour force by highlighting the barriers they face linked to conflicting time demands. In particular, it flags possible limits in the degree to which mothers’ participation in full‑time work can be increased, given the reality of their child care responsibilities. Mothers could remain unable to respond to policy encouragement to increase participation, or move from part to full‑time work, unless child care, industrial relations and workplace policy address these time demands.

Please note that in this section we analyse mothers’ and fathers’ data separately and so do not test for statistically significant differences between the samples of employed mothers and employed fathers. Any commentary on gender differences is therefore based on the patterns and trends and awaits future research to definitively test for their statistical significance. Further, the small sample size of the adolescent cohort means that we cannot answer many of the research questions for parents in this group. For some groups, such as breadwinner mothers, we could not conduct multivariate analyses comparing dual‑earner and breadwinner parents because of small samples.

Page 66: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

54 Social Policy research Paper No. 44

4.2 do mothers and fathers work in jobs with similar conditions?

Work hours and conditions: employed mothers and fathers

The first set of analyses compares work hours and conditions of employed mothers and fathers from the three cohorts without consideration of different family types or earner arrangements. Table 12 describes work hours and conditions for mothers and fathers with infants, preschoolers and adolescents.

Mothers and fathers showed starkly different working patterns, both in terms of hours worked per week and in the proportions working part time or full time. Employed fathers of infants and preschoolers devoted, on average, about twice the number of hours to paid work as mothers. Smaller differences in work hours were observed in the adolescent cohort, where mothers worked an average of two‑thirds of the hours fathers worked. Close to three‑quarters of employed mothers worked part time in the infant and preschooler cohort and just over half worked part time in adolescent cohort. Few fathers worked part time in any cohort and similarly small percentages of mothers worked long full‑time hours. This trend for mothers remained fairly stable across the three cohorts, with a slightly greater proportion in the adolescent sample.

More fathers than mothers reported job security, job control and flexible work hours in all cohorts. This suggests that while part‑time work may give mothers more time, it does not deliver them more control over their work time, or greater job security.

Along with work hours, the nature of the employment tenure showed clear‑cut differences between mothers and fathers. Mothers were three to four times more likely than fathers to work in a casual job in all cohorts (where there is no formal obligation for continuing work on the part of employers). Fathers were more likely to work in a permanent position in all cohorts, although this difference narrowed in the adolescent cohort. Interestingly, while 20 to 25 per cent of both mothers and fathers were self‑employed in the infant and preschooler cohorts, 10 per cent of mothers and 25 per cent of fathers were self‑employed in the adolescent cohort. This suggests that self‑employment is a relatively short‑term strategy used by mothers with younger children for managing work and family demands. On the other hand, fathers’ self‑employment does not appear to vary according to the age of their children, suggesting that it is less closely linked to child care or family demands.

Page 67: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

55

MOThErS’ ANd FAThErS’ ExPErIENCE OF WOrk

Table 12: Employed mothers’ and fathers’ work hours and conditions by cohort, with comparisons across cohorts

Infants Preschoolers Adolescents

Mothersn=1,867

Fathersn=2,832

Mothersn=2,083

Fathersn=2,633

Mothersn=277

Fathersn=270

Work hours(per week)

Mean 22.66+ 46.18+ 23.92 47.26 31.21# 47.90

Part time(≤35 hours)

% 73.55 6.45 75.00 5.80 53.40# 5.20

Full time(36–45 hours)

% 21.45 50.55 19.20 48.30 37.20# 45.50

Long full time(≥46 hours)

% 5.00 43.00 5.80 45.90 9.40 49.30

Job security % 76.20 81.50 76.60 80.40 87.00# 91.50#

Employment tenure^

Permanent % 58.90 70.70 52.40 68.80 58.50 64.80

Fixed term % 3.00 3.40 5.40 2.50 10.10 5.20

Casual % 17.60 4.50 19.20 3.60 21.30 4.80

Self‑employed % 20.50 21.40 23.00 25.10 10.10 25.20

Job control % 80.20 84.40 82.00 85.00 85.20 86.70

Flexible hours(a) % 80.00 82.30 82.20 82.20 51.30# 55.20#

(a) Different items assessed flexible hours in the adolescent sample than in the infant and preschooler samples.Notes: + Difference between infants and preschoolers at p<0.01.

# Difference between adolescent cohort and both infant and preschooler cohort at p<0.01.^ Across‑cohort comparison employment tenure:

�� mothers’: infant/preschoolers 2(3)=25.40, p<0.01; infants/adolescents 2(3)=46.42, p<0.01;preschoolers/adolescents 2(3)=30.28, p<0.01

�� fathers’: infant/preschoolers 2(3)=15.96, p<0.01; infants/adolescents not significant;preschoolers/adolescents not significant.

Are work conditions related?

We examined Pearson’s product moment correlations of the work conditions of mothers and fathers across the three cohorts. Examining the correlations matrix (see Table 13) can give insight into whether positive job conditions tend to ‘hang together’. The table shows that the correlations that were significant tended to be only weak to moderate associations. There were several correlation patterns that are worth noting.

Casual employment was weakly and negatively associated with control and flexibility for fathers in the infant and preschooler cohort, and for mothers in the preschool cohort. Casual employment was also negatively associated with low job security, for both mothers and fathers (excepting mothers in the preschool cohort). Conversely, self‑employment is associated with both job control and flexibility for mothers and fathers (with the exception of mothers in the adolescent cohort). Self‑employment may deliver many of the job conditions that are important when combining paid work with care.

Unsurprisingly, work hours were negatively associated with casual employment for mothers and fathers across all three cohorts and the associations were stronger for mothers than fathers, particularly in the adolescent cohort. Work hours were also negatively associated with mothers’ self‑employment in the younger cohorts, but positively associated with fathers’ self‑employment in all three cohorts. Self‑employed fathers tend to work longer hours while mothers appear to use it more as a part‑time work option. There appears to be no consistent relationship between work hours and job control and flexibility. Job security is weakly associated with mothers’ work hours in the infant and adolescent cohort.

Page 68: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

56 Social Policy research Paper No. 44

Flexibility and job control are moderately positively associated for both mothers and fathers across all three cohorts. Job control is also significantly, though weakly, associated with security for mothers and fathers in the younger cohorts and moderately correlated for fathers in the adolescent cohort. There is no clear pattern of association for flexibility and security; they are related for fathers but not mothers in the infant and adolescent cohorts and mothers but not fathers in the preschooler cohort. Though there are significant associations for these job quality conditions, the associations are small.

The positive work conditions described in this report do co‑occur in some jobs, but not for everyone. Many parents are likely to be working in a job that delivers some positive conditions but not others; for example, a mother working in a job that is flexible but does not allow much control.

do parents’ job conditions improve as their children grow up?

Table 12 also displays the results of significance tests that assess the differences in means and proportions of work hours and conditions between the three cohorts. While fathers’ work hours did not significantly differ between the cohorts, mothers’ work hours varied between cohorts in line with the age of the study child. This was true for both the average number of work hours per week and for the proportion of parents working part time and full time. Mothers worked 7 to 8 hours more per week on average in the adolescent cohort than in the infant and preschooler cohorts. Mothers were also more likely to work full time in the adolescent cohort than in the infant and preschooler cohorts, with 37 per cent of mothers in the adolescent cohort working full time compared to 21 per cent in the infant cohort and 19 per cent in the preschooler cohort.

Both employed mothers and employed fathers from the adolescent cohort reported higher levels of job security and job control than mothers and fathers from the infant and preschooler cohorts, although the differences for job control were non‑significant. This suggests that parents may be able to access better quality jobs as their children get older. In contrast, fewer parents reported flexible hours in the adolescent cohort than in the infant and preschooler cohorts; however, this is likely to reflect differences in question wording (see Section 2.3 ‘Caveats’).

do earner arrangements vary in families across the age cohorts?

This section, a descriptive analysis, considers whether families vary in earner arrangements (breadwinner and dual earner) across the infant, preschooler and adolescent cohorts. Typically, fathers are breadwinners when children are young, but families often change to dual earner when children are older and mothers return to work. As there is a small group of breadwinner mothers in our analysis, we consider whether there is any evidence that their partners might also return to work in the older cohort. Note that these analyses only include two‑parent families and that, although the cohorts supply evidence of differences that might be linked to child age, they do not show whether individual parents changed their work patterns over time. Such analysis would require longitudinal data.

The proportion of two‑parent families with a mother as a breadwinner was consistently low at around 3 per cent in all cohorts. In contrast, the proportion of two‑parent families with a father as breadwinner was smaller in each successive older cohort: about 40 per cent of employed fathers were breadwinners in the infant (44 per cent) and preschooler (37 per cent) cohorts, while 20 per cent were breadwinners in the adolescent cohort.

Conversely, the proportion of fathers in dual‑earner families increased with child age, with 80 per cent of employed fathers in dual‑earner arrangements in the adolescent cohort compared with 63 per cent in the preschooler and 56 per cent in the infant cohort.

The division of work hours between mothers and fathers in dual‑earner families also differed depending on child age. In all three cohorts, the most common arrangement was for fathers to be working full time and mothers part time. This arrangement (often termed a 1.5 earner household) became less common as the children got older. Seventy per cent of dual‑earner families with infants and 73 per cent of dual‑earner families with preschoolers had this arrangement, as compared to only 55 per cent of families with adolescents.

Page 69: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

57

MOThErS’ ANd FAThErS’ ExPErIENCE OF WOrk

Tabl

e 13

: In

terc

orre

latio

ns o

f job

con

ditio

ns fo

r mot

hers

and

fath

ers

by c

ohor

t

Job

secu

rity

Casu

al te

nure

Self

‑em

ploy

edJo

b co

ntro

lFl

exib

ility

(yes

/no)

Mot

her

Fath

erM

othe

rFa

ther

Mot

her

Fath

erM

othe

rFa

ther

Mot

her

Fath

er

Infa

nt c

ohor

t

1. W

ork

hour

s (w

eekl

y av

erag

e)0.

06**

0.03

–0.3

0**

–0.3

0**

–0.2

3**

0.16

**–0

.03

0.04

*–0

.05*

0.01

2. Jo

b se

curi

ty (l

ow/h

igh)

–0.1

0**

–0.1

1**

–0.0

1–0

.03

0.07

**0.

10**

0.04

0.07

**

3. C

asua

l ten

ure

(no/

yes)

–0.2

4**

–0.1

1**

–0.0

3–0

.08*

*–0

.04

–0.0

8**

4. S

elf‑

empl

oyed

(no/

yes)

0.10

**0.

16**

0.15

**0.

11**

5. Jo

b co

ntro

l (lo

w/h

igh)

0.23

**0.

23**

Pres

choo

ler c

ohor

t

1. W

ork

hour

s (w

eekl

y av

erag

e)0.

000.

05*

–0.2

9**

–0.2

2**

–0.1

2**

0.19

**0.

010.

04*

0.00

–0.0

1

2. Jo

b se

curi

ty (l

ow/h

igh)

–0.0

4–0

.10*

*–0

.01

0.00

0.06

**0.

08**

0.06

**0.

01

3. C

asua

l ten

ure

(no/

yes)

–0.2

7**

–0.1

1**

–0.0

3–0

.09*

*–0

.05*

–0.0

6**

4. S

elf‑

empl

oyed

(no/

yes)

0.12

**0.

12**

0.16

**0.

08**

5. Jo

b co

ntro

l (lo

w/h

igh)

0.22

**0.

24**

Ado

lesc

ent c

ohor

t

1. W

ork

hour

s (w

eekl

y av

erag

e)0.

13*

0.06

–0.4

3**

–0.2

1**

0.04

0.25

**0.

100.

05–0

.02

–0.0

6

2. Jo

b se

curi

ty (l

ow/h

igh)

–0.2

5**

–0.1

8**

0.09

–0.0

10.

050.

19**

–0.0

60.

13*

3. C

asua

l ten

ure

(no/

yes)

–0.1

7**

–0.1

3–0

.06

–0.1

7**

0.07

–0.1

1

4. S

elf‑

empl

oyed

(no/

yes)

0.11

0.15

*0.

28**

0.13

*

5. Jo

b co

ntro

l (lo

w/h

igh)

0.24

**0.

30**

Not

e:

*p<

0.05

; **p

<0.

001.

Page 70: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

58 Social Policy research Paper No. 44

Mothers in dual‑earner families also appeared to move from part‑time to full‑time work as their children got older. Almost twice as many families in the adolescent cohort had both parents working full time (41 per cent) as in the infant (24 per cent) and preschooler cohorts (21 per cent). In all three cohorts, families where both parents worked part time or with mothers working full time and fathers working part time were relatively rare (6 per cent for infant cohort, 6 per cent for the preschooler cohort and 4 per cent for the adolescent cohort).

4.3 breadwinner and dual‑earner families across the three cohorts

Are families with breadwinner mothers similar to families with breadwinner fathers in size and other socioeconomic characteristics? How do breadwinner families differ from dual‑earner families across the three cohorts? The next series of analyses compares two‑parent family characteristics according to different earner arrangements. Because of the very small numbers of breadwinner mothers with adolescents, all inferences involving this group of mothers are tentative.

As illustrated in Table 14, family size was similar for dual‑earner families and families with a breadwinner mother in each of the three cohorts. In contrast, families with a breadwinner father were larger than dual‑earner families. Children also tended to be younger in families with a breadwinner father, when compared to dual‑earner families. Unsurprisingly, the younger cohorts were much more likely to have infants in the family. Working fathers, both breadwinner and dual earners, were more likely to have an infant in the family than working mothers. This suggests that families that have a father‑as‑breadwinner arrangement may be at a slightly earlier family life‑cycle stage than dual‑earner families or families with a breadwinner mother.

Child care use also varied by earner arrangements. In the infant and preschooler cohorts, dual‑earner families were more likely to use child care than breadwinner families. In particular, over 50 per cent of dual‑earner families used child care in the infant and preschooler cohorts and these families were almost three times more likely to use child care than families with a father as breadwinner.

Consistent across the three cohorts, families with a breadwinner mother were relatively financially disadvantaged compared with two‑parent, dual‑earner families (see Table 15). This hardship gap is particularly marked when mothers are the breadwinners and it appears to widen as the children grow up. One in four breadwinner mothers with an infant report financial hardship, compared with one in two in the adolescent cohort. Note, however, that the very small sample size for breadwinner mothers means that findings for this cohort must be tentative.

Table 15 shows that families with breadwinner fathers had lower median household income compared with dual‑earner families in the preschooler and adolescent cohorts, but not in the infant cohort. In the infant and preschooler cohorts, families with breadwinner fathers were also more likely to report experiencing financial hardship in the past year. Note that we cannot statistically test whether the income gap between breadwinner and dual‑earner families widens across child developmental stage because income data in LSAC is reported in ranges. However, similarities in the income of dual‑earner families and those with a breadwinner father may be because some employed dual‑earner mothers are on unpaid leave.

There was very little difference in education attainment of mothers and fathers across the three cohorts. Fathers in dual‑earner families had higher average numbers of years of education than fathers in breadwinner families, but this difference was small.

Page 71: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

59

MOThErS’ ANd FAThErS’ ExPErIENCE OF WOrk

Tabl

e 14

: Fa

mily

cha

ract

eris

tics

by e

arne

r arr

ange

men

ts: w

ithin

‑coh

ort c

ompa

rison

Mot

hers

Fath

ers

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Bre

ad‑

win

ner

Dua

l ea

rner

Bre

ad‑

win

ner

Dua

l ea

rner

Bre

ad‑

win

ner(a

)D

ual

earn

erB

read

‑w

inne

rD

ual

earn

erB

read

‑w

inne

rD

ual

earn

erB

read

‑w

inne

rD

ual

earn

ern=

63n=

1,73

0n=

64n=

1,82

1n=

8n=

224

n=1,

244

n=1,

588

n=96

8n=

1,66

5n=

54n=

216

Youn

gest

chi

ld

(yea

rs/m

onth

s)M

ean

9 m

9 m

3 y,

3 m

3 y,

4 m

13 y

12 y

, 4 m

8 m

9 m

*2

y, 9

m3

y, 4

m*

10 y

, 4 m

12 y

, 1 m

*

Infa

nt in

fam

ily%

92.1

085

.70

21.9

017

.70

0.00

0.90

94.4

097

.20

29.6

018

.40*

7.40

1.40

No.

of c

hild

ren

Mea

n1.

901.

772.

392.

324.

624.

822.

051.

782.

672.

32*

5.66

4.88

1%

38.1

045

.80

15.6

09.

800.

003.

1033

.30

45.2

0*5.

009.

70*

3.70

2.80

2%

38.1

036

.80

45.3

057

.20

25.0

014

.30

39.3

036

.80*

47.7

057

.60*

7.40

11.6

0

3–4

%23

.80

16.2

035

.90

31.7

050

.00

39.3

025

.40

16.6

0*41

.80

31.4

0*35

.20

41.2

0

5+%

0.00

1.20

3.10

1.30

25.0

043

.30

2.00

1.40

*5.

501.

30*

53.7

044

.40

Use

s ch

ild

care

(b)

%28

.60

57.1

0*37

.50

53.7

0*12

.50

17.0

016

.90

54.7

0*20

.20

52.4

0*3.

7018

.50*

(a)

Not

e th

e ve

ry s

mal

l sam

ple

size

for b

read

win

ner m

othe

rs in

the

adol

esce

nt c

ohor

t. A

ll in

fere

nces

are

tent

ativ

e.(b

) In

the

adol

esce

nt s

ampl

e, o

nly

fam

ilies

wit

h ch

ildre

n un

der 1

4 ye

ars

wer

e as

ked

abou

t chi

ld c

are

use.

We

repo

rt th

e pe

rcen

tage

of t

he e

ntir

e ad

oles

cent

sam

ple

usin

g ch

ild c

are.

Not

es:

Sign

ifica

nce

test

s co

mpa

re b

read

win

ner a

nd d

ual‑e

arne

r fam

ilies

wit

hin

each

coh

ort.

Due

to ro

undi

ng, p

erce

ntag

es m

ay n

ot a

dd to

100

per

cen

t. m

refe

rs to

mon

ths,

y to

yea

rs o

f ag

e. *

p<0.

01.

Page 72: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

60 Social Policy research Paper No. 44

Tabl

e 15

: So

cioe

cono

mic

cha

ract

eris

tics

by m

othe

rs’ a

nd fa

ther

s’ e

arne

r arr

ange

men

ts: w

ithin

‑coh

ort c

ompa

rison

Mot

hers

Fath

ers

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Infa

nts

Pres

choo

lers

Ado

lesc

ents

Bre

ad‑

win

ner

Dua

l ea

rner

Bre

ad‑

win

ner

Dua

l ea

rner

Bre

ad‑

win

ner(a

)D

ual

earn

erB

read

‑w

inne

rD

ual

earn

erB

read

‑w

inne

rD

ual

earn

erB

read

‑w

inne

rD

ual

earn

ern=

63n=

1,73

0n=

64n=

1,82

1n=

8n=

224

n=1,

244

n=1,

588

n=96

8n=

1,66

5n=

54n=

216

Hou

seho

ld

inco

me

(p.a

.)

Med

ian

$36,

400–

$41,

599

$52,

000–

$77,

999

$36,

400–

$41,

599

$78,

000–

$103

,999

$59,

144

($45

,000

)(b)

$99,

697

($82

,750

)(b)

$52,

000–

$77,

999

$52,

000–

$77,

999

$52,

000–

$77,

999

$78,

000–

$103

,999

$80,

772

($55

,000

)(b)

$97,

974

($80

,756

)(b)

Fina

ncia

l ha

rdsh

ip%

28.6

021

.00*

43.8

019

.70*

*50

.00

18.3

026

.90

20.4

0*25

.90

18.7

0*24

.10

20.4

0

Educ

atio

nM

ean

15.9

215

.59

15.0

915

.18

13.8

714

.09

14.8

415

.08

14.8

915

.03

14.3

714

.85

Inco

mpl

ete

seco

ndar

y%

6.30

6.50

15.6

012

.80

50.0

032

.60

11.7

08.

8014

.40

11.4

027

.80

18.5

0

Com

plet

e se

cond

ary

%6.

3013

.60

6.30

13.4

00.

0011

.60

10.5

09.

907.

709.

801.

905.

10

Cert

ifica

te/

Dip

lom

a%

34.9

034

.00

39.1

034

.40

0.00

25.9

048

.50

46.9

046

.30

45.2

040

.70

42.6

0

Bac

helo

r or

high

er%

52.4

045

.80

39.1

039

.40

50.0

029

.90

30.0

033

.80

31.6

033

.70

29.6

033

.80

(a)

Not

e th

e ve

ry s

mal

l sam

ple

size

for b

read

win

ner m

othe

rs in

the

adol

esce

nt c

ohor

t. A

ll in

fere

nces

are

tent

ativ

e.(b

) H

ouse

hold

inco

me

in th

e ad

oles

cent

sam

ple

incl

uded

inco

me

from

all

fam

ily m

embe

rs. C

ombi

ned

pare

nt a

nnua

l sal

ary

(but

not

incl

udin

g be

nefit

s or

oth

er in

com

e) is

not

ed in

pa

rent

hese

s.N

otes

: Si

gnifi

canc

e te

sts

com

pare

bre

adw

inne

r and

dua

l‑ear

ner f

amili

es w

ithi

n ea

ch c

ohor

t. D

ue to

roun

ding

, per

cent

ages

may

not

add

to 1

00 p

er c

ent.

m re

fers

to m

onth

s, y

to y

ears

of

age.

*p<

0.01

; **p

<0.

001.

Page 73: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

61

MOThErS’ ANd FAThErS’ ExPErIENCE OF WOrk

4.4 breadwinner and dual‑earner work conditions

Are breadwinner mothers’ jobs equally flexible, autonomous and secure as dual‑earner mothers’ jobs, or as fathers’? Are mothers who are dual earners working in jobs with poorer conditions than dual‑earner fathers? Is there any evidence that mothers’ or fathers’ job conditions improve when their children become older? In this section we compare mothers’ and fathers’ jobs in terms of security (both perceived and by type of employment tenure), level of autonomy or ‘say’ at work, and their flexibility of start and stop times.

Work hours and conditions: breadwinner and dual‑earner families

Table 16 displays mothers’ and fathers’ average work hours per week and the proportion of mothers reporting each working condition by earner arrangement. The table also presents the results of within and across‑cohort significance tests of working hours and conditions.

As shown in Table 16, breadwinner mothers in the infant and preschooler cohorts worked significantly more hours per week than mothers from dual‑earner families. These mothers were also more likely to work full time or long hours in the infant cohort. No differences in work hours were observed between breadwinner and dual‑earner mothers in the adolescent cohort (but note that small samples sizes mean inferences are tentative for this group). In contrast, fathers’ work hours were similar for both breadwinners and dual earners across all three cohorts.

For job security in the adolescent cohort, mothers from dual‑earner families were nearly 40 per cent more likely to report having a secure future in their jobs than breadwinner mothers. This contrasted with mothers in the infant and preschooler cohorts who reported similar levels of job security regardless of whether they were breadwinner or dual earner. For employment tenure, breadwinner mothers in the infant cohort were more likely than dual‑earner mothers to be on permanent or fixed‑term tenures and were less likely to be casually or self‑employed.

Cross‑cohort comparisons reveal that dual‑earner mothers in the adolescent cohort work, on average, more hours than dual‑earner mothers in the infant and preschooler cohorts. Dual‑earner mothers in the adolescent cohort were more likely to work long hours and less likely to work part time than their counterparts in the younger cohorts. Dual‑earner mothers in the adolescent cohort were significantly more likely to report feeling secure in their jobs and less likely to work flexible hours compared to dual‑earner mothers in both young cohorts. (Note that breadwinner mothers could only be compared across the infant and preschooler cohort due to the very small number of breadwinner mothers in the adolescent cohort.)

Page 74: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

62 Social Policy research Paper No. 44

Table 16: Mothers’ and fathers’ work hours and conditions by earner arrangements

Infants Preschoolers Adolescents

Bread‑winnern=63

Dual earner

n=1,730

Bread‑winnern=64

Dual earner

n=1,821

Bread‑winner

n=8

Dual earnern=224

Mothers

Work hours (per week) Mean 29.59 22.49* 29.37 23.59* 32.00 30.97#

Part time (≤35 hours) % 54.00 74.00* 60.90 75.90 37.50 55.40#

Full time (36–45 hours) % 34.90 21.30* 28.10 18.50 50.00 35.30#

Long full time (≥46 hours) % 11.10 4.70* 10.90 5.60 12.50 9.40@

Job security % 77.80 76.50 79.70 77.60 50.00 89.70*#

Employment tenure^

Permanent % 69.80 58.40* 56.30 51.90 75.00 58.00

Fixed term % 7.90 2.70* 6.20 5.50 0.00 10.30

Casual % 15.90 17.30* 23.40 17.80 25.00 20.10

Self‑employed % 6.40 21.60* 14.10 24.80 0.00 11.60

Job control % 85.70 80.20 82.80 82.50 87.50 86.60

Flexible hours % 82.50 80.10 84.40 82.10 62.50 50.40#

Fathers

Work hours (per week) Mean 45.75+ 46.52 47.43 47.16 47.81 47.92

Part time (≤35 hours) % 6.40 6.40 4.90 6.30 7.40 4.60

Full time (36–45 hours) % 52.30 49.20 47.60 48.70 40.70 46.80

Long full time (≥46 hours) % 41.30+ 44.40 47.50 45.00 51.90 48.60

Job security % 80.90 82.10 80.00 80.70 88.90 92.10#

Employment tenure^

Permanent % 74.40 67.80* 76.00 64.60* 72.20 63.00

Fixed‑term % 4.60 2.50 2.80 2.30 0.00 6.50

Casual % 5.40 3.80 4.10 3.20 3.70 5.00

Self‑employed % 15.60 25.90 17.10 29.90 24.10 25.50

Job control % 82.60 85.90* 83.30 86.10 85.20 87.00

Flexible hours % 80.40 86.90* 79.60 83.70* 59.30# 54.20#

Notes: Breadwinner mothers could only be compared across the infant and preschooler cohorts. Due to rounding, percentages may not add to 100 per cent. *p<0.01 within‑cohort comparisons; dual earner compared with breadwinner within each cohort.

+p<0.01 cross‑cohort comparison; difference between infants and preschoolers.#p<0.01 cross‑cohort comparison; adolescents significantly different to both infants and preschoolers@p<0.01 cross‑cohort comparison; significant difference between infants and adolescents^ Cross‑cohort comparison:

�� dual‑earner mothers employment tenure: infant/preschoolers 2(3)=26.93, p<0.01;infants/adolescents 2(3)=41.51, p<0.01; preschoolers/adolescents 2(3)=25.75, p<0.01

�� breadwinner fathers employment tenure: infant/preschoolers not significant;infants/adolescents 2(3)=31.11, p<0.01; preschoolers/adolescents 2(3)=11.49, p<0.01

�� dual‑earner fathers employment tenure: infant/preschoolers not significant;infants/adolescents 2(3)=11.49, p<0.01; preschoolers/adolescents 2(3)=15.42, p<0.01.

Page 75: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

63

MOThErS’ ANd FAThErS’ ExPErIENCE OF WOrk

There were variations in work conditions and employment tenure for breadwinner and dual‑earner fathers (see Table 16). In all cohorts, breadwinner fathers were more likely to be permanently employed than dual‑earner fathers. Breadwinner fathers were less likely to report flexible hours than fathers from dual‑earner families in the infant and preschooler cohorts and were less likely to report job control in the infant cohort. Similar patterns were observed for fathers in the adolescent cohort, although these were not statistically significant.

Cross‑cohort comparisons show that dual‑earner fathers in the adolescent cohort are more likely to report feeling secure in their job than dual‑earner fathers in both younger cohorts. Both dual‑earner and breadwinner fathers in the adolescent cohort were less likely to have access to flexible work hours than fathers in both younger cohorts.

Breadwinner fathers in the infant cohort report slightly lower average work hours than breadwinner fathers in the preschool cohort and are also less likely to work long hours.

4.5 gender equality in the workforce: family time as a resource

Taken together, descriptive findings thus far indicate that, compared with fathers, mothers in the labour market do not have the same access to a key resource—time—and this is likely to shape employment opportunities and success. One way to think about this is in terms of how work time is shared in two‑parent families (both mothers and fathers).

The vast majority of mothers who are employed also have partners who are working. In most instances these fathers work full time and nearly half work long full‑time hours. On the other hand, fathers’ jobs are usually worked in a much more ‘time‑rich’ family context—it is more likely that their partner is at home or their partner is working part time. The notable exception is the one‑fifth to two‑fifths of dual‑earner fathers whose partner works full time.

Figure 6 depicts average household work time for dual‑earner and breadwinner families across the three cohorts. There is nearly a 25‑hour difference in household work time between breadwinner and dual‑earner families with infants and preschoolers and this widens to approximately 32 hours in the adolescent cohort. Families where both parents work in paid employment face significantly greater time pressure for balancing work with the care needs of children.

Page 76: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

64 Social Policy research Paper No. 44

Figure 6: Household work hours per week by earner arrangements per cohort

Dual earners Breadwinners

Cohort

Mea

n ho

urs

0

10

20

30

40

50

60

70

80

AdolescentPreschoolerInfant

Figure 7 shows mothers’ and fathers’ work hours by earner arrangement across the three cohorts. The figure shows that mothers, but not fathers, appear to tailor their work time according to the needs and commitments of family members. Mothers work more hours when they are solely responsible for earning income, particularly when their children are young, and work fewer hours in dual‑earner arrangements. Further, mothers appear to tailor their work time according to the age (and presumably the care needs) of their children, increasing their hours as their children age. Mothers’ paid work time increases from an average of 22 hours per week for the infant cohort to over 30 hours per week in the adolescent cohort (slightly more for breadwinner compared to dual‑earner mothers).

Fathers’ work time, however, appears much more fixed. Fathers do not appear to change their work hours when their children get older. Fathers in both dual‑earner and breadwinner arrangements work an average of 45 to 47 hours a week across the infant, preschooler and adolescent cohorts. This suggests that their partner’s work status does not influence how many hours fathers devote to paid work.

Page 77: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

65

MOThErS’ ANd FAThErS’ ExPErIENCE OF WOrk

Figure 7: Mothers’ and fathers’ work hours per week by earner arrangements per cohort

Dual earners Breadwinners

Cohort and sex of parentFather Mother

Mea

n ho

urs

0

10

20

30

40

50

AdolescentPreschoolerInfant AdolescentPreschoolerInfant

4.6 Summary

This section explored differences in mothers’ and fathers’ work arrangements and work conditions. Conditions were examined in dual‑earner and breadwinner families for the infant, preschooler and adolescent cohorts. There is a large body of work that has investigated parents’ work hours, but until now there has been much less evidence on the quality and conditions of the jobs in which mothers and fathers work.

The descriptive analysis revealed that:

�� Employed mothers work fewer hours and appeared to have less job security and control than fathers.

�� Mothers were more likely to be casually employed, while fathers were more likely to be permanently employed.

�� Both mothers and fathers reported more job security in the adolescent cohort than parents from the younger cohorts. However, there was no evidence that mothers moved from casual to permanent jobs in the older cohort, or that gender gaps in job security or job control were lower once children were older.

�� Being the breadwinner is common for fathers when children are young, but many two‑parent families appear to shift to dual‑earner arrangements when they have adolescents in the family. This suggests father‑as‑breadwinner arrangements are often relatively short term, linked to children’s development and care needs.

�� The proportion of mothers with sole responsibility for earning income appears relatively fixed and independent of the age of their children.

�� Part‑time work remains the main option for employed mothers in our sample. Even in the adolescent cohort, part‑time work was still the most common work pattern, although the proportions were lower.

�� Part‑time work does not appear to be a viable option for fathers. Irrespective of their child’s developmental stage or partners’ work hours, relatively few fathers worked part time.

Page 78: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

66 Social Policy research Paper No. 44

�� Twice as many dual‑earner families in the adolescent cohort had both parents working full time (41.2 per cent) compared with the infant cohort (24.2 per cent) and the preschooler cohort (20.9 per cent).

�� Compared with mothers, fathers’ jobs are often worked in a more ‘time‑rich’ context; either their partner is at home, or their partner is working part time.

�� Breadwinner mothers work more hours per week compared to dual‑earner mothers and both groups of mothers appear to work fewer hours when children are young.

�� Fathers do not appear to calibrate their work time according to either their partners’ work status, or their child’s age, whereas mothers do. Mothers’ and fathers’ time resources are therefore asymmetrical, and this could limit gender equity in earnings and career advancement.

�� Proportionally more breadwinner families report financial hardship than families where both parents are employed. This hardship gap is particularly marked when mothers are the breadwinners and it appears to widen as the children grow up. One in four breadwinner mothers with an infant report financial hardship, compared with one in two in the adolescent cohort (note, however, the small sample for the latter).

�� Both mothers and fathers who were breadwinners tended to have more secure job tenures compared with their dual‑earner counterparts.

�� Compared with dual‑earner fathers, breadwinner fathers in the younger cohorts had less flexible work hours.

Page 79: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

67

WOrk ANd WEllbEINg

5 Work and wellbeing

5.1 Overview and aims

This next section considers how employed mothers’ and fathers’ work hours and conditions are associated with their own wellbeing and that of their children. Until now this report has focused largely on mothers’ and fathers’ work hours. Indeed most research on work and family has focused solely on time‑related aspects of parents’ jobs, such as part‑time jobs, flexible hours, paid maternity or paternity leave, and time off to care for sick children. These time‑related conditions could be important to wellbeing because they remove some of the time barriers faced by employed mothers and fathers and so reduce pressures on family life and on parents. But not all jobs are equal and there is now several decades of epidemiological evidence linking other working conditions such as job insecurity and low control to adult wellbeing. The following section considers both mothers’ and fathers’ hours and work conditions across the three cohorts.

Optimising parents’ wellbeing may yield benefits to business, communities and reduce burdens on the nation’s health care system. Health and wellbeing is a key component of the nation’s human capital and health problems lead to staff turnover, sickness absence and reduced productivity. Poor workforce health therefore affects the bottom line for business and leads to increased health care costs. Mental health problems are particularly costly. In terms of productivity and absences, depression has as great or greater impact than nearly all other common illnesses. The estimated costs of depression to US business are in the region of US$33 billion annually (Druss, Rosenheck & Sledge 2000; Greenberg et al. 1996). Depression and anxiety are leading causes of disability and disease burden and globally, unipolar depressive disorder is projected to become the second most important cause of years of productive life lost to disability (Mathers & Loncar 2006).

Over and above its national and economic importance, parent wellbeing is an essential resource for families and children because it underpins the quality of family relationships. For example, depressed parents are more withdrawn, angry and sad, and this affects the quality of marital and parent–child relationships (Coyne, Thompson & Palmer 2002; Downey & Coyne 1990; Lovejoy et al. 2000). Thus if work hours or job conditions affect parents’ mental health and wellbeing, their impact may reach beyond parents to children. In both longitudinal and cross‑sectional research, job insecurity and low control are associated with a range of adverse mental health outcomes, especially depression and anxiety (D’Souza et al. 2003; Ferrie et al. 2001; Marmot et al. 1999; Stansfeld, North & Marmot 1995). Australia has high rates of casual employment and most workers in temporary or casual jobs worry about their job future (Hesselink & van Vuuren 1999). In the context of restructuring and downsizing, even workers not directly under threat report heightened concerns about their futures and worsening job conditions, including lesser control (De Witte & Näswell 2003; Kivimaki et al. 2000).

Children’s wellbeing is an indicator of how they are faring, and strategies to optimise their wellbeing have the potential to yield benefits into the future. Maximising wellbeing in the early years of life is likely to improve outcomes and achievements in later life (Keating & Hertzmann 1999). Therefore, ensuring optimal wellbeing when children are young may be particularly important. It is essential to identify factors influencing children’s wellbeing, and determine how to best support protective factors while minimising harmful factors. The LSAC and HILDA data deliver an important opportunity to study the influence of parent work and wellbeing on children’s wellbeing, as it contains data on a large amount of factors that may influence child wellbeing.

Most research investigating the relationship between parental work and child wellbeing has focused on mothers’ work hours and child wellbeing. This is because for women, time spent at work often represents time away from children. However, longer work hours do not automatically translate into less time spent with children, as parents strive to protect their time with children, rearranging their schedules and compensating for time lost (Craig 2007; Sayer, Bianchi & Robinson 2004). For example, in Australia, university educated mothers

Page 80: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

68 Social Policy research Paper No. 44

spend more time in both paid work and child care activities than other mothers, choosing to address their time constraints by reducing time spent sleeping (Craig 2007 ).

Other aspects of parent’s work, such as how much control they have in their job and how secure they feel about their job and career future, may also impact children’s wellbeing. As well as increasing the likelihood of parents being punitive or harsh with their children (Galinsky 1999), stress brought home from work by one spouse can influence the wellbeing of the other, and thereby potentially confound the negative effect on children (Jones & Fletcher 1996). There is also some evidence to suggest that job insecurity and low job control can lead to reduced levels of self‑esteem and personal control, making parents less able to show warmth to their children or provide a stimulating environment (Cooksey, Menaghan & Jekielek 1997). For example, fathers’ self‑esteem and interactions with their children (including showing warmth and acceptance) are linked to their experiences of control and support at work (Grimm‑Thomas & Perry‑Jenkins 1994).

The section first examines links between employed mothers’ and fathers’ work hours/conditions and their wellbeing. The findings are then extended to examine the relationship between parents’ work and children’s wellbeing. Parents’ mental health and wellbeing is a key determinant of children’s wellbeing and of strong family relationships, so our findings could inform family and child policy and mental health interventions focusing on family and children’s wellbeing and on social sustainability. The section has three aims, to:

�� test for links between parents’ work conditions and hours and their own mental health and wellbeing

�� test whether employed parents’ work conditions and hours are associated with children’s wellbeing

�� examine the association between parents’ wellbeing and their children’s wellbeing.

5.2 Methods

Analyses in this section were conducted using linear multiple regression analyses.1 The analyses were run first with parent wellbeing as the outcome of interest and then again with child wellbeing as the dependent variable. Parent wellbeing is measured in the infant and preschooler cohort via the K6, where higher scores represent more psychological distress. Parent wellbeing in the adolescent cohort is measure via the SF‑36, where higher scores correspond to more positive mental health. Child wellbeing was assessed as socio‑emotional functioning, measured via assessment of irritability and approachability for infants, the SDQ for preschoolers and the SF‑36 for adolescents.

For the first set of analyses, all six work‑related predictors were entered simultaneously into a regression analysis predicting parent mental health. For the second set of analyses, the same six work‑related variables were used to predict child mental health. The work hours and conditions variables are work hours, job security, casual employment, self‑employment, job control and flexible hours. The third set of analyses examined the relationship between parent and child mental health.

All analyses controlled for the following key confounding sociodemographic factors:

�� equivalised household income

�� age of youngest child in the family

�� family type (mother analysis only, lone or couple)

�� parent years of education

�� parent with a chronic medical condition

�� parent age

�� parent self‑employment

Page 81: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

69

WOrk ANd WEllbEINg

�� age and sex of study child

�� regular use of child care

�� mother on maternity or parental leave from employment (infant cohort only).

The first part of this section presents results of analyses examining the relationship between work and parent wellbeing for each cohort analyses. Note analyses were performed separately for mothers and fathers. Results in the second section are presented on the infant and preschooler cohorts only as the small sample size of the adolescent cohort did not permit the accurate calculation of parameter estimates for the relationship between parent work circumstances and adolescent distress.

Sensitivity analysis was undertaken for each work condition where each work condition was modelled in a separate analysis. These analyses revealed comparable results to those presented here. Results modelling each condition separately are available from the authors upon request.

The results for this section are presented in three parts, the first of which examines the relationship between work conditions and parent wellbeing. The second section investigates direct relationships between parents’ work hours and conditions and child wellbeing. Here we ask, for example, whether children of mothers who work part time or who have secure employment report better wellbeing than those of mothers working full time or who do not have secure employment. The final section explores the relationship between parent and child mental health to determine whether parent mental health could be the mechanism linking parent job conditions to child wellbeing.

5.3 Wellbeing and employed mothers’ and fathers’ work hours and conditions

Table 17 summarises a series of multiple regressions analyses (full analyses are presented at Appendix E). Beta values in this table indicate the size of the association between work hours and conditions and parent wellbeing, after controlling for sociodemographic factors.

Job security emerged as a significant correlate of parent wellbeing in the infant and preschooler cohorts. Job security was associated with decreased psychological distress for both mothers and fathers in both cohorts, with a larger effect size for fathers. Job security was not associated with parent wellbeing in the adolescent cohort. However, due to low statistical power (due to a much smaller sample size), we cannot assume that there is no relationship.

Job control also emerged as an important predictor of parent wellbeing. Job control was significantly associated with better mental health for fathers in all three cohorts and also for mothers in the younger two cohorts. Interestingly, the effect sizes were stronger for mothers than fathers in the younger cohorts, suggesting that control may be more strongly linked to wellbeing for mothers. For both mothers and fathers, holding constant any effect of income or how many hours they are working, having control over the work they do is associated with decreased psychological distress.

Page 82: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

70 Social Policy research Paper No. 44

Table 17: Mothers’ and fathers’ wellbeing by work hours and conditions: summary of linear multiple regression analyses

Infant cohortK6 distress

Preschooler cohortK6 distress

Adolescent cohortSF‑36 mental health

Mothers

Work hours (part/full time) 0.06* 0.07** –0.13*

Job security (low/high) –0.07*** –0.06** 0.04

Casual tenure (no/yes) 0.00 0.02 –0.07

Self‑employed (no/yes) 0.10*** 0.04 0.04

Job control (low/high) –0.09*** –0.11*** 0.01

Flexible hours (no/yes) –0.06*** –0.04 –0.01

Fathers

Work hours

Up to 45 hours – – –

Long (45–55 hours) 0.01 –0.01 0.04

Very long (>55 hours) 0.01 0.02 –0.02

Job security (low/high) –0.12*** –0.13*** 0.09

Casual tenure (no/yes) 0.02 –0.05* 0.13*

Self‑employed (no/yes) 0.02 0.01 0.01

Job control (low/high) –0.08*** –0.08*** 0.14*

Flexible hours (no/yes) –0.03 –0.08*** 0.02

Notes: Work hours: 0=part time, 1=full time; job security: 0=low, 1=high; casual tenure: 0=no, 1=yes; self‑employed: 0=no, 1=yes; job control: 0=low, 1=high; flexible hours: 0=no, 1=yes. *p<0.05; **p<0.01; ***p<0.001.

For mothers in all three cohorts working full time is associated with poorer mental health. Fathers’ work hours did not appear to be associated with their wellbeing at any age group. Non‑flexible work hours and self‑employment are both associated with poorer mental health for working mothers of infants. Fathers with preschoolers who work in non‑flexible job also show significantly poorer mental health.

Multivariate models were also run for each work condition separately, and the patterns of results were largely the same as the simultaneous model. The only condition that showed different associations when modelled separately was flexible hours. Flexible work hours were associated with less psychological distress for mothers and fathers in the infant and the preschooler cohorts when modelled separately. These findings (not presented) suggest that the ability to negotiate start and stop times may be related to improve wellbeing for parents.

Page 83: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

71

WOrk ANd WEllbEINg

5.4 Parents’ work hours and conditions and children’s wellbeing

Table 18 presents the standardised beta coefficients for the models predicting children’s wellbeing from mothers’ and fathers’ work circumstances. Mothers’ casual employment was related to infant irritability, such that those in casual employment were less likely to have irritable infants. Fathers with flexible hours tended to have more approachable infants, and fathers working very long hours had less irritable infants. For the preschooler cohort, mothers’ work hours and job security were significantly related to preschooler wellbeing. Mothers in full‑time employment or in insecure jobs reported having preschoolers with more emotional and behavioural difficulties. Self‑employed fathers had preschoolers with better social and emotional wellbeing. Full details of these analyses are presented at Appendix F. Analyses that reported non‑significant results are not reported here but are available from the authors on request.

We explored the role of fathers’ work hours in the relationship between mothers’ full‑time work and preschoolers’ SDQ scores (analyses not presented). In two‑parent families, poorer preschooler wellbeing was observed only when mothers worked full time and fathers worked long or very long hours. Preschoolers whose mothers worked full time but whose fathers worked fewer than 45 hours a week did not show higher SDQ scores than children whose mothers worked part time. These analyses indicated that mothers’ work hours must be considered in the context of fathers’ work hours.

The finding that fathers working very long hours had less irritable infants was unexpected, given the assumption of previous research that long work hours operate to the detriment of child wellbeing. To investigate further we compared fathers’ work hours at a household level. These analyses revealed that, of the 602 families in which the father worked very long hours (55 hours or more per week), 262 (or 44 per cent) had a mother who was unemployed or not in the labour force. Further, 75 per cent of the remaining 340 families had a mother working part time. When we controlled for mothers’ work hours in these analyses, the association between fathers’ very long work hours and infant irritability became non‑significant. The reduced irritability of infants whose fathers were working very long hours may therefore be due to their mothers’ reduced working hours. This finding underlines the importance of considering total work hours at the household level, rather than at the individual level. In two‑parent families, there appears to be a co‑regulation of work hours between parents.

Page 84: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

72 Social Policy research Paper No. 44

Table 18: Association between children’s wellbeing and mothers’ and fathers’ work hours and conditions: summary of linear regression analyses

Infant cohort Preschooler cohort

Approachability (+) Irritability (–) SDQ (total) (–)

Mothers

Work hours (part/full time) –0.03 0.03 0.06**

Job security (low/high) –0.01 0.02 –0.05*

Casual tenure (no/yes) 0.03 –0.06* 0.02

Self‑employed (no/yes) 0.05 0.01 –0.04

Job control (low/high) 0.03 –0.00 –0.04

Flexible hours (no/yes) 0.03 –0.04 –0.01

Fathers

Work hours

Up to 45 hours – – –

Long work hours (45–55 hours) 0.03 –0.04 0.00

Very long (>55 hours) 0.03 –0.05* 0.03

Job security (low/high) –0.01 0.03 0.01

Casual tenure (no/yes) 0.04 –0.02 –0.01

Self‑employed (no/yes) 0.03 0.01 –0.09***

Job control (low/high) 0.01 –0.01 –0.02

Flexible hours (no/yes) 0.04* 0.01 –0.04

Notes: (+) High score indicates positive outcome, (–) High score indicates negative outcome. Work hours: 0=part time, 1=full time; job security: 0=low, 1=high; casual tenure: 0=no, 1=yes; self‑employed: 0=no, 1=yes; job control: 0=low, 1=high; flexible hours: 0=no, 1=yes. *p<0.05; **p<0.01; ***p<0.001.

In order to compare the associations between mothers’ work hours and conditions and child wellbeing (above) with other, better‑known predictors of child wellbeing, we compared the significant beta coefficients of mothers’ work hours and conditions with those of the factors for which we controlled (see Appendix F for full models). Effects sizes for work hours and conditions were small, ranging from 0.04 to 0.06. However, effect sizes of this magnitude are comparable to other well‑known predictors of child wellbeing such as household income. Beta values for the relationship between fathers’ work and their children’s wellbeing ranged from 0.04 to 0.09. Similar to mothers’ work, these coefficients were often stronger than for known risk factors of poor child wellbeing. For example, fathers’ flexible hours were comparable in magnitude to household income and was a stronger predictor of infants’ approachability than were fathers’ years of education, father having a chronic medical condition and regular use of child care.

Though all beta values were relatively weak, these population‑level predictors give an idea of the relative contribution of work conditions to wellbeing compared with other known risk factors for poor child health. These findings indicate that parents’ work hours and conditions could influence child wellbeing independently of, and in some cases more strongly than, other known predictors of child wellbeing.

Page 85: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

73

WOrk ANd WEllbEINg

5.5 Parent and children’s wellbeing

We now examine the relationship between parent and child wellbeing. One mechanism through which parents’ work conditions may relate to children’s wellbeing is via parents’ mental health and mood. There is considerable evidence suggesting that parents’ mental health is a key determinant of their children’s wellbeing, and epidemiological research has established associations between low job control and perceived insecurity at work and adult mental health. However, explicit examinations about how work‑related mental health problems in parents translate into poorer child wellbeing are currently inconclusive.

A summary of the standardised beta values derived from a series of multiple linear regression analyses investigating the associations between parent wellbeing and child wellbeing are presented in Table 19. This table indicates that, overall, there is a relationship between parents’ wellbeing and child wellbeing in the infant and preschooler cohorts. This is evident for both mothers and fathers, but the finding is more consistent for mothers. Measures of parent wellbeing appear unrelated to the wellbeing of their adolescent children, despite considerable evidence showing a well‑established relationship between parent and child wellbeing. Consequently, the lack of association between parent and adolescent wellbeing in this sample is likely due to low statistical power.

Table 19: Associations between child and parents’ wellbeing scores: summary of multiple linear regressions

Infants Preschoolers Adolescents(a)

Approachability Irritability SDQ SF‑36

Mothers –0.07* 0.18** 0.26** 0.12

Fathers –0.03 0.08** 0.26** –0.01

(a) Adolescents’ parent wellbeing was assessed with the SF‑36, not the K6.Note: *p<0.01; **p<0.001.

5.6 Summary

This section explored associations between parents’ work conditions and their own wellbeing, for parents of infants, preschoolers and adolescents. Parents’ mental health and wellbeing is a vital resource for children and has a strong influence on children’s wellbeing. Less than optimal working conditions may have an effect on more than just the worker. Thus, we also examined the links between parent work and their children’s wellbeing.

The results indicate that:

�� For both mothers and fathers with infants and preschool‑aged children, having job control is associated with decreased psychological distress. Job control is also associated with better mental health for working fathers of adolescents.

�� Feeling secure in their job is related to the mental health of mothers and fathers of infants and preschoolers.

�� Work hours are related to mothers’ mental health, with mothers working full time in each cohort reporting poorer mental health than mothers who work part time.

�� Casual employment appears to be associated with poorer wellbeing for fathers.

�� Although there were few direct relationships between mothers’ work hours and conditions and child wellbeing, those that did exist were often of equal magnitude to known risk factors for child wellbeing.

Page 86: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

74 Social Policy research Paper No. 44

�� Fathers working very long hours tended to have less irritable infants than did those working standard hours. Preliminary evidence suggests that this may be because these fathers had partners who were working less than full‑time hours. This finding highlights the importance of considering work hours as a household, rather than an individual, resource.

�� Fathers’ self‑employment was a significant independent predictor of better outcomes for preschoolers’ wellbeing, and flexible hours appear to be related to infant wellbeing. The associations among these factors were at least as large as the associations between known risk factors and child outcomes, suggesting that fathers’ work conditions may be an important source of influence on child wellbeing.

�� Both mothers’ and fathers’ wellbeing predicted child wellbeing in the infant and preschooler cohort. However, parent wellbeing appeared unrelated to adolescent wellbeing.

Page 87: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

75

VulNErAblE FAMIlIES

6 Vulnerable families

6.1 Overview and aims

What sorts of jobs do low‑SEP and lone mothers have and are they more vulnerable to poor working conditions? In this section we examine the relationships between job quality and wellbeing for vulnerable families. By vulnerable families, we mean those in the lowest quartile of socioeconomic position, or families headed by a lone mother. Thus, we examine job conditions and wellbeing for low versus medium/high‑SEP families, as well as families with employed mothers where the mother is a lone parent versus being in a couple family.

There is substantial evidence that parents from families in the lowest quartile of socioeconomic position have poorer quality jobs than parents in the medium/high quartiles. In Australia, low‑income jobs tend to lack security and upward mobility, offer fewer benefits and have higher job turnover as they are less stable and have poorer working conditions (Bernstein 2004; SCARC 2004). Further, workers in these low‑income jobs are disproportionately women, including lone mothers, and people who do not have a post‑secondary educational qualifications (SCARC 2004). Employees in low‑income jobs are also less likely to control when they can start and finish, more likely to report work overload, and less likely to have autonomy in their work (Fenwick & Tausig 2001). Thus, it is likely that the people with greater susceptibility to negative spillover from work conditions into family life will be women, lone mothers and low‑income earners.

In Section 3 we found poorer wellbeing in families with low socioeconomic position. In particular, low‑SEP fathers fared badly in comparison with fathers in medium/high‑SEP quartiles in terms of their mental health. Children from low‑SEP families also had poorer social and emotional functioning in the preschooler cohort. Lone‑mother families also had poorer wellbeing than two‑parent families. In the infant and preschooler cohorts, employed lone mothers showed nearly double the rate of distress of employed mothers in two‑parent households. Preschool‑aged children of employed lone mothers also showed nearly double the rate of social and emotional problems compared with children of employed mothers in two‑parent families, although there were few differences in the infant and adolescent cohorts (see Section 3 for more details). This section extends the investigation of socioeconomic position and family type presented in Section 3 by looking at the vulnerability of families with low‑SEP and lone‑mother families in terms of their work hours and conditions.

This section also extends investigations of the links between work conditions and parent and child wellbeing presented in Sections 4 and 5. These sections showed clear relationships between work conditions and wellbeing. In particular, they demonstrated that aspects of work other than work hours, such as job control, flexible hours and job security, have implications for parent and child wellbeing. The current section extends these investigations by examining whether these relationships differ between low‑SEP families and those in medium/high quartiles, or between lone and two‑parent families.

Analyses in this section may have policy implications in the areas of social inclusion, welfare reforms and industrial relations: social inclusion because we identify families whose needs in society are not being met; welfare reforms because the quality of a job is not currently a criteria for work placement; and industrial relations because we investigate how work hours and conditions relate to wellbeing. They may also inform policy relating to children’s wellbeing, as they identify the family‑level risk factors that could be detrimental to children’s wellbeing and undermine efforts to give children the best start to life (also discussed in Section 5).

In Section 3 we found that, overall, employed lone mothers were more socially disadvantaged than employed mothers from two‑parent households. Employed lone mothers showed a much greater risk of experiencing financial hardship, had less household income and were less educated than employed mothers in two‑parent families, and this was true in all three child‑age cohorts. Although lone mothers share many characteristics

Page 88: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

76 Social Policy research Paper No. 44

of socioeconomically disadvantaged families, there are also key differences. For example, lone mothers have more time pressures, are not supported by a partner and have a greater reliance on child care. Thus our analyses distinguish between lone mothers and low‑SEP families.

We hypothesise that low‑SEP and lone‑mother families could be vulnerable in two ways. First, they could be vulnerable because relatively disadvantaged parents are less able to access good quality jobs, possibly due to their poorer education and lower bargaining power. Second, they could be more vulnerable because they have fewer resources to mitigate any stresses or difficulties associated with poor working conditions. For example, in jobs where flexible hours are not available, lone mothers may become more distressed than couple mothers because they do not have any fallback to help them look after sick children or otherwise juggle their work and family commitments. Further, lone parents and low‑SEP families may not be able to afford services that could help them manage work‑to‑family conflict or pressures. In this sense, socioeconomic position and family type are moderators of the way work conditions affect wellbeing (that is, they may strengthen the degree to which some work conditions are associated with parent and child wellbeing).

Just as not all jobs are equal (and some offer worse conditions that others), not all families have the same access to resources. The aim of this section therefore is to investigate whether:

�� low‑SEP parents have similar or different work conditions compared with medium/high‑SEP families

�� lone mothers have similar or different work conditions compared with employed mothers in couple families

�� low‑SEP parents show similar or different associations between work conditions and wellbeing, compared with medium/high‑SEP families

�� lone mothers show similar or different associations between work conditions and wellbeing, compared with employed mothers in couple families.

Note that investigations by family type for fathers and all multivariate analyses involving the adolescent cohort were not conducted due to small sample sizes (see Section 2.3 ‘Caveats’).

6.2 Methods

This section is presented in two parts. The first contains descriptive analyses for work hours and conditions by family socioeconomic positions and family type. The second presents multivariate analyses examining whether socioeconomically disadvantaged parents and lone mothers are more vulnerable to poor work hours and conditions than other parents.

descriptive analyses

The descriptive analyses examine differences both within and across cohorts. The within‑cohort comparisons allow us to directly examine the association between socioeconomic position and family type on the type of jobs parents have when their children are at specific ages. For example, they allow us to determine the work hours of low versus medium/high‑SEP parents within the infant cohort. The across‑cohort comparisons address the key question of whether there is any change in the work hours and conditions of families according to child age. They allow us to examine differences in parents’ experience of work hours and conditions by child age, and whether these patterns change by socioeconomic position and family type. For example, we investigate whether there is a statistically significant difference in the number of lone mothers working casually in the infant compared with the preschooler cohort.

To illustrate the findings we present graphs of work hours and conditions by socioeconomic position (Figures 8 and 9). Significance test results for comparisons between socioeconomic position and family‑type groups are presented in Tables 20 and 21. Note that the groupings of socioeconomic position and family type in these analyses are not mutually exclusive and so we do not test for differences between socioeconomic position and family‑type groups.

Page 89: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

77

VulNErAblE FAMIlIES

Note that analyses in this section are at the bivariate level and therefore do not control for any potential confounding factors. Multivariate analyses, which do control for potential confounding factors, are presented in the next section. (See Section 2.2 for more details on bivariate and multivariate analyses.)

Multivariate analyses

In this section we explore, at a multivariate level, whether families from low‑SEP and lone‑mother families are more vulnerable to the stressors associated with poor working conditions than their medium/high‑SEP and partnered‑mother counterparts. We do this by investigating the moderating effect of socioeconomic position and family type on the relationship between parents’ working conditions and parent and child wellbeing. We also investigate these moderating effects on the relationship between parents’ wellbeing and children’s wellbeing, as we expect that parents’ work conditions influence children’s wellbeing via parent wellbeing.

Each table in this section displays adjusted means and standard errors of parent and child wellbeing by work hours and conditions, and socioeconomic position or family type, as well as the results of interaction testing. The interaction tests investigate whether the relationship between parents’ work hours and conditions and wellbeing change depending on socioeconomic position or family type, with significant interactions confirming such a change. The tables also show the results of main effect analyses for work hours and conditions, and socioeconomic position or family type. The main effects show the average association with parent or child wellbeing irrespective of the influence of socioeconomic position or family type. The main effects of socioeconomic position and family type show the average association between socioeconomic position or family type and parent or child wellbeing irrespective of levels of work conditions. The mean score differences from the multivariate analyses are also included (adjusted means were calculated to take account of confounding factors such as parents’ age or income affecting the mean outcome scores).

One method of evaluating the size of differences in mean scores is to use the standard deviation, which tells you the average amount of variation from the mean in a dataset (the greater the standard deviation, the greater the variation in the data). Standard deviations for each parent and child outcome variable in each cohort are presented in Section 3.

For relationships between parents’ work conditions and parent and child wellbeing, interactions were tested using ANOVA,2 which yields adjusted marginal mean estimates, useful for comparison with the unadjusted means reported in previous sections. For relationships between parent wellbeing and child wellbeing, we used linear multiple regressions to test for interaction significance because parent wellbeing outcomes are continuous measures and so are not suitable for an ANOVA approach. Note that we applied a significance level of p<0.01 to main effect and interaction analyses, and that we were also not able to conduct investigations on the adolescent cohort or on lone fathers due to small sample sizes (see Section 2.3 ‘Caveats’).

All analyses also controlled for key confounding sociodemographic factors. These sociodemographic factors were:

�� equivalised household income

�� age of youngest child in the family

�� family type (mothers only, lone or couple)

�� number of children in the household

�� parent years of education

�� parent with a chronic medical condition

�� parent age

�� parent self‑employment

Page 90: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

78 Social Policy research Paper No. 44

�� age and sex of study child

�� regular use of child care

�� mother on maternity or parental leave from employment (infant cohort only)

�� parent work hours per week

�� parent self‑employment.

6.3 Jobs in low‑SEP and lone‑mother families

With globalisation, the nature and conditions of jobs in Australia are changing and there is evidence that there is a widening gap between high‑skilled, high‑paid and flexible jobs and low‑skilled, low‑paid jobs with poor conditions. This polarisation in the labour market is partly driven by global competition for low‑skilled jobs, technological advances and the shift to a service and knowledge economy (with the emergence of very highly skilled knowledge jobs and low‑skilled service jobs) (Grzywacz & Dooley 2003). Downsizing, labour market flexibility and deregulation places pressure on job security, and technology has led to the automatisation of many jobs, resulting in job loss for some low‑skilled workers (Glass 2000). Globalisation has also made it easier for business to move across nations, placing downward pressure on wages and conditions of some workers (Glass 2000). Consequently, some segments of the labour force are vulnerable to poor work conditions and job insecurity. We therefore ask, in our three cohorts of families, whether there are differences in the quality and security of jobs held by low compared with medium/high‑SEP families, and between lone‑mother and two‑parent families.

6.4 Work hours and conditions by socioeconomic position and family type

Socioeconomic position

Table 20 displays mothers’ and fathers’ average work hours per week and the proportion with each work condition by socioeconomic position. It also presents the results of within‑cohort and across‑cohort significance tests for work hours and conditions.

Work hours

Figure 8 further illustrates mothers’ and fathers’ work hours by cohort and socioeconomic position. Mothers from socioeconomically advantaged families worked an average of three to four hours more per week than mothers from disadvantaged families in all cohorts, although this was non‑significant in the adolescent cohort. In the infant and preschooler cohorts, fathers from advantaged families worked two hours more per week on average than those from disadvantaged families.

Mothers’ work hours appear to be linked to the age of their child, more so than their socioeconomic position. Mothers in the adolescent cohort were significantly less likely to work part time and significantly more likely to work full time than mothers of infants and preschoolers, regardless of socioeconomic position. Though more mothers from advantaged families report working long hours than their disadvantaged counterparts, this difference did not reach significance.

Fathers were significantly more likely to work part time in low‑SEP families compared with medium/high‑SEP families for both younger cohorts, though this difference was not observed in the adolescent cohort. There were no significant differences in the proportion of low versus medium/high‑SEP fathers working full time or long full time in the infant or preschooler cohorts. In the across‑cohort comparisons, low‑SEP fathers in the adolescent cohort were more likely to work long hours than low‑SEP fathers in the infant cohort.

Page 91: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

79

VulNErAblE FAMIlIES

Regardless of their socioeconomic position, mothers typically work part time when their children are young, with a small proportion working full time. Mothers are more likely to increase to full‑time work hours when their children are adolescents. Consequently, mothers’ work hours appear to be more closely tied to the care needs of their children than the socioeconomic position of the family. Fathers are just as likely to work full time or long hours, regardless of the age of their child. Part‑time work is relatively uncommon for fathers, particularly fathers from medium/high‑SEP families. Fathers from low‑SEP families did report working approximately five hours more per week in the adolescent cohort than in the infant and preschooler cohorts. This could be due to a heightened need for income in low‑SEP families when children reach adolescence or, possibly, the ability to work longer hours once child care responsibilities have reduced.

Figure 8: Proportion of mothers’ and fathers’ working part time, full time or long full‑time hours by socioeconomic position, in each cohort

Infant Preschooler Adolescent

Mother

Part-time Full-time Long Part-time Full-time LongL M–H M–HL M–HL M–HL M–HL M–HL

Father

per c

ent

0

10

20

30

40

50

60

70

80

90

100

Note: L=low socioeconomic position (lowest quintile, lowest 25 per cent); M–H=medium to high socioeconomic position (highest 75 per cent).

Employment tenure

In the preschooler cohort more low‑SEP mothers (compared with medium/high‑SEP mothers) were employed casually and fewer were employed on permanent or fixed‑term tenure. Similar patterns occurred in the infant and adolescent cohorts, but were non‑significant.

Fathers from medium/high‑SEP families in both younger cohorts were more likely to have better quality employment tenure than fathers from low‑SEP families. Medium/high‑SEP fathers were more likely to have permanent tenure and less likely to be employed casually. A similar pattern was found in the adolescent cohort, but the difference was not significant.

The employment tenure of mothers in the adolescent cohort was significantly different to that of mothers in the infant and preschooler cohort, but only for medium/high‑SEP mothers. Mothers from medium/high‑SEP families with infants and preschoolers were more than twice as likely to be self‑employed than similar mothers from the adolescent cohort. Interestingly, the proportion of mothers permanently employed was higher in the infant and adolescent cohorts than in the preschooler cohort. There were no differences in the employment contracts of low‑SEP mothers across the three cohorts.

Page 92: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

80 Social Policy research Paper No. 44

Figure 9: Proportion of mothers and fathers in each type of employment tenure by socioeconomic position, in each cohort

Infant Preschooler Adolescent

Mother

Permanent Fixed-term Casual Self-employed

Permanent Fixed-term Casual Self-employed

L M–H M–HL M–HL M–HL M–HL M–HM–H M–HLLL

Father

per c

ent

0

10

20

30

40

50

60

70

80

90

100

Note: L=low socioeconomic position (lowest quintile, lowest 25 per cent); M–H=medium to high socioeconomic position (highest 75 per cent).

There was a significant difference in the employment tenure of fathers from medium/high‑SEP families in the infant compared with the preschooler cohort. Fathers of infants appear slightly more likely to be permanently employed and slightly less likely to be self‑employed than fathers from the preschooler cohort. For fathers the proportion of permanently employed medium/high‑SEP fathers remained relatively stable across cohorts. However, the proportion of low‑SEP fathers employed permanently dropped somewhat in the adolescent cohort (see Figure 9). Low‑SEP fathers of adolescents were more likely to be self‑employed than their counterparts in the younger cohorts.

Job security

Within all three cohorts, the proportion of mothers and fathers reporting job security did not differ as a function of socioeconomic position. That is, low‑SEP parents were just as likely to report feeling secure in their job as medium/high‑SEP families. However, job security did differ by cohort. Mothers and fathers of adolescents in medium/high‑SEP families were more likely to feel secure in their jobs than mothers and fathers in the infant and adolescent cohorts. This was only evident for parents from medium/high‑SEP families—low‑SEP parents did not differ across the cohorts.

Job control

Mothers’ perceptions of job control did not differ as a function of family socioeconomic position within the cohorts. There were also no differences in the proportion of mothers with control over their work across the cohorts. Regardless of socioeconomic position and age of their child, 77 to 88 per cent of mothers reported feeling they have control over their work.

Fathers from medium/high‑SEP families in the infant and preschooler cohorts were significantly more likely to report job control then their low‑SEP counterparts. There was no difference in the proportion of fathers reporting control across the cohorts.

Page 93: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

81

VulNErAblE FAMIlIES

Flexible hours

Mothers from low and medium/high‑SEP families had similar levels of flexible hours in all three cohorts. However, mothers in the adolescent cohort were significantly less likely to have flexible hours than mothers in both younger cohorts; and this was true for mothers low‑SEP and medium/high‑SEP families. Approximately 80 per cent of mothers in the younger two cohorts felt they have flexible hours, compared with just half of the mothers in the adolescent cohort. Note, however, that this could reflect different measures used in adolescent versus the younger cohorts (see Section 2).

Flexibility did differ by socioeconomic position for fathers in the younger cohorts. Fathers from medium/high‑SEP families in both the infant and preschooler cohort were significantly more likely to have flexible work hours compared with their low‑SEP counterparts. Fathers of adolescents were also significantly less likely to report having flexible hours compared with fathers of both infants and preschoolers. Like mothers, this difference in flexible hours across cohorts was found regardless of the socioeconomic position of the family.

Family type (lone and partnered mothers)

Table 21 displays mothers’ average work hours per week and the proportion of mothers reporting each work condition by family type. Note that these comparisons are only presented for lone mothers as there are very few lone fathers in the samples. Table 21 also presents the results of within and across‑cohort significance tests.

Within each cohort, lone and partnered mothers reported similar work hours, job control and flexible hours, but differed in type of employment tenure and job security. Lone mothers were more likely than partnered mothers to be employed casually and less likely to be self‑employed, although these differences were not significant in the adolescent cohort. Interestingly, lone mothers were more likely to be permanently employed than their partnered counterparts in both of the younger cohorts. Lone mothers were also less likely to report feeling secure in their job in every cohort, although the difference was significant in the preschooler cohort only.

Across‑cohort comparisons revealed that mothers from the adolescent cohort, whether partnered or single, differed in work hours compared with their counterparts in both younger cohorts. On average, mothers from the adolescent cohort worked seven to 11 hours more per week. Lone and couple mothers in the adolescent cohort were more likely to work full‑time or long work hours and less likely to work part‑time hours than mothers from the younger cohorts.

Couple mothers from the adolescent cohort were also more likely to feel secure in their job and less likely to have flexible work hours than couple mothers in the infant and preschooler cohorts. Lone mothers in the adolescent cohort also reported less flexible hours than the preschooler cohort.

Page 94: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

82 Social Policy research Paper No. 44

Table 20: Employed mothers’ and fathers’ work hours and conditions by socioeconomic position: within and across‑cohort comparisons

Infants Preschoolers Adolescents

Lown=283

Med/highn=1,584

Lown=365

Med/highn=1,718

Lown=63

Med/highn=214

Mothers

Work hours (weekly) Mean 19.79 23.18*+ 20.70 24.61* 28.68# 31.95#

Part time (≤35 hours) % 78.10 72.70 80.00 73.90 55.60# 52.80#

Full time (36–45 hours) % 19.10 21.80 17.00 19.60 39.70# 36.40#

Long full time (≥46 hours) % 2.80 5.40 3.00 6.50 4.80 10.80

Job security % 73.90 76.60 75.10 76.90 87.30 86.90#

Employment tenure^

Permanent % 46.60 61.10 39.20 55.20* 52.40 60.30

Fixed term % 3.20 3.00 1.60 6.20 4.80 11.70

Casual % 30.70 15.30 35.90 15.60 28.60 19.20

Self‑employed % 19.40 20.60 23.30 23.00 14.30 8.90

Job control % 78.80 80.40 79.20 82.60 77.80 87.40

Flexible hours % 79.90 80.00 84.10 81.80 49.20# 51.90#

Fathers

Work hours (weekly) Mean 44.14 46.68*+ 45.49 47.67* 50.63# 47.20#

Part time (≤35 hours) % 10.50 5.40* 9.30 5.00* 5.30 4.80

Full time (36–45 hours) % 52.30 50.10 49.90 47.90 38.60 48.10

Long full time (≥46 hours) % 37.20 44.50 40.80 47.10 56.10† 47.10

Job security % 78.50 82.30 80.00 80.50 84.20 93.30#

Employment tenure^

Permanent % 66.60 71.70* 68.10 69.00* 56.10 67.60

Fixed term % 2.20 3.80 1.00 2.80 3.50 5.70

Casual % 10.80 2.90 8.50 2.40 7.10 4.30

Self‑employed % 20.40 21.60 22.40 25.80 33.30 22.40

Job control % 79.60 85.60* 76.60 87.00* 78.90 89.00

Flexible hours % 73.50 84.50* 69.30 85.20* 50.90# 56.70#

Notes: *p<0.01 within‑cohort comparisons; low compared with medium/high‑SEP families.#p<0.01 across‑cohort comparison; adolescents significantly different to both infants and preschoolers.+p<0.01 across‑cohort comparison; significant difference between infants and preschoolers.†p<0.01 across‑cohort comparison; significant difference between infants and adolescents.^Across‑cohort comparisons:

�� mothers’ med/high‑SEP: infant/preschoolers not significant�� infants/adolescents 2(3)=51.15, p<0.001; preschoolers/adolescents 2(3)=28.37, p<0.001�� fathers’ med/high‑SEP: infant/preschoolers 2(3)=13.03, p<0.01; infants/adolescents not significant�� preschoolers/adolescents not significant�� low‑SEP cross‑cohort comparisons of employment tenure for mothers and fathers all not significant.

Page 95: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

83

VulNErAblE FAMIlIES

Table 21: Employed mothers’ work hours and conditions by family type: within and across‑cohort comparisons

Infants Preschoolers Adolescents

Lonen=74

Couplen=1,793

Lonen=198

Couplen=1,885

Lonen=45

Couplen=232

Work hours (per week) Mean 20.92 22.73 25.23 23.78 32.28# 31.00#

Part time (≤35 hours) % 79.70 73.20 70.70 75.40 46.70# 54.70#

Full time (36–45 hours) % 13.50 21.80 22.70 18.80 44.40# 35.80#

Long full time (≥46 hours) % 6.80 5.00 6.60 5.80 8.90# 9.50#

Job security % 68.90 76.50 66.20 77.70* 80.00 88.40#

Employment tenure^

Permanent % 60.80 58.80* 55.10 52.00* 57.80 58.60

Fixed term % 5.40 2.90* 5.10 5.50* 11.10 9.90

Casual % 27.00 17.30* 30.30 18.00* 26.70 20.30

Self‑employed % 6.80 21.00* 9.50 24.50* 4.40 11.20

Job control % 74.30 80.40 77.30 82.50 77.80 86.60

Flexible hours % 75.70 80.10 82.30 82.20 53.30~ 50.90#

Notes: *p<0.01.~p<0.01 across‑cohort comparison; significant difference between preschoolers and adolescents.#p<0.01 across‑cohort comparison; adolescents significantly different to both infants and preschoolers.^Across‑cohort comparisons:

�� couple mothers: infant/preschoolers 2(3)=26.92, p<0.001; infants/adolescents 2(3)=38.60, p<0.001; preschoolers/adolescents 2(3)=24.89, p<0.001

�� lone mothers: all not significant—there is no difference between lone mothers of infants, preschoolers or adolescents on any measure of job characteristic.

6.5 low‑SEP families and lone mothers and job conditions

Relationships between work and wellbeing are well established in epidemiological research. Low job control, high work demands and job insecurity are demonstrated as being associated with poor physical and mental health outcomes in cross‑sectional and longitudinal studies (D’Souza et al. 2003; Ferrie 1999; Sanderson et al. 2005; Stansfeld, North & Marmot 1995). This section investigates whether, in addition to their increased participation in poor quality jobs, low‑SEP parents and lone mothers are also more susceptible to the negative consequences of these jobs. It extends previous research by investigating whether the negative influence of poor job conditions may extend beyond individual employees, across generations, to their children.

6.6 Interactions with socioeconomic position

Parents’ working conditions and wellbeing by socioeconomic position

Tables 22 and 23 display adjusted means and standard errors of parent wellbeing by work hours or conditions, socioeconomic position and the results of interaction testing. They also show the results of main effect analyses for work hours and conditions and socioeconomic position.

Page 96: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

84 Social Policy research Paper No. 44

Table 22: Association between employed mothers’ psychological distress and mothers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions

Infant cohort Preschooler cohort

Low SEP Med/high SEP Low SEP Med/high SEP

Work hours

Part time (SE) 3.30 (0.22) 3.11 (0.09) 3.65 (0.20) 3.63 (0.09)

Full time (SE) 3.20 (0.40) 3.57 (0.16) 4.07 (0.39) 4.00 (0.16)

Interaction F(1,1841)=1.33, ns F(1,2057)=0.01, ns

Job security

No (SE) 4.04 (0.37) 3.70† (0.16) 3.98 (0.35) 4.20 (0.17)

Yes (SE) 3.07 (0.22) 3.09 (0.09) 3.67 (0.20) 3.58 (0.09)

Interaction F(1,1840)=0.59, ns F(1,2056)=0.50, ns

Self‑employed

No (SE) 3.11 (0.22) 3.10† (0.09) 3.78 (0.21) 3.66 (0.09)

Yes (SE) 4.08 (0.42) 3.75 (0.18) 3.60 (0.36) 3.92 (0.17)

Interaction F(1,1841)=0.41, ns F(1,2057)=1.00, ns

Job control

No (SE) 3.74 (0.40) 4.05† (0.18) 4.69 (0.38) 4.52†† (0.19)

Yes (SE) 3.20 (0.22) 3.04 (0.09) 3.50 (0.20) 3.55 (0.09)

Interaction F(1,1840)=0.95, ns F(1,2056)=0.23, ns

Flexible hours

No (SE) 3.61 (0.42) 3.89 (0.18) 4.94 (0.43) 4.07†† (0.19)

Yes (SE) 3.24 (0.21) 3.07 (0.09) 3.50 (0.20) 3.65 (0.09)

Interaction F(1,1840)=0.79, ns F(1,2056)=4.00, ns

Notes: †p<0.01; ††p<0.001; daggers indicate significance level of a work‑related main effect.ns=not significant; SE=standard error; SEP=socioeconomic position.

There were no significant main effects of socioeconomic position and no interactions between work hours and conditions and socioeconomic position for mothers. This indicates that socioeconomic position was not a significant predictor of mothers’ distress in the infant or preschooler cohorts and that socioeconomic position did not change the relationships between work hours and conditions and mothers’ distress. There was a tendency for mothers from low‑SEP families to be more distressed than mothers from medium/high‑SEP families, but this result should be interpreted with caution as it was inconsistent and did not reach significance in any analysis. It is worth noting that the interaction between flexible hours and socioeconomic position approached significance in the preschooler cohort. This interaction was such that the difference in distress between mothers with and without access to flexible hours was more pronounced among low‑SEP mothers. This suggests that low‑SEP mothers may be more vulnerable to this poor work condition.

Consistent with Section 5, there were main effects for the job security, self‑employment and job control in the infant cohort and job control and flexible work hours in the preschooler cohort. These job conditions were associated with better psychological function, except for self‑employment, which is associated with higher distress for mothers. There were no main effects for socioeconomic position, demonstrating that mother psychological wellbeing is not associated with the socioeconomic position of her family.

Page 97: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

85

VulNErAblE FAMIlIES

Table 23: Association between employed fathers’ psychological distress and fathers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions

Infant cohort Preschooler cohort

Low SEP Med/high SEP Low SEP Med/high SEP

Long hours (45–55 hrs/week)

No (SE) 3.18 (0.18) 3.31 (0.09) 3.67 (0.19) 3.34 (0.09)

Yes (SE) 3.19 (0.32) 3.37 (0.14) 3.77 (0.33) 3.22 (0.15)

Interaction F(1,2803)=0.02, ns F(1,2598)=0.35, ns

Very long hours (>55 hrs/week)

No (SE) 3.28 (0.16) 3.32 (0.08) 3.34 (0.17) 3.27 (0.08)

Yes (SE) 3.10 (0.35) 3.36 (0.16) 4.10 (0.36) 3.29 (0.16)

Interaction F(1,2803)=0.32, ns F(1,2598)=3.31, ns

Job security

No (SE) 4.45 (0.28) 4.09†† (0.15) 4.58 (0.32) 4.17†† (0.16)

Yes (SE) 2.89 (0.15) 3.15 (0.07) 3.21 (0.16) 3.09 (0.08)

Interaction F(1,2804)=2.94, ns F(1,2599)=0.58, ns

Self‑employed

No (SE) 3.04 (0.15) 3.33 (0.07) 3.52 (0.17) 3.29 (0.08)

Yes (SE) 3.94 (0.29) 3.27 (0.14) 3.35 (0.30) 3.32 (0.14)

Interaction F(1,2805)=6.96, p<0.01 F(1,2600)=0.27, ns

Job control

No (SE) 3.81 (0.29) 4.04†† (0.18) 4.08 (0.30) 4.21†† (0.19)

Yes (SE) 3.08 (0.15) 3.20 (0.07) 3.31 (0.17) 3.16 (0.07)

Interaction F(1,2804)=0.10, ns F(1,2599)=0.51, ns

Flexible hours

No (SE) 3.73 (0.26) 3.65† (0.17) 4.20 (0.26) 3.94†† (0.18)

Yes (SE) 3.05 (0.16) 3.26 (0.07) 3.16 (0.17) 3.19 (0.08)

Interaction F(1,2804)=0.66, ns F(1,2599)=0.56, ns

Notes: †p<0.01; ††p<0.001; daggers indicate significance level of the working arrangement main effect.ns=not significant; SE=standard error; SEP=socioeconomic position.

Table 23 shows a significant interaction between self‑employment and socioeconomic position in the infant cohort, which is displayed in Figure 10. This figure shows that, among fathers from low‑SEP families, self‑employed fathers reported higher rates of distress than fathers who were not self‑employed. This pattern was reversed among medium/high‑SEP fathers, who reported slightly lower distress when self‑employed than when not. Independent t‑tests confirmed the difference between self‑employed and non–self employed fathers from low‑SEP families (t(549)=–2.32, p<0.05), and indicated that there was no corresponding difference for fathers from medium/high‑SEP families (t(2261)=0.55, ns). This interaction may reflect a poorer quality of self‑employment for fathers from low‑SEP families, or indicate that low‑SEP fathers are more vulnerable than medium/high‑SEP fathers to the increased risks and responsibilities of self‑employment.

Page 98: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

86 Social Policy research Paper No. 44

Figure 10: Employed fathers’ psychological distress by self‑employment and socioeconomic position (infant cohort)

Low SEP Med–high SEP

Not self-employed Self-employed

Mea

n de

pres

sion

2.0

2.5

3.0

3.5

4.0

A similar (although non‑significant) result occurred for very long work hours in the preschooler cohort. Here, medium/high‑SEP fathers showed almost no difference in distress levels when working very long work hours compared with shorter work hours. However, low‑SEP fathers working very long work hours were more distressed than fathers not working these hours. Again, these results suggest that low‑SEP fathers may be more vulnerable to the increased distress associated with very long work hours. Long work hours may also function as ‘exposure’ to poor quality jobs for low‑SEP fathers who tend to hold poorer quality jobs.

When accounting for sample stratification and clustering the interaction between fathers’ self‑employment and socioeconomic position predicting fathers’ distress (infant cohort) became marginally non‑significant (p=0.058). It was also non‑significant when accounting for under or over‑representation of some groups. Thus this finding should be treated with caution. Appendix B presents further information and the results of these analyses.

Table 23 also shows that there were main effects for job security, job control and flexible hours in both cohorts. All of these conditions are associated with improved wellbeing for fathers of infants and preschoolers. Like mothers, results showed that fathers’ psychological wellbeing is not associated with the socioeconomic position of his family.

Parents’ wellbeing and children’s wellbeing by socioeconomic position

There were no significant interactions between mothers’ wellbeing and socioeconomic position and infants’ or preschoolers’ wellbeing (see Table 24 for a summary of results), indicating that the relationship between mothers’ wellbeing and preschoolers’ wellbeing did not differ between low and medium/high‑SEP families. Further, there were no significant main effects of mothers’ wellbeing or socioeconomic position on infant wellbeing, suggesting that infant wellbeing was not directly affected by mothers’ wellbeing or socioeconomic position in these analyses. Preschoolers’ wellbeing, however, appeared to be influenced by mothers’ distress, with higher distress associated with higher SDQ scores (and therefore poorer preschooler wellbeing). Preschoolers’ SDQ scores were also consistently associated with socioeconomic position, with preschoolers from low‑SEP families having higher SDQ scores than those from medium/high‑SEP families.

Page 99: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

87

VulNErAblE FAMIlIES

Table 24: Standardised beta coefficients for the effect of parent wellbeing, socioeconomic position and interactions on infant and preschooler wellbeing

Infants Preschoolers

Approachability Irritability SDQ

Mothers

Parent wellbeing 0.02 0.11 0.37**

Socioeconomic position 0.00 0.03 –0.12**

Interaction –0.11 0.08 –0.12

Fathers

Parent wellbeing –0.01 0.07 0.04

Socioeconomic position 0.05** 0.02 –0.19**

Interaction 0.02 0.01 0.05

Note: **p<0.001.

There were also no significant interactions between fathers’ wellbeing and infant or preschooler wellbeing (see Table 24). This means that the relationship between fathers’ wellbeing and infants’ and preschoolers’ wellbeing did not differ by socioeconomic position.

Socioeconomic position showed a direct association with infants’ approachability and preschoolers’ SDQ scores: children from low‑SEP families were less approachable and had more behavioural difficulties than those from medium/high‑SEP families.

Work hours and conditions and children’s wellbeing by socioeconomic position

There were no significant main effects of mothers’ work hours and conditions or socioeconomic position on infants’ approachability scores, indicating that mothers’ work hours and conditions and socioeconomic position did not significantly influence infants’ approachability scores. (See Appendix H for adjusted means and standard errors for each work condition, along with interaction terms for the infant cohort.) There were also no significant interactions between mothers’ work hours and conditions and socioeconomic position, demonstrating that the relationship between work hours and conditions and approachability scores did not differ for mothers from low or medium/high‑SEP families.

As shown in Table 25, infants with mothers working part time were significantly less irritable than infants with mothers working full time. This finding may be due to the greater amount of time that mothers working part time are able to spend with their children, or could be related to an increase in distress felt by mothers working full time. Additional analyses (not shown) indicated that this association was reduced (although remained marginally significant) once mothers’ psychological distress was added into the equation. No other work arrangement was significantly related to infant irritability. Socioeconomic position did not significantly influence infant irritability in either the main effects or interactions.

Page 100: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

88 Social Policy research Paper No. 44

Table 25: Association between infants’ irritability and employed mothers’ work hours by socioeconomic position: adjusted means (standard errors) and interaction

Low socioeconomic position Med/high socioeconomic position

Part time work hours (SE) 2.36 (0.06) 2.46† (0.02)

Full time work hours (SE) 2.65 (0.10) 2.50 (0.04)

Interaction F(1,1853)=4.67, ns

Notes: †p<0.05; daggers indicate significance level of the working arrangement main effect.ns=not significant; SE=standard error.

Table 26 shows mean SDQ scores for preschoolers by mothers’ work arrangement and socioeconomic position. Note that a high SDQ score indicates that a preschooler displays more problem behaviours. The main effect of socioeconomic position was significant for every work hour and condition variable, demonstrating that preschoolers from low‑SEP families displayed more problem behaviours than preschoolers from medium/high‑SEP families. There were no significant interactions between work hours and conditions and socioeconomic position, indicating that mothers’ work hours and conditions did not change the relationship between socioeconomic position and preschoolers’ SDQ scores.

Table 26: Association between preschoolers’ SDQ (total) scores and employed mothers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions

Low socioeconomic position Med/high socioeconomic position

Part time work hours (SE) 9.69 (0.28) 8.04** (0.13)

Full time work hours (SE) 10.04 (0.55) 8.55 (0.22)

Interaction F(1,2069)=0.05, ns

Job security

No (SE) 9.80 (0.50) 8.87** (0.23)

Yes (SE) 9.79 (0.29) 7.96 (0.13)

Interaction F(1,2068)=2.10, ns

Self‑employed

No (SE) 9.82 (0.29) 8.23** (0.13)

Yes (SE) 9.65 (0.51) 7.96 (0.24)

Interaction F(1,2069)=0.03, ns

Job control

No (SE) 10.71 (0.54) 8.56** (0.27)

Yes (SE) 9.55 (0.28) 8.09 (0.13)

Interaction F(1,2068)=1.08, ns

Flexible hours

No (SE) 10.74 (0.62) 8.23** (0.27)

Yes (SE) 9.60 (0.28) 8.16 (0.13)

Interaction F(1,2068)=2.18, ns

Notes: **p<0.001; asterisks indicate significance level socioeconomic position main effect.ns=not significant. SDQ=Strengths and Difficulties Questionnaire; SE=standard error.

There was no relationship between infant irritability scores and fathers’ work hours/conditions and socioeconomic position. Interactions between fathers’ work hours/conditions and socioeconomic position were not significant in any analysis. This suggests that socioeconomic position was not associated with a change in the relationship between work hours/conditions and infants’ irritability.

Page 101: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

89

VulNErAblE FAMIlIES

Fathers with flexible work hours have infants who are more approachable than infants of fathers with non‑flexible hours (see Table 27). There was also a main effect for socioeconomic position in this analysis, where infants from medium/high‑SEP families tended to be more approachable than infants from low‑SEP families. There were no other main effects nor interactions between fathers’ work hours or conditions and socioeconomic position on infants’ approachability.

Table 27: Association between infants’ approachability and employed fathers’ flexible work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions

Low socioeconomic position Med/high socioeconomic position

Flexible hours

No (SE) 4.50 (0.07) 4.76*† (0.04)

Yes (SE) 4.75 (0.04) 4.79 (0.02)

Interaction F(1,2819)=6.47, ns

Notes: *p<0.01; asterisks indicate significance level socioeconomic position main effect; †p<0.01; daggers indicate significance level of the working arrangement main effect. ns=not significant; SE=standard error.

Preschoolers from low‑SEP families had consistently higher SDQ scores than preschoolers from medium/high‑SEP families, indicating that they had more emotional difficulties (see Table 28). Further, the interactions between socioeconomic position and work hours and conditions were not significant in any analysis, suggesting that the association between socioeconomic position and preschoolers’ SDQ scores is a stand‑alone association that is not moderated by fathers’ work arrangements. Also consistent with Section 5, there were also no main effects for any work condition and preschoolers’ wellbeing.

Page 102: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

90 Social Policy research Paper No. 44

Table 28: Association between preschoolers’ SDQ (total) scores and employed fathers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions

Low socioeconomic position Med/high socioeconomic position

Long hours (45–55 hrs/week)

No (SE) 10.64 (0.28) 7.39** (0.13)

Yes (SE) 11.08 (0.48) 7.97 (0.22)

Interaction F(1,2618)=1.64, ns

Very long hours (>55 hrs/week)

No (SE) 10.49 (0.25) 8.11** (0.12)

Yes (SE) 11.22 (0.53) 8.10 (0.24)

Interaction F(1,2618)=1.50, ns

Job security

No (SE) 10.55 (0.48) 8.13** (0.23)

Yes (SE) 10.56 (0.24) 8.15 (0.11)

Interaction F(1,2619)=0.00, ns

Self‑employed

No (SE) 10.79 (0.25) 8.32**† (0.12)

Yes (SE) 9.87 (0.45) 7.65 (0.21)

Interaction F(1,2620)=0.18, ns

Job control

No (SE) 10.45 (0.44) 8.69** (0.28)

Yes (SE) 10.60 (0.25) 8.07 (0.11)

Interaction F(1,2619)=1.78, ns

Flexible hours

No (SE) 10.69 (0.38) 8.67** (0.27)

Yes (SE) 10.51 (0.26) 8.06 (0.11)

Interaction F(1,2619)=0.65, ns

Notes: **p<0.001; asterisks indicate significance level socioeconomic position main effect. †p<0.01; daggers indicate significance level of the working arrangement main effect. ns=not significant; SE=standard error.

6.7 Interactions with family type

Mothers’ work hours and conditions and wellbeing by family type

Table 29 displays results of interaction testing for the association between mothers’ work hours and conditions, family type and her wellbeing. The table shows the adjusted means and standard errors of mothers’ wellbeing by work hours and conditions and family type, and the results of main effects and interaction testing. Note that there are fewer significant results for the associations between work hours and conditions on mothers’ wellbeing in this section than in the overall analyses in Section 5. This is due to the more stringent significance level (p<0.01) applied to the current analyses and the inclusion of the interaction variable.

Lone mothers reported significantly higher rates of distress than partnered mothers, particularly in the infant cohort (see Table 29). Analyses also identified main effects for job control (infant cohort) and flexible hours (infant and preschooler cohort); both job conditions are associated with decreased distress.

Page 103: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

91

VulNErAblE FAMIlIES

There was a significant interaction between mothers’ job control and family type, as illustrated in Figure 11. Job control measures how much freedom people feel they have over how they do their job. Independent t‑tests showed that, for both lone and partnered mothers, high job control was associated with less distress than low job control (t(72)=3.82, p<0.001 for lone mothers; t(1776)=3.97, p<0.001 for partnered mothers). As shown in Figure 11, the greater distress experienced by mothers with high compared with low job control was more pronounced for lone mothers than for partnered mothers. While partnered mothers without job control had distress scores only marginally higher than those without job control, lone mothers without job control had vastly elevated distress scores. This pattern suggests that lone mothers are particularly vulnerable to a lack of control over their working environments.

Table 29: Associations between employed mothers’ distress and employed mothers’ work hours and conditions by family type: adjusted means (standard errors) and interactions

Infant cohort Preschooler cohort

Lone Couple Lone Couple

Work hours

Part time (SE) 4.37 (0.42) 3.07* (0.09) 4.11 (0.30) 3.56 (0.09)

Full time (SE) 5.13 (0.81) 3.52 (0.15) 4.90 (0.43) 3.98 (0.16)

Interaction F(1,1840)=0.12, ns F(1,2057)=0.48, ns

Job security

Low (SE) 5.35 (0.66) 3.66* (0.15) 4.29 (0.42) 4.15 (0.16)

High (SE) 4.20 (0.44) 3.05 (0.08) 4.41 (0.30) 3.52 (0.09)

Interaction F(1,1839)=0.46, ns F(1,2056)=1.97, ns

Self‑employed

No (SE) 4.32 (0.39) 3.06 (0.08) 4.24 (0.26) 3.63 (0.09)

Yes (SE) 5.99 (1.39) 3.71 (0.17) 5.12 (0.76) 3.77 (0.16)

Interaction F(1,1840)=0.49, ns F(1,2057)=0.84, ns

Job control

Low (SE) 7.28 (0.71) 3.79**†† (0.17) 4.33 (0.50) 4.58 (0.18)

High (SE) 3.65 (0.43) 3.04 (0.08) 4.42 (0.28) 3.46 (0.08)

Interaction F(1,1840)=11.61, p<0.01 F(1,2056)=4.18, ns

Flexible hours

No (SE) 6.40 (0.74) 3.68**†† (0.17) 5.24 (0.57) 4.05* (0.18)

Yes (SE) 3.96 (0.43) 3.07 (0.08) 4.18 (0.28) 3.58 (0.08)

Interaction F(1,1839)=4.54, ns F(1,2056)=0.81, ns

Notes: *p<0.01; **p<0.001; asterisks indicate significance level of the main effects of family type. ††p<0.001; daggers indicate significance level of the main effect of working arrangement. ns=not significant.

Page 104: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

92 Social Policy research Paper No. 44

Figure 11: Employed mothers’ distress by job control and family type (infant cohort)

Single Couple

LowJob control

High

Mea

n de

pres

sion

0

1

2

3

4

5

6

7

8

When accounting for sample stratification, clustering and under or over‑representation of some groups, the interaction between mothers’ job control and family type (infant cohort) became marginally non‑significant (p=0.079). This suggests that we should interpret the interaction with caution. Appendix B presents further information and the results of these analyses.

Mothers’ wellbeing and children’s wellbeing by family type

There were no significant interactions between mothers’ wellbeing and family type on children’s wellbeing (see Table 30). This indicates that family type was not related to the relationship between mothers’ wellbeing and children’s wellbeing.

Table 30: Standardised beta coefficients for the effect of parent wellbeing, socioeconomic position and interactions on infant and preschooler wellbeing

Parent wellbeingSocioeconomic

positionInteraction

Infant approachability –0.09 –0.01 0.01

Infant irritability –0.02 –0.04 0.20

Preschooler emotional and behavioural problems

0.24 –0.01 0.03

Mothers’ hours and conditions and children’s wellbeing by family type

There were no significant interactions between mothers’ work hours/conditions and family type on infant and children’s wellbeing. This indicated that children in lone and two‑parent families were affected in a similar way by their mother’s working conditions. See Appendix I for a summary of all models.

6.8 Summary

This section has discussed the particular vulnerabilities created by poor quality work conditions in two groups, low‑SEP families and lone mothers. In doing so, it has demonstrated that these groups tend to have poorer

Page 105: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

93

VulNErAblE FAMIlIES

quality jobs than their medium/high‑SEP and partnered‑mother counterparts, and it has provided suggestive evidence that they are more susceptible to the stress associated with these poor working conditions. We also investigated the wellbeing of children in low‑SEP and lone‑mother families compared with those in medium/high‑SEP and coupled‑mother families, but found few differences between these groups.

Multivariate analysis in this section found that:

�� Work hours are associated with socioeconomic position. Both mothers and fathers from medium/high‑SEP families worked more hours per week than their low‑SEP counterparts.

�� Employment tenure is also related to socioeconomic position. Parents from low‑SEP families were more likely to be employed casually and less likely to be employed on permanent tenure than those from medium/high‑SEP families.

�� As with parents from low‑SEP families, lone mothers were more likely than partnered mothers to be employed casually.

�� Lone mothers were less likely to have a secure job future in their job than employed partnered mothers.

�� Medium/high‑SEP fathers worked similar hours in each cohort while their low‑SEP counterparts worked more hours in the adolescent cohort than in the infant and preschooler cohorts.

�� While over one‑third of low‑SEP fathers worked long full‑time hours in the infant cohort, this increased to more than half in the adolescent cohort. There was no significant corresponding difference for medium/high‑SEP fathers.

�� The proportions of medium/high‑SEP and coupled mothers reporting job control increased in each successive cohort. In contrast, low‑SEP and lone mothers showed no significant increase in their job control between the cohorts. Lone mothers and low‑SEP mothers and fathers were less likely to report job control than their partnered and medium/high‑SEP counterparts.

�� The distress associated with inflexible work hours was more pronounced among low‑SEP mothers than among medium/high‑SEP mothers in the preschooler cohort, suggesting that low‑SEP mothers are more vulnerable to this work condition. However, this trend should be interpreted with caution, as it did not reach statistical significance.

�� Among fathers from low‑SEP families, self‑employed fathers reported higher rates of distress than fathers who were not self‑employed. This pattern was reversed among medium/high‑SEP fathers, who reported slightly lower distress when they were self‑employed than when they were not. Future research should investigate the quality of self‑employment when parents are relatively socioeconomically disadvantaged.

�� The wellbeing of employed lone mothers may be particularly supported by good quality jobs. The mental health of both coupled and lone mothers was optimised when they had some control over their job tasks. However, when jobs did not offer mothers some freedom over how they did their job, lone mothers showed relatively higher psychological distress compared with coupled mothers.

�� Neither socioeconomic position nor family type changed the relationship between parents’ wellbeing and children’s wellbeing for employed mothers or fathers in any cohort.

�� Relationships between both mothers’ and fathers’ work conditions and infants’ temperament were the same, regardless of family socioeconomic position.

�� However, preschoolers’ socio‑emotional issues were associated with family socioeconomic position. Children in medium/high‑SEP families had significantly better wellbeing than children in low‑SEP families.

�� Children in lone and partnered‑mother families were affected in similar ways by their mother’s work hours and conditions.

Page 106: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

94 Social Policy research Paper No. 44

Page 107: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

95

WOrk ANd FAMIly WEllbEINg—WhAT dOES IT MEAN?

7 Work and family wellbeing—what does it mean?

There is currently a national policy priority to maximise workforce participation, particularly for women, to help support Australia’s ageing population. Concurrent with this is the aim to give children a healthy start to life to support the development of their mental, physical and academic capabilities, which will also aid Australia’s social and economic future. Jobs, parents, children and family life lie at the centre of these two priorities.

This study investigated how work and family life are intertwined. Work and the quality of work have the capacity to influence family health and wellbeing. The nature and quality of parental work is linked to their own wellbeing and, through this, to their children’s outcomes. Access to good quality jobs (that is, those with flexibility, security and control) may help optimise both parent and child wellbeing. What happens at work and the sorts of jobs Australian parents have access to could have consequences reaching across generations.

This project used data from LSAC and the HILDA survey to provide new evidence about the influence of work on parents’ and children’s wellbeing. We investigated this link for children at three different age points (infants, preschoolers and adolescents) and also for vulnerable families. Families from low socioeconomic circumstances were compared to their higher socioeconomic position counterparts. We also compared breadwinner families with dual‑earner families and lone parents compared with two‑parent families.

Results indicate that the quality and conditions of employed parents’ jobs matter to their own and their children’s wellbeing. Socioeconomic disadvantage was found to be persistently linked to poorer access to good quality jobs, especially for mothers. Mothers (but not fathers) modulate their work hours according to their children’s age and partner’s work and hold down jobs in the context of having partners with heavy work‑time commitments. Mothers’ employment appears to be performed in a different ‘time context’ to fathers.

7.1 Families across the cohorts

We found evidence that parents’ wellbeing varied across age cohorts. Mothers’ psychological distress was higher when their children were preschooler‑aged compared with mothers of infants or adolescents. Fathers’ wellbeing did not vary across cohort.

Within cohorts, socioeconomic circumstance was linked to both parent and child wellbeing, especially for children. Thus, socioeconomic disadvantage appears to couple with health and wellbeing disadvantages, with possible compounding impacts over the life course. This suggests that inventions to remedy child disadvantage should target family socioeconomic circumstances and family wellbeing.

Compared with couple mothers, employed lone mothers of preschoolers reported more psychological distress and were more likely to be above the threshold for distress. Similar trends were also apparent in the infant and adolescent cohorts, though they did not quite reach statistical significance. These findings indicate that although employment may help lone mothers financially, these mothers continue to face other hardships and difficulties such as managing their work and care responsibilities. Unaddressed, continuing poor mental health is likely to erode these mothers’ capacity to both work and provide care, with likely adverse consequences for their sustained work force participation and, possibly, for their children. Improving the assistance available to mothers may improve their own and their children’s wellbeing and maximise mothers’ workforce participation.

Page 108: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

96 Social Policy research Paper No. 44

7.2 Mothers’ and fathers’ jobs—different or equal?

Our analyses indicate that the conditions and quality of jobs available to parents may be a key driver of their wellbeing. Although the sample size limits any clear conclusion, the findings from the adolescent cohort suggest that job conditions continue to be important for employed parents with older children. Further, while jobs help parents via income and social inclusion, poor conditions could undermine these benefits. Recent Australian analyses indicate that jobs that combine insecurity with poor control and heavy workloads could be equally damaging to mental health as unemployment (Broom et al. 2006). Some international studies classify people along an ‘employment continuum’ of unemployment and exposure to adverse job conditions. For example, unemployed people in the United States showed the worst health, but poor quality jobs with adverse conditions such as low control conferred only small health benefits, while adults in high‑quality jobs showed the best health (Grzywacz & Dooley 2003). Our findings point to similar conclusions. Our findings also underline the need for work–family policy to think across the whole spectrum of children’s developmental needs. For example, while policies that allow access to leave when parents have infants are clearly necessary, our findings suggest that shorter work hours and on‑the‑job control could benefit parents of adolescents. Given the small sample size for this cohort, future research is required to further test the role played by other work conditions.

On the one hand, while much policy focus has been on the benefits delivered by jobs to families, our research suggests that poor conditions could erode parents’ wellbeing. This is in turn likely to lower productivity and increase healthcare costs. Our findings also indicate that maximising the security, control and flexibility of jobs will optimise the benefits of employment to families.

We found that mothers and fathers do not always work in jobs with similar conditions. Although access to flexible work hours was similar for mothers and fathers, mothers reported more job insecurity and less control than fathers. Similarly, although mothers reported working more hours (on average) when their children were older, there was no evidence that the gender gap narrowed for job security, casual versus permanent employment, or job control as their children aged.

Being a dual‑earner parent, or the breadwinner, is a different experience depending on whether you are a mother or father. Mothers in the labour force do not have the same access to time that fathers do. In two‑parent families, mothers’ jobs are nearly always worked in the context of a full‑time employed partner (who often works long hours). By contrast fathers are more likely to have a partner who is employed part time or not in the labour force (particularly when children are young). Fathers do not appear to change their work hours when their children get older, nor do they work fewer hours (on average) if their partner is also employed.

Our findings indicate that mothers and fathers hold down jobs in the context of different time resources. Looking across the cohorts, it seems that mothers do not experience the same ‘flexibility’ from their partner’s work hour commitments that fathers do. Mothers arrange their paid work around their children and their partners’ jobs and this is apparent across all three cohorts. At the same time, expectations and requirements to work long hours are on the rise. Australia has one of the longest work hour cultures in the OECD. Long hours are a feature of senior jobs in particular and so mothers are competing for, or working in, these jobs from a different time‑resource base. The recent OECD report on usual working hours by sex illustrates how long‑hour cultures drive a gender gap in work time (OECD 2005). In countries where long hours are expected, like Australia, there is a long ‘tail’ (largely men) working long full‑time hours and an equally long tail of women in the opposite direction, working part‑time hours. Countries where work hour expectations for full‑time jobs are more or less confined within 35 to 40 hours a week (for example, Denmark, Finland and France) show a different gender distribution. In these countries, the majority of men and women are working full time, within this time band, with much smaller long hour or part‑time ‘tails’.

There has been a large focus on increasing fathers’ time commitments to caring. However, tackling expectations around full‑time jobs is also important. A long work hour culture is linked to work intensification

Page 109: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

97

WOrk ANd FAMIly WEllbEINg—WhAT dOES IT MEAN?

and most developed countries have reported a steady increase in work intensity, linked to global competition in labour costs. If left unaddressed, a culture of long full‑time hours could continue to underpin a less visible form of gender discrimination in the labour market, as success in jobs, especially senior jobs, requires access to time.

7.3 Work and parent and child wellbeing

To date, little large‑scale epidemiological research has investigated the relationship between parental employment circumstances and child wellbeing. These findings highlight the link and add new knowledge about child development and wellbeing that should be considered when developing policy.

Our analyses indicate that the conditions and quality of parents’ jobs may be a key driver of their wellbeing. Results also indicate that children have a stake in the nature of their parents’ work arrangements and the quality of their parents’ jobs. We have shown that different sorts of jobs are systematically associated with different levels of child wellbeing. Much research has documented a link between parent and child wellbeing, and we have confirmed the relationship in these samples. The link between parent and child mental health may hold the key to explain why parent jobs are related to children’s wellbeing: the effect that jobs have on children’s wellbeing may be a function of parent mental health.

Our findings may inform the ways in which jobs could be designed to give children and families the opportunity to flourish. These findings suggest important links from parent work hours and job conditions to child wellbeing. In particular, they suggest that work arrangements may have important implications for child wellbeing. Policy designed to maximise child wellbeing could be enhanced by including a focus on improving parental working conditions.

A likely mechanism through which parents’ work conditions may indirectly relate to children’s wellbeing is via parents’ mental health and mood. There is considerable evidence suggesting that parents’ mental health is a key determinant of their children’s wellbeing. Epidemiological research has established associations between low job control and perceived insecurity at work and adult mental health. However, explicit examinations of work‑related mental health problems in parents influencing child wellbeing are currently inconclusive. Several studies have shown that work stress, potentially linked to work conditions and work–family strain, affects family functioning and how parents interact with their children.

Work can be satisfying as well as stressful, energising as well as exhausting, and positive work‑related experiences can also influence the quality of family life (Galinsky 1999). Thus, our findings may also flag which aspects of jobs are beneficial for wellbeing. For example, an alternative interpretation of the finding that non‑flexible hours relate to poor wellbeing is that flexible work hours benefit wellbeing. Policy advances in improving employment‑related conditions may reduce parental distress. Our research indicates that any such advances would, in turn, flow on to increase the wellbeing of children at all stages of development.

Jobs deliver clear benefits to families and increasing workforce participation (particularly for women) is an important policy priority. However, our research suggests that poor job conditions could erode parents’ wellbeing and may lower productivity and increase healthcare costs. Our findings indicate that maximising the security, control and flexibility of jobs will optimise the benefits of employment to families.

Findings also suggest that encouraging participation in the workforce would not necessarily improve child wellbeing; care would need to be taken to ensure that parents enter jobs that are consistent with their goals and responsibilities for their children. To optimise child wellbeing, such jobs would be flexible in their required hours, secure in their future and allow individual control over the tasks performed. This would help minimise negative trade‑offs parents currently face between earning income and maximising their children’s wellbeing. Policy aimed at supporting family health and wellbeing and increasing workforce participation need to be considered in tandem and aim for complimentary outcomes.

Page 110: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

98 Social Policy research Paper No. 44

7.4 Vulnerable families

There are some groups in society for whom the labour market is a particularly harsh environment. We examined the relationship between work and wellbeing for two disadvantaged groups: lone mothers and families with low socioeconomic position. Parents from low‑SEP families are disadvantaged in terms of their household income, education and occupational prestige. Our results indicate that they are also disadvantaged because they work in jobs with less security, control and with poorer conditions than their higher socioeconomic position peers. Similarly, lone mothers are disadvantaged not only because they lack the support of a partner, but also because they tend to work in less secure jobs.

Lone parents and low‑SEP parents are more likely to encounter poor employment circumstances than are other parents and these circumstances appear to hit them harder. Such circumstances also mean the children of these parents are more vulnerable to poorer outcomes. Our research suggests that there may be ways to short‑circuit such compounding jeopardy. Vulnerable parents (and therefore, also, their children) may be particularly responsive to supportive employment circumstances. Raising minimum standards would ensure that even the most vulnerable are employed under more flexible conditions, with better job control and security. However, our findings also suggest that wages and income support remain a critical part of the picture; even when employed, many lone mothers still report high rates of financial hardship.

Negative employment experiences and conditions appear to be particularly damaging for vulnerable parents who already live with underlying disadvantage. These are also the very people who are most likely to encounter poor employment circumstances. This is particularly pertinent given that our findings suggest underlying disadvantage interacts with poor employment circumstances. Positive employment circumstances for vulnerable parents could deliver clear advantages to their own and their children’s wellbeing.

7.5 limitations

This study has several limitations that must be acknowledged. First, our analyses are based on cross‑sectional data, which does not permit causation to be inferred. However, both LSAC and the HILDA survey are longitudinal datasets. Research that extends on the findings presented here would benefit from capitalising on this to look at these relationships over time. Future research should also take steps towards testing for a mediating relationship between parental work conditions, parental wellbeing and child wellbeing.

Despite the large cohort sizes for LSAC, there are relatively few employed lone mothers or breadwinner mothers in either the infant or preschooler cohort. This limits the reliability of our estimates and the ability to detect statistically significant findings when we conduct analyses on these subgroups of employed parents. Similarly, the adolescent cohort from the HILDA survey was too small to conduct complex multivariate analyses, especially on subgroups in the sample. Ideally future work would also examine families where the mother is breadwinner and more fully examine the link between parent work and adolescent wellbeing.

Another important limitation concerns the measurement of parent work conditions in LSAC and the HILDA survey. Most are based on one or two items, which are less precise than multi‑item measures and so may underestimate the strength of association among the key variables. (Compare, for example, the effect sizes reported in D’Souza et al. (2003) and Strazdins et al. (2004) from Australian data using multi‑item measures.) In population‑level datasets there is a trade‑off between the depth and breadth of measures. Both the LSAC and the HILDA survey briefly measure a wide range of important family characteristics, attitudes, experiences and functions. Consequently, it is likely that all estimates of effect size in this report are underestimated due to the brief measures used in the surveys.

Page 111: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

99

WOrk ANd FAMIly WEllbEINg—WhAT dOES IT MEAN?

7.6 Conclusion

In investigating whether parents’ work impacts children’s wellbeing, this report raises important questions about the intergenerational consequences of parents’ work. It also explores variations in the relationship between work and wellbeing at different points in children’s development by examining this relationship for three cohorts of children. Further, by distinguishing different types of families, it recognises that some families may be more vulnerable to particular work conditions than others. Rather than simply considering children’s wellbeing when parents are either employed or unemployed, this report provides a detailed investigation of how the conditions of employed parents’ work impacts on their own and their children’s wellbeing.

Parents’ employment circumstances are related to their own and their children’s wellbeing, and there is considerable variation around the details of this relationship. Some features of jobs affect mothers but not fathers, or vulnerable parents but not those living with more favourable socioeconomic circumstances, and so on. Some features of jobs show direct links to children, while other pathways are indirect, through their impact on parents. Despite these subtleties, the main story from this research is that parents’ employment circumstances may have impacts on their own wellbeing and on the wellbeing of their children.

Insecure jobs, inflexible hours and poor job control are associated with an increase in parental distress. Parents’ unfavourable employment experiences may undermine children’s wellbeing—largely because they erode parents’ wellbeing. Conversely, ensuring that jobs give parents some flexibility, control and security may help optimise parents’ and children’s wellbeing, yielding both short and long‑term benefits to families.

Negative employment experiences and conditions appear to be particularly damaging for vulnerable families, who are the very people who are most likely to encounter poor employment circumstances. The greater the pressure on parents, including work‑related pressures, the greater is the potential harm to children’s wellbeing. Positive employment circumstances for parents could deliver clear advantages to children’s wellbeing and development, particularly to disadvantaged children.

Page 112: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

100 Social Policy research Paper No. 44

Page 113: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

101

APPENdIx A

Appendix A: Missing data analysisVariables with significant missing data

For both LSAC and the HILDA survey, we computed an indicator of whether a case had complete data on all the variables in the study or was missing one or more of the variables in the study. Table A1 shows the number of parents with complete data used for analysis in the current study. These numbers reflect the number of parents currently in paid employment and, for the adolescent sample, those who had parenting responsibilities for children aged 17 years or younger.

Table A1: Number of parents with complete data in each cohort

Infant cohort Preschooler cohort Adolescent cohort

Employed mothers with complete data 1,867 2,083 91

Employed fathers with complete data 2,832 2,633 279

LSAC infant cohort (mothers)

We performed independent sample t‑tests to check for differences between employed mothers with complete versus incomplete data on age, years of education, combined yearly income and mental health.

The means of employed mothers with and without complete data are reported in Table A2. This table shows that, compared with mothers with incomplete data, mothers with complete data were slightly younger, had slightly more years of education and were slightly less depressed.

Table A2: Means (and standard deviations) for employed mothers with and without complete data (infant cohort)

Complete data Incomplete data

Mean age, years (SD) 31.80*** (4.63) 34.73 (5.01)

Mean years of education (SD) 15.58*** (2.34) 15.05 (2.51)

Mean depression score (SD) 3.24*** (3.15) 3.69 (3.35)

Notes: ***p<0.001. SD=standard deviation. K6 scores grouped as follows: low (0–3); moderate (4–7); high (8–12); and very high (13–24).

LSAC infant cohort (fathers)

We performed independent sample t‑tests to check for differences between employed fathers with and without complete data on age, years of education, and combined yearly income. Means and standard deviations are presented in Table A3. This table shows that fathers with complete data were slightly younger, had slightly more years of education and had slightly higher levels of combined yearly income.

Page 114: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

102 Social Policy research Paper No. 44

Table A3: Means and (standard deviations) for employed fathers with and without complete data (infant cohort)

Complete data Incomplete data

Mean age, years (SD) 33.69** (5.55) 34.47 (6.53)

Mean years of education (SD) 14.98** (2.32) 14.58 (2.33)

Mean combined yearly income bracket (SD) 11.00*** (2.11) 10.28 (2.21)

Notes: **p<0.01; ***p<0.001. SD=standard deviation.

LSAC preschooler cohort (mothers)

We performed independent sample t‑tests to check for differences between employed mothers with and without complete data on years of education and combined yearly income. Means and standard deviations are presented in Table A4. This table shows that, compared to those with incomplete data, mothers with complete data had slightly more years of education and slightly higher combined yearly incomes. This suggests that, in the preschooler cohort, the sample of employed mothers with complete data may be slightly biased towards mothers who are better educated and who have slightly higher incomes and therefore a slightly higher socioeconomic position.

Table A4: Means and (standard deviations) for employed mothers with and without complete data (preschooler cohort)

Complete data Incomplete data

Mean years of education (SD) 15.09* (2.49) 14.75 (2.61)

Mean combined yearly income bracket (SD) 11.27*** (2.35) 10.36 (2.45)

Notes: *p<0.05; ***p<0.001. SD=standard deviation.

LSAC preschooler cohort (fathers)

We performed independent sample t‑tests to check for differences between employed fathers with complete versus incomplete data on age, years of education, combined yearly income and mental health. Means and standard deviations are presented in Table A5. This table shows that, compared to those with incomplete data, fathers with complete data had slightly more years of education, slightly higher combined yearly incomes and were slightly more depressed.

Table A5: Means and (standard deviations) for employed fathers with and without complete data (preschooler cohort)

Complete data Incomplete data

Mean years of education (SD) 14.98** (2.46) 14.57 (2.54)

Mean combined yearly income bracket (SD) 11.42*** (2.06) 10.45 (2.52)

Mean mental health score (SD) 3.33* (3.20) 2.82 (2.68)

Notes: *p<0.05; **p<0.01; ***p<0.001. SD=standard deviation. K6 scores grouped as follows: low (0–3); moderate (4–7); high (8–12); and very high (13–24).

HILDA survey adolescent sample (households with 15 to 17 year‑old children only)

There was no significant bias of age, years of education, income or mental health in the sample of employed mothers with complete data or employed fathers with complete data.

Page 115: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

103

APPENdIx b

Appendix B: Technical appendix—study strata, clustering and weightingSample design and non‑response effects

We conducted comparative analyses of selected key findings to check for design effects and non‑response in the LSAC cohorts. LSAC used a complex sampling design that was stratified by state (‘strata’) and clustered by postcode (‘primary sampling units’). This sampling design violates the assumption of the independence of observations and may inflate the standard errors of estimates. We used the SPSS Complex Samples module to rerun selected key findings, adjusted first for the design effects (primary sampling unit and strata) and then by also weighting the variables using the sample weights. In the LSAC cohorts, sample weights were calculated using post‑stratification using the 2001 Australian Census data; the sample who completed the interview were compared with census data for families with similar aged children and two main predictors of non‑response were observed: families where the mother spoke a language other than English, and families where the mother did not complete Year 12. Sample weights do not take into account missing data or non‑response on the self‑complete questionnaire (from which many of our variables are drawn), although initial comparisons reported in the technical paper found little additional benefit in calculating weights to address this source of non‑response. Readers are referred to two publications that detail sample design and approaches to sample weighting:

�� Soloff, C, Lawrence, D & Johnson, R 2005, Sample design, LSAC Technical Paper No. 1, Australian Institute of Family Studies, Melbourne.

�� Soloff, C, Lawrence, D, Misson, S & Johnson, R 2006, Wave 1 weighting and non‑response, LSAC Technical Paper No. 3, Australian Institute of Family Studies, Melbourne.

Path C—impact of parent work conditions on child wellbeing

Table B1: Influence of mothers’ wellbeing outcomes on children’s outcomes: comparison of outcomes from original, clustered and stratified, and weighted, clustered and stratified data

Original Clustered and stratifiedWeighted, clustered and

stratified

estimate SE p estimate SE p estimate SE p

Casual employment on infant’s irritability

–0.14 0.05 0.012 –0.14 0.05 0.013 –0.12 0.09 0.04

Ordinal job control on infant’s approachability

0.06 0.02 0.020 0.06 0.02 0.023 0.05 0.03 0.08

Job security on work on preschooler’s SDQ

–0.65 0.24 0.007 –0.65 0.26 0.012 –0.64 0.27 0.019

Ordinal job control on preschooler’s SDQ

–0.27 0.13 0.046 –0.27 0.14 0.05 –0.33 0.14 0.019

Notes: Figures in bold indicate changes in significance from original findings. SDQ=Strengths and Difficulties Questionnaire. SE=standard error.

Page 116: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

104 Social Policy research Paper No. 44

Table B2: Influence of fathers’ wellbeing outcomes on children’s outcomes: comparison of outcomes from original, clustered and stratified, and weighted, clustered and stratified data

Original Clustered and stratifiedWeighted, clustered and

stratified

Estimate SE p Estimate SE p Estimate SE p

Flexible hours on infant’s approachability

0.10 0.04 0.017 0.10 0.04 0.023 0.13 0.04 0.004

Flexible hours on preschooler’s SDQ

–0.54 0.02 0.028 0.54 0.24 0.028 –0.59 0.27 0.032

Notes: Figures in bold indicate if there are changes in significance from original findings. SDQ=Strengths and Difficulties Questionnaire. SE=standard error.

Path b—impact of parent wellbeing on child wellbeing

Table B3: Influence of mothers’ wellbeing outcomes on infants’ irritability: comparison of outcomes from original, clustered and stratified, and weighted, clustered and stratified data

Original Clustered and stratifiedWeighted, clustered and

stratified

Estimate SE p Estimate SE p Estimate SE p

Psychological distress 0.04 0.01 <0.000 0.04 0.01 <0.000 0.04 0.01 <0.000

Work–family gain 0.03 0.03 ns 0.03 0.03 ns 0.04 0.03 ns

Work–family strain 0.11 0.02 <0.000 0.11 0.03 <0.000 0.17 0.03 <0.000

Notes: Figures in bold indicate if there are changes in significance from original findings. ns=not significant. SE=standard error.

Table B4: Influence of fathers’ wellbeing outcomes on infants’ irritability: comparison of outcomes from original, clustered and stratified, and weighted, clustered and stratified data

Original Clustered and stratifiedWeighted, clustered and

stratified

Estimate SE p Estimate SE p Estimate SE p

Psychological distress 0.01 0.01 0.005 0.01 0.01 0.009 0.02 0.01 0.004

Work–family gain 0.02 0.03 ns 0.02 0.02 ns 0.03 0.03 ns

Work–family strain 0.09 0.02 <0.000 0.09 0.02 <0.000 0.08 0.02 0.001

Notes: Figures in bold indicate if there are changes in significance from original findings. ns=not significant. SE=standard error.

Page 117: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

105

APPENdIx b

Table B5: Influence of mothers’ wellbeing outcomes on preschoolers’ SDQ scores: comparison of outcomes from original, clustered and stratified, and weighted, clustered and stratified data

Original Clustered and stratifiedWeighted, clustered and

stratified

Estimate SE p Estimate SE p Estimate SE p

Psychological distress 0.32 0.03 <0.000 0.32 0.03 <0.000 0.33 0.04 <0.000

Work–family gain –0.31 0.16 0.045 –0.31 0.15 0.39 –0.37 0.16 0.22

Work–family strain 0.47 0.14 0.001 0.47 0.14 0.001 0.43 0.14 0.002

Notes: Figures in bold indicate if there are changes in significance from original findings. SDQ=Strengths and Difficulties Questionnaire. SE=standard error.

Table B6: Influence of fathers’ wellbeing outcomes on preschoolers’ SDQ scores: comparison of outcomes from original, clustered and stratified, and weighted, clustered and stratified data

Original Clustered and stratifiedWeighted, clustered and

stratified

Estimate SE p Estimate SE p Estimate SE p

Psychological distress 0.09 0.04 0.024 0.08 0.04 0.03 0.10 0.04 0.019

Work–family gain –0.26 0.18 ns –0.26 0.19 ns –0.22 0.21 ns

Work–family strain 0.55 0.17 0.001 0.55 0.16 0.001 0.53 0.17 0.002

Notes: Figures in bold indicate if there are changes in significance from original findings. ns=not significant. SDQ=Strengths and Difficulties Questionnaire. SE=standard error.

Table B7: Interaction between fathers’ self‑employment and socioeconomic position on fathers’ psychological distress (infant cohort)

Original Clustered and stratifiedWeighted, clustered and

stratified

Low SEP Med/high SEP Low SEP Med/high SEP Low SEP Med/high SEP

Self‑employed 3.04 3.33 3.04 3.33 3.11 3.32

Not self‑employed 3.94 3.27 3.94 3.27 3.86 3.34

Interaction F(1,2805)=6.96, p<0.01 Wald F(1,268)=3.63, p=0.058

Wald F(1, 268)=2.16, p=0.143

Notes: Figures in bold indicate changes in significance from original findings. SEP=socioeconomic position.

Page 118: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

106 Social Policy research Paper No. 44

Table B8: Interaction between mothers’ job control and family type on mothers’ psychological distress (infant cohort)

Original Clustered and stratifiedWeighted, clustered and

stratified

Lone Couple Lone Couple Lone Couple

Low job control 7.28 3.79 7.28 3.79 6.58 3.89

High job control 3.65 3.04 3.63 3.04 3.86 3.08

Interaction F(1,1840)=11.61, p<0.01 Wald F(1,265)=7.625, p<0.01

Wald F(1,265)=3.01, p=0.079

Note: Figures in bold indicate changes in significance from original findings.

Table B9: Interaction between mothers’ flexible hours and family type and psychological distress (preschooler cohort)

Original Clustered and stratifiedWeighted, clustered and

stratified

Lone Couple Lone Couple Lone Couple

No flexible hours 3.44 3.62 3.44 3.62 3.49 3.59

Flexible hours 3.85 3.67 3.85 3.67 3.87 3.66

Interaction F(1,2071)=6.99, p<0.01 Wald F(1,263)=3.54, p=0.061

Wald F(1,263)=2.58, p=0.109

Note: Figures in bold indicate changes in significance from original findings.

results

We note that adjusting for clustering and/or using sample weighting has minimal impact on the point estimates or significance of main effects estimates in the multiple regressions, except where estimates are already marginal or borderline. As per the main report, we recommend caution when interpreting such marginal or borderline findings.

More substantial differences are observed in analyses from the section on vulnerable families in the significance testing for the interaction effects. The point estimates and standard errors changed little after adjustment for the design effect. After also adjusting for the sample weights, we found that there were only moderate changes in the standard errors, while the point estimates and test statistics for the interactions were more variable. This is not surprising, as the interaction term is a function of changes in four point estimates. We therefore recommend further caution in interpreting these interactions.

Conclusion

Our results reinforce the robustness of key findings in most multiple regression analyses except where estimates are already marginal or are based on small group numbers. They also highlight the need for more detailed and sophisticated approaches to analysing complex surveys.

Page 119: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

107

APPENdIx C

Appendix C: Employed mothers on leaveWe created an indicator to identify mothers in the infant and preschooler cohort who reported being ‘employed’ but who were also away from work in the week prior to the study interview due to being on paid or unpaid maternity leave. Note that we do not know how long these mothers have been on leave or when the leave began.

Table C1: Employed mothers on leave

Infants Preschoolers

Mothers(n=1,867)

Mothers(n=2,083)

Total no. employed and on leave n%

35819.2

452.2

Disadvantaged % 17.3 2.2

Advantaged % 19.5 2.2

Lone employed % 14.9 1.0

Couple employed % 19.4 2.3

Breadwinner % 19.0 1.6

Dual earner % 19.4 2.3

We performed independent samples t‑tests and chi‑square tests comparing mothers who were employed and currently working with mothers who were employed and on leave on work hours and conditions, depression and work–family conflict.

In the infant cohort, employed mothers on leave had consistently different working hours and conditions to employed mothers who were currently working at the time of the LSAC interview. Specifically, employed mothers on leave:

�� worked on average 12 hours more per week

�� had more job security (82 per cent versus 75 per cent)

�� had more secure employment tenure (93 per cent permanent versus 51 per cent permanent)

�� had less job control (68 per cent versus 83 per cent) and less flexible hours (69 per cent versus 83 per cent).

Employed mothers on leave also had less work–family gain and overall work–family synergy and more work–family strain than employed mothers who were currently working; however, their levels of depression were the same.

In the preschooler cohort, employed mothers on leave had more secure employment tenure only with 84 per cent on permanent tenure compared to 52 per cent of employed mothers who were currently working.

Page 120: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

108 Social Policy research Paper No. 44

Page 121: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

109

APPENdIx d

Appendix D: Scale itemsKessler6

Parent Mental Health—LSAC infant and preschooler cohorts

In the past 4 weeks how often …�� Did you feel nervous?

�� Did you feel hopeless?

�� Did you feel restless or fidgety?

�� Did you feel that everything was an effort?

�� Did you feel so sad that nothing could cheer you up?

�� Did you feel worthless?

SF‑36 Mental Health subscale

Parent and Adolescent Mental Health—HILDA survey

How much of the time during the past 4 weeks have you...�� Been a nervous person?

�� Felt so down in the dumps nothing could cheer you up?

�� Felt calm and peaceful?

�� Felt down?

�� Been a happy person?

Short Temperament Scale for Infants (STSI)

Infant temperament—LSAC infant cohort

Approachability items�� This baby is pleasant (smiles, laughs) when first arriving unfamiliar places (friend’s house, shops).

�� This baby’s first reaction (at home) to approach by strangers is acceptance.

�� This baby accepts within a few minutes a change in place of bath or person giving bath.

�� This baby’s first reaction to seeing a doctor or infant welfare sister is acceptance (smiles, coos).

Irritability items�� This baby is fretful on waking up and/or going to sleep (frowns, cries).

�� This baby amuses self for 1/2 hour or more in cot or playpen (looking at mobile, playing with toy etc).

�� This baby continues to cry in spite of several minutes of soothing.

�� This baby cries when left to play alone.

Page 122: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

110 Social Policy research Paper No. 44

Strengths and Difficulties Questionnaire (SDQ)

Preschooler behavioural and emotional difficulties—LSAC preschooler cohort

Pro‑social behaviour subscale�� Considerate of other people’s feelings

�� Shares readily with other children (treats, toys, pencils, etc.)

�� Helpful if someone is feeling hurt, upset or ill

�� Kind to younger children

�� Often volunteers to help others (parents, teachers, other children)

Hyperactivity/inattention subscale�� Restless, overactive, cannot stay still for long

�� Constantly fidgeting or squirming

�� Easily distracted, concentration wanders

�� Can stop and think things out before acting

�� Sees tasks through to end, good attention span

Emotional symptoms subscale�� Often complains of headaches, stomach aches or sickness

�� Many worries, often seems worried

�� Often unhappy, downhearted or tearful

�� Nervous or clingy in new situations, easily loses confidence

�� Many fears, easily scared

Conduct/oppositional problems subscale�� Often has temper tantrums or hot tempers

�� Generally obedient, usually does what adults request

�� Often fights with other children or bullies them

�� Often argumentative with other adults

�� Can be spiteful to others

Peer relationship problems subscale�� Rather solitary, tends to play alone

�� Has at least one good friend

�� Generally liked by other children

�� Picked on or bullied by other children

�� Gets on better with adults than with other children

Page 123: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

111

APPENdIx E

Appendix E: Work and parent wellbeing: full regression modelsMothers

Table E1: Mothers’ psychological distress (K6) and job conditions—LSAC infant cohort

B SE B p 95% CI for B

Work hours (part/full time) 0.07 0.03 0.06 0.015 0.01 0.13

Job security –0.09 0.03 –0.07 0.001 –0.15 –0.04

Casual employment 0.00 0.03 0.00 0.917 –0.06 0.07

Self‑employed 0.13 0.03 0.10 0.000 0.07 0.19

Job control –0.11 0.03 –0.09 0.000 –0.17 –0.05

Flexible hours –0.08 0.03 –0.06 0.007 –0.14 –0.02

Household equivalised income 0.00 0.00 –0.07 0.009 0.00 0.00

Age of youngest child 0.07 0.05 0.03 0.148 –0.03 0.17

Family type (single or couple) –0.22 0.06 –0.08 0.000 –0.34 –0.10

Years of education 0.01 0.01 0.04 0.123 0.00 0.02

Chronic medical condition 0.12 0.03 0.09 0.000 0.06 0.17

Age 0.01 0.00 0.05 0.056 0.00 0.01

Constant 0.57 0.14 0.000 0.30 0.84

R2 0.24

Adjusted R2 0.05

Page 124: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

112 Social Policy research Paper No. 44

Table E2: Mothers’ psychological distress (K6) and job conditions—LSAC preschooler cohort

B SE B p 95% CI for B

Work hours (part/full time) 0.08 0.03 0.07 0.003 0.03 0.14

Job security –0.08 0.03 –0.06 0.007 –0.13 –0.02

Casual employment 0.02 0.03 0.02 0.453 –0.04 0.09

Self‑employed 0.05 0.03 0.04 0.073 –0.01 0.11

Job control –0.15 0.03 –0.11 0.000 –0.21 –0.09

Flexible hours –0.05 0.03 –0.04 0.110 –0.11 0.01

Household equivalised income 0.00 0.00 –0.06 0.018 0.00 0.00

Age of youngest child 0.00 0.01 0.01 0.560 –0.01 0.02

Family type (single or couple) –0.12 0.04 –0.06 0.007 –0.20 –0.03

Years of education 0.01 0.01 0.04 0.093 0.00 0.02

Chronic medical condition 0.18 0.03 0.14 0.000 0.13 0.24

Age 0.00 0.00 –0.04 0.081 –0.01 0.00

Constant 0.88 0.13 0.000 0.63 1.13

R2 0.24

Adjusted R2 0.06

Table E3: Mothers’ mental health (SF‑36) and job conditions—HILDA adolescent cohort

B SE B p 95% CI for B

Work hours (part/full time) –3.56 1.81 –0.13 0.05 –7.13 0.01

Job security 1.60 2.56 0.04 0.53 –3.45 6.64

Casual employment –2.46 2.27 –0.07 0.28 –6.93 2.02

Self‑employed 1.78 3.01 0.04 0.56 –0.42 7.71

Job control 0.33 2.46 0.01 0.89 –4.51 5.18

Flexible hours –0.08 1.79 –0.01 0.96 –3.61 3.45

Household equivalised income –4.28E–5(a) 0.00 –0.10 0.12 0.00 0.00

Age of youngest child 0.80 0.27 0.21 0.01 0.28 1.33

Family type (single or couple) 3.39 2.41 0.09 0.16 –1.36 8.13

Years of education 0.22 0.34 0.04 0.52 –0.44 0.88

Chronic medical condition –6.57 2.41 –0.16 0.01 –11.30 –1.83

Age –0.18 0.20 –0.06 0.37 –0.58 0.22

Constant 12.21

R2 0.08

Adjusted R2 0.04

(a) –4.28E–5 means –0.0000428.

Page 125: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

113

APPENdIx E

Fathers

Table E4: Fathers’ psychological distress (K6) and job conditions—LSAC infant cohort

B SE B p 95% CI for B

Work hours

Full time (35–45 hrs/wk) Reference group

Long hours (45–55 hrs/wk) 0.02 0.02 0.01 0.47 –0.03 0.06

Very long (55+ hrs/wk) 0.01 0.03 0.01 0.65 –0.04 0.06

Job security –0.16 0.03 –0.12 0.00 –0.21 –0.11

Casual employment 0.04 0.05 0.02 0.40 –0.05 0.14

Self‑employed 0.03 0.03 0.02 0.24 –0.02 0.08

Job control –0.11 0.03 –0.08 0.00 –0.17 –0.06

Flexible hours –0.04 0.03 –0.03 0.10 –0.10 0.01

Household equivalised income 0.00 0.00 –0.05 0.03 0.00 0.00

Age of youngest child –0.03 0.04 –0.01 0.49 –0.11 0.05

Years of education 0.01 0.00 0.06 0.00 0.01 0.02

Chronic medical condition 0.11 0.02 0.08 0.00 0.06 0.15

Age 0.00 0.00 0.01 0.75 0.00 0.00

Constant 0.61 0.09 0.00 0.43 0.79

R2 0.19

Adjusted R2 0.04

Table E5: Fathers’ psychological distress (K6) and job conditions—LSAC preschooler cohort

B SE B p 95% CI for B

Work hours

Full time (35–45 hrs/wk) Reference group

Long hours (45–55 hrs/wk) –0.01 0.02 –0.01 0.699 –0.06 0.04

Very long (55+ hrs/wk) 0.03 0.03 0.02 0.315 –0.03 0.08

Job security –0.18 0.03 –0.13 0.000 –0.23 –0.13

Casual employment –0.13 0.06 –0.05 0.017 –0.24 –0.02

Self‑employed 0.01 0.02 0.01 0.604 –0.04 0.06

Job control –0.13 0.03 –0.08 0.000 –0.18 –0.07

Flexible hours –0.12 0.03 –0.08 0.000 –0.17 –0.06

Household equivalised income 0.00 0.00 –0.04 0.062 0.00 0.00

Age of youngest child 0.00 0.01 –0.01 0.779 –0.01 0.01

Years of education 0.01 0.00 0.04 0.054 0.00 0.02

Chronic medical condition 0.14 0.02 0.11 0.000 0.09 0.19

Age 0.00 0.00 –0.02 0.457 –0.01 0.00

Constant 0.85 0.09 0.000 0.66 1.03

R2 0.23

Adjusted R2 0.05

Page 126: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

114 Social Policy research Paper No. 44

Table E6: Fathers’ mental health (SF‑36) and job conditions—HILDA adolescent cohort

B SE B p 95% CI for B

Work hours

Full time (35–45 hrs/wk) Reference group

Long hours (45–55 hrs/wk) 1.30 2.16 0.04 0.55 –2.95 5.55

Very long (55+ hrs/wk) –0.49 2.43 –0.02 0.84 –5.29 4.30

Job security 4.78 3.31 0.09 0.15 –1.73 11.29

Casual employment 8.92 4.36 0.13 0.04 0.33 17.51

Self‑employed 0.24 2.24 0.01 0.92 –4.18 4.65

Job control 5.99 2.84 0.14 0.04 0.40 11.58

Flexible hours 0.67 1.92 0.02 0.73 –3.10 4.45

Household equivalised income 1.84E–5(a) 0.00 0.04 0.52 0.00 0.00

Age of youngest child –0.19 0.24 –0.05 0.44 –0.67 0.29

Years of education –0.41 0.36 –0.07 0.26 –1.13 0.30

Chronic medical condition –5.24 2.28 –0.14 0.02 –9.74 –0.75

Age 0.22 0.21 0.07 0.29 –0.19 0.63

Constant 10.60

R2 0.08

Adjusted R2 0.04

(a) 1.84E–5 means 0.0000184.

Page 127: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

115

APPENdIx F

Appendix F: Work and child wellbeing: full regression modelsMothers

Table F1: Infant approachability (STSI) and mothers’ job conditions—LSAC infant cohort

B SE B p 95% CI for B

Work hours (part/full time) –0.05 0.05 –0.03 0.262 –0.14 0.04

Job security –0.02 0.04 –0.01 0.658 –0.10 0.07

Casual employment 0.07 0.05 0.03 0.203 –0.04 0.18

Self‑employed 0.09 0.05 0.05 0.082 –0.01 0.19

Job control 0.05 0.05 0.03 0.252 –0.04 0.15

Flexible hours 0.07 0.05 0.03 0.148 –0.02 0.16

Household equivalised income 0.00 0.00 0.03 0.208 0.00 0.00

Age of youngest child 0.06 0.17 0.02 0.737 –0.27 0.39

Family type (single or couple) –0.01 0.10 0.00 0.883 –0.21 0.18

Years of education (mother) 0.00 0.01 –0.01 0.583 –0.02 0.01

Chronic medical condition (mother) 0.04 0.04 0.02 0.381 –0.05 0.13

Age (mother) –0.01 0.00 –0.06 0.008 –0.02 0.00

Age (child) –0.09 0.04 –0.06 0.013 –0.16 –0.02

Sex (child) –0.06 0.02 –0.20 0.000 –0.10 –0.03

Regular child care –0.05 0.04 –0.03 0.288 –0.13 0.04

On paid/unpaid maternity leave 0.04 0.06 0.02 0.513 –0.08 0.16

Constant 5.79 0.24 0.000 5.33 6.26

R2 0.22

Adjusted R2 0.05

Page 128: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

116 Social Policy research Paper No. 44

Table F2: Infant irritability (STSI) and mothers’ job conditions—LSAC infant cohort

B SE B p 95% CI for B

Work hours (part/full time) 0.05 0.04 0.03 0.298 –0.04 0.13

Job security 0.03 0.04 0.02 0.447 –0.05 0.12

Casual employment –0.12 0.05 –0.06 0.027 –0.22 –0.01

Self‑employed 0.01 0.05 0.01 0.793 –0.09 0.11

Job control 0.00 0.05 0.00 0.950 –0.10 0.09

Flexible hours –0.07 0.05 –0.04 0.127 –0.16 0.02

Household equivalised income 0.00 0.00 –0.04 0.149 0.00 0.00

Age of youngest child 0.11 0.17 0.03 0.524 –0.22 0.43

Family type (single or couple) –0.17 0.10 –0.04 0.076 –0.36 0.02

Years of education (mother) 0.01 0.01 0.04 0.076 0.00 0.03

Chronic medical condition (mother) –0.02 0.04 –0.01 0.570 –0.11 0.06

Age (mother) 0.00 0.00 0.02 0.354 0.00 0.01

Age (child) –0.10 0.04 –0.07 0.004 –0.18 –0.03

Sex (child) 0.00 0.02 0.00 0.963 –0.03 0.03

Regular child care 0.02 0.04 0.01 0.662 –0.06 0.10

On paid/unpaid leave 0.14 0.06 0.07 0.018 0.02 0.25

Constant 2.23 0.23 0.000 1.77 2.68

R2 0.16

Adjusted R2 0.02

Table F3: Child behavioural and emotional problems (SDQ) and mothers’ job conditions—LSAC preschooler cohort

B SE B p 95% CI for B

Work hours (part/full time) 0.71 0.25 0.06 0.004 0.23 1.19

Job security –0.60 0.24 –0.05 0.013 –1.08 –0.13

Casual employment 0.28 0.28 0.02 0.331 –0.28 0.83

Self‑employed –0.30 0.26 –0.03 0.253 –0.82 0.22

Job control –0.48 0.27 –0.04 0.080 –1.01 0.06

Flexible hours –0.11 0.28 –0.01 0.691 –0.65 0.43

Household equivalised income 0.00 0.00 –0.09 0.000 0.00 0.00

Age of youngest child –0.11 0.07 –0.04 0.100 –0.24 0.02

Family type (single or couple) –0.49 0.38 –0.03 0.197 –1.22 0.25

Years of education (mother) –0.17 0.04 –0.09 0.000 –0.26 –0.08

Chronic medical condition (mother) 0.72 0.24 0.06 0.002 0.25 1.19

Age (mother) –0.08 0.02 –0.08 0.000 –0.13 –0.04

Age (child) –1.29 0.20 –0.13 0.000 –1.69 –0.89

Sex (child) –0.01 0.04 0.00 0.822 –0.09 0.07

Regular child care –0.35 0.21 –0.04 0.099 –0.76 0.07

Constant 18.59 2.49 0.000 13.70 23.48

R2 0.28

Adjusted R2 0.08

Page 129: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

117

APPENdIx F

Fathers

Table F4: Infant approachability (STSI) and fathers’ job conditions—LSAC infant cohort

B SE B p 95% CI for B

Work hours

Full time (35–45 hrs/wk) Reference group

Long hours (45–55 hrs/wk) 0.06 0.04 0.03 0.117 –0.01 0.12

Very long (55+ hrs/wk) 0.06 0.04 0.03 0.131 –0.02 0.14

Job security –0.02 0.04 –0.01 0.525 –0.10 0.05

Casual employment 0.14 0.07 0.04 0.063 –0.01 0.29

Self‑employed 0.06 0.04 0.03 0.095 –0.01 0.14

Job control 0.02 0.04 0.01 0.567 –0.06 0.11

Flexible hours 0.09 0.04 0.04 0.022 0.01 0.17

Household equivalised income 0.00 0.00 0.04 0.040 0.00 0.00

Age of youngest child 0.23 0.12 0.07 0.050 0.00 0.46

Years of education (father) –0.01 0.01 –0.03 0.121 –0.02 0.00

Chronic medical condition (father) –0.01 0.04 0.00 0.864 –0.08 0.07

Age (father) –0.01 0.00 –0.06 0.001 –0.01 0.00

Age (child) –0.14 0.03 –0.08 0.000 –0.20 –0.08

Sex (child) –0.08 0.01 –0.25 0.000 –0.11 –0.06

Regular child care –0.07 0.03 –0.04 0.021 –0.14 –0.01

Constant 5.91 0.16 0.000 5.59 6.23

R2 0.23

Adjusted R2 0.05

Page 130: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

118 Social Policy research Paper No. 44

Table F5: Infant irritability (STSI) and fathers’ job conditions—LSAC infant cohort

B SE B p 95% CI for B

Work hours

Full time (35–45 hrs/wk) Reference group

Long hours (45–55 hrs/wk) –0.07 0.04 –0.04 0.069 –0.14 0.01

Very long (55+ hrs/wk) –0.10 0.04 –0.05 0.016 –0.18 –0.02

Job security 0.06 0.04 0.03 0.120 –0.02 0.14

Casual employment –0.06 0.08 –0.02 0.409 –0.21 0.09

Self‑employed 0.03 0.04 0.01 0.454 –0.05 0.11

Job control –0.03 0.04 –0.01 0.495 –0.12 0.06

Flexible hours 0.02 0.04 0.01 0.682 –0.06 0.10

Household equivalised income 0.00 0.00 –0.05 0.022 0.00 0.00

Age of youngest child 0.00 0.12 0.00 0.988 –0.24 0.23

Years of education (father) 0.02 0.01 0.06 0.005 0.01 0.03

Chronic medical condition (father) 0.09 0.04 0.04 0.020 0.01 0.17

Age (father) 0.00 0.00 0.02 0.204 0.00 0.01

Age (child) –0.05 0.03 –0.03 0.122 –0.11 0.01

Sex (child) 0.01 0.01 0.02 0.574 –0.02 0.03

Regular child care 0.05 0.03 0.03 0.144 –0.02 0.11

Constant 2.08 0.17 0.000 1.75 2.40

R2 0.11

Adjusted R2 0.01

Page 131: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

119

APPENdIx F

Table F6: Child behavioural and emotional problems (SDQ) and fathers’ job conditions—LSAC preschooler cohort

B SE B p 95% CI for B

Work hours

Full time (35–45 hrs/wk) Reference group

Long hours (45–55 hrs/wk) 0.03 0.22 0.00 0.877 –0.40 0.46

Very long (55+ hrs/wk) 0.31 0.24 0.03 0.207 –0.17 0.78

Job security 0.07 0.23 0.01 0.751 –0.38 0.53

Casual employment –0.22 0.51 –0.01 0.666 –1.22 0.78

Self‑employed –0.96 0.23 –0.09 0.000 –1.40 –0.52

Job control –0.26 0.27 –0.02 0.327 –0.79 0.26

Flexible hours –0.48 0.25 –0.04 0.053 –0.97 0.01

Household equivalised income 0.00 0.00 –0.13 0.000 0.00 0.00

Age of youngest child –0.07 0.06 –0.03 0.182 –0.18 0.04

Years of education (father) –0.20 0.04 –0.10 0.000 –0.28 –0.12

Chronic medical condition (father) 0.45 0.22 0.04 0.041 0.02 0.89

Age (father) –0.02 0.02 –0.02 0.291 –0.05 0.02

Age (child) –1.36 0.18 –0.14 0.000 –1.72 –1.00

Sex (child) –0.08 0.04 –0.04 0.019 –0.15 –0.01

Regular child care –0.41 0.19 –0.04 0.030 –0.78 –0.04

Constant 22.14 2.22 0.000 17.79 26.48

R2 0.28

Adjusted R2 0.07

Page 132: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

120 Social Policy research Paper No. 44

Page 133: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

121

APPENdIx g

Appendix G: Parent wellbeing and child wellbeing: full regression modelsMothers

Table G1: Infant approachability (STSI) and mothers’ psychological distress (K6)—LSAC infant cohort

B SE B p 95% CI for B

Psychological distress –0.12 0.04 –0.07 0.001 –0.19 –0.05

Household equivalised income 0.00 0.00 0.02 0.512 0.00 0.00

Age of youngest child 0.07 0.17 0.02 0.670 –0.26 0.40

Family type (single or couple) –0.02 0.10 0.00 0.845 –0.21 0.17

Years of education (mother) –0.01 0.01 –0.02 0.514 –0.02 0.01

Chronic medical condition (mother) 0.05 0.04 0.03 0.230 –0.03 0.14

Age (mother) –0.01 0.00 –0.06 0.012 –0.02 0.00

Age (child) –0.10 0.04 –0.06 0.008 –0.17 –0.03

Sex (child) –0.06 0.02 –0.20 0.000 –0.10 –0.03

Regular child care –0.04 0.04 –0.02 0.356 –0.11 0.04

Constant 5.95 0.21 0.000 5.54 6.37

R2 0.22

Adjusted R2 0.05

Table G2: Infant irritability (STSI) and mothers’ psychological distress (K6)—LSAC infant cohort

B SE B p 95% CI for B

Psychological distress 0.27 0.03 0.18 0.000 0.20 0.34

Household equivalised income 0.00 0.00 –0.01 0.728 0.00 0.00

Age of youngest child 0.04 0.16 0.01 0.800 –0.28 0.36

Family type (single or couple) –0.12 0.09 –0.03 0.190 –0.31 0.06

Years of education (mother) 0.02 0.01 0.05 0.060 0.00 0.03

Chronic medical condition (mother) –0.06 0.04 –0.03 0.152 –0.15 0.02

Age (mother) 0.00 0.00 0.02 0.301 0.00 0.01

Age (child) –0.09 0.04 –0.06 0.008 –0.16 –0.03

Sex (child) 0.00 0.02 –0.01 0.894 –0.03 0.03

Regular child care 0.07 0.04 0.04 0.071 –0.01 0.14

Constant 2.13 0.20 0.000 1.73 2.53

R2 0.21

Adjusted R2 0.05

Page 134: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

122 Social Policy research Paper No. 44

Table G3: Child emotional and behavioural issues (SDQ) and mothers’ psychological distress (K6)—LSAC preschooler cohort

B SE B p 95% CI for B

Psychological distress 2.23 0.18 0.26 0.000 1.88 2.59

Household equivalised income 0.00 0.00 –0.07 0.002 0.00 0.00

Age of youngest child –0.10 0.06 –0.03 0.124 –0.22 0.03

Family type (single or couple) –0.37 0.36 –0.02 0.305 –1.08 0.34

Years of education (mother) –0.19 0.04 –0.10 0.000 –0.27 –0.10

Chronic medical condition (mother) 0.32 0.23 0.03 0.171 –0.14 0.77

Age (mother) –0.08 0.02 –0.08 0.000 –0.12 –0.03

Age (child) –1.20 0.20 –0.12 0.000 –1.58 –0.81

Sex (child) –0.02 0.04 –0.01 0.657 –0.09 0.06

Regular child care –0.48 0.20 –0.05 0.016 –0.87 –0.09

Constant 17.49 2.37 0.000 12.85 22.14

R2 0.37

Adjusted R2 0.13

Fathers

Table G4: Infant approachability (STSI) and fathers’ psychological distress (K6)—LSAC infant cohort

B SE B p 95% CI for B

Psychological distress –0.04 0.03 –0.03 0.132 –0.10 0.01

Household equivalised income 0.00 0.00 0.04 0.029 0.00 0.00

Age of youngest child 0.23 0.12 0.07 0.055 0.00 0.46

Years of education (father) –0.01 0.01 –0.03 0.156 –0.02 0.00

Chronic medical condition (father) 0.00 0.04 0.00 0.965 –0.08 0.07

Age (father) –0.01 0.00 –0.06 0.002 –0.01 0.00

Age (child) –0.14 0.03 –0.09 0.000 –0.20 –0.08

Sex (child) –0.08 0.01 –0.25 0.000 –0.10 –0.06

Regular child care –0.08 0.03 –0.05 0.018 –0.14 –0.01

Constant 6.03 0.16 0.000 5.73 6.34

R2 0.23

Adjusted R2 0.05

Page 135: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

123

APPENdIx g

Table G5: Infant irritability (STSI) and fathers’ psychological distress (K6)—LSAC infant cohort

B SE B p 95% CI for B

Psychological distress 0.12 0.03 0.08 0.000 0.07 0.18

Household equivalised income 0.00 0.00 –0.05 0.012 0.00 0.00

Age of youngest child 0.00 0.12 0.00 0.998 –0.24 0.24

Years of education (father) 0.02 0.01 0.05 0.010 0.00 0.03

Chronic medical condition (father) 0.08 0.04 0.04 0.047 0.00 0.15

Age (father) 0.00 0.00 0.02 0.201 0.00 0.01

Age (child) –0.04 0.03 –0.03 0.147 –0.10 0.02

Sex (child) 0.01 0.01 0.02 0.581 –0.02 0.03

Regular child care 0.05 0.03 0.03 0.146 –0.02 0.11

Constant 2.04 0.16 0.000 1.73 2.35

R2 0.12

Adjusted R2 0.02

Table G6: Child emotional and behavioural issues (SDQ) and fathers’ psychological distress (K6)—LSAC preschooler cohort

B SE B p 95% CI for B

Psychological distress 2.23 0.18 0.26 0.000 1.88 2.59

Household equivalised income 0.00 0.00 –0.07 0.002 0.00 0.00

Age of youngest child –0.10 0.06 –0.03 0.124 –0.22 0.03

Years of education (father) –0.37 0.36 –0.02 0.305 –1.08 0.34

Chronic medical condition (father) –0.19 0.04 –0.10 0.000 –0.27 –0.10

Age (father) 0.32 0.23 0.03 0.171 –0.14 0.77

Age (child) –0.08 0.02 –0.08 0.000 –0.12 –0.03

Sex (child) –1.20 0.20 –0.12 0.000 –1.58 –0.81

Regular child care –0.02 0.04 –0.01 0.657 –0.09 0.06

Constant –0.48 0.20 –0.05 0.016 –0.87 –0.09

R2 0.27

Adjusted R2 0.08

Page 136: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

124 Social Policy research Paper No. 44

Page 137: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

125

APPENdIx h

Appendix H: Parent and infant wellbeing by socioeconomic position: summary regression modelsTable H1: Association between infants’ approachability scores and employed mothers’ work hours and conditions

by socioeconomic position: adjusted means (standard errors) and interactions

Socioeconomic position

Low Med/high

Work hours Part time(SE)

4.83(0.06)

4.80(0.02)

Full time(SE)

4.58(0.10)

4.75(0.04)

Interaction F(1,1853)=2.661, ns

Job security No(SE)

4.74(0.09)

4.80(0.04)

Yes(SE)

4.78(0.06)

4.78(0.02)

Interaction F(1,1852)=0.29, ns

Self‑employed No(SE)

4.76(0.06)

4.76(0.02)

Yes(SE)

4.77(0.11)

4.86(0.05)

Interaction F(1,1853)=0.51, ns

Job control No(SE)

4.91(0.10)

4.69(0.05)

Yes(SE)

4.73(0.06)

4.81(0.02)

Interaction F(1,1852)=5.64, ns

Flexible hours No(SE)

4.83(0.11)

4.70(0.05)

Yes(SE)

4.75(0.06)

4.81(0.02)

Interaction F(1,2852)=2.18, ns

Notes: Analyses are within cohort. All analyses controlled for: mothers on leave from employment; mothers having a chronic medical condition; mothers’ age; number of children in household; age of study child; family type; mothers’ work hours per week; age of youngest child in family; mothers’ self‑employment; sex of study child; regular use of child care. There were no significant main effects for socioeconomic position or work hours and conditions in any analysis.

ns=not significant; SE=standard error.

Page 138: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

126 Social Policy research Paper No. 44

Table H2: Association between infants’ irritability and employed mothers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions

Socioeconomic position

Low Med/high

Work hours Part time(SE)

2.36(0.06)

2.46†

(0.02)

Full time(SE)

2.65(0.10)

2.50(0.04)

Interaction F(1,1853)=4.67, ns

Job security No(SE)

2.41(0.09)

2.45(0.04)

Yes(SE)

2.43(0.06)

2.48(0.02)

Interaction F (1,1852)=0.01, ns

Self‑employed No(SE)

2.42(0.05)

2.47(0.02)

Yes(SE)

2.44(0.11)

2.50(0.05)

Interaction F(1,1853)=0.02, ns

Job control No(SE)

2.37(0.10)

2.50(0.05)

Yes(SE)

2.44(0.05)

2.47(0.02)

Interaction F(1,1852)=0.70, ns

Flexible hours No(SE)

2.40(0.10)

2.54(0.04)

Yes(SE)

2.43(0.05)

2.46(0.02)

Interaction F(1,1852)=1.01, ns

Notes: Analyses are within cohort. All analyses controlled for: mothers on leave from employment; mothers having a chronic medical condition; mothers’ age; number of children in household; age of study child; family type; mothers’ work hours per week; age of youngest child in family; mothers’ self‑employment; sex of study child; regular use of child care. The main effect of working arrangement was not significant in any analysis. †p<0.05; daggers indicate significance level of the working arrangement main effect.

ns=not significant; SE=standard error.

Page 139: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

127

APPENdIx h

Table H3: Association between infants’ approachability and employed fathers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions

Socioeconomic position

Low Med/high

Long hours (45–55 hrs/week) No(SE)

4.70(0.05)

4.78(0.02)

Yes(SE)

4.68(0.08)

4.86(0.04)

Interaction F(1,1818)=0.96, ns

Very long hours (55+ hrs/week) No(SE)

4.67(0.04)

4.78(0.02)

Yes(SE)

4.70(0.09)

4.86(0.04)

Interaction F(1,1818)=0.28, ns

Job security No(SE)

4.70(0.07)

4.80(0.04)

Yes(SE)

4.69(0.04)

4.78(0.02)

Interaction F(1,2819)=0.01, ns

Self‑employed No(SE)

4.67(0.04)

4.77(0.02)

Yes(SE)

4.750.08)

4.84(0.04)

Interaction F(1,2820)=0.01, ns

Job control No(SE)

4.67(0.08)

4.74(0.04)

Yes(SE)

4.69(0.04)

4.79(0.02)

Interaction F(1,2819)=0.08, ns

Flexible hours No(SE)

4.50(0.07)

4.76*†

(0.04)

Yes(SE)

4.75(0.04)

4.79(0.02)

Interaction F(1,2819)=6.47, ns

Notes: Analyses are within cohort. All analyses controlled for: fathers having a chronic medical condition; fathers’ age; number of children in household; age of study child; fathers’ work hours per week; age of youngest child in family; fathers’ self‑employment; sex of study child; regular use of child care. *p<0.01; asterisks indicate significance level socioeconomic position main effect. †p<0.01; daggers indicate significance level of the working arrangement main effect.

ns=not significant; SE=standard error.

Page 140: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

128 Social Policy research Paper No. 44

Table H4: Association between infants’ irritability and employed fathers’ work hours and conditions by socioeconomic position: adjusted means (standard errors) and interactions

Socioeconomic position

Low Med/high

Long hours (45–55 hrs/week) No(SE)

2.49(0.05)

2.48(0.02)

Yes(SE)

2.39(0.08)

2.41(0.04)

Interaction F(1,2818)=0.13, ns

Very long hours (55+ hrs/week) No(SE)

2.43(0.04)

2.52(0.02)

Yes(SE)

2.45(0.09)

2.37(0.04)

Interaction F(1,2818)=2.70, ns

Job security No(SE)

2.50(0.08)

2.42(0.04)

Yes(SE)

2.43(0.04)

2.52(0.02)

Interaction F(1,2819)=2.75, ns

Self‑employed No(SE)

2.44(0.04)

2.49(0.02)

Yes(SE)

2.46(0.08)

2.53(0.04)

Interaction F(1,2820)=0.70, ns

Job control No(SE)

2.58(0.08)

2.48(0.05)

Yes(SE)

2.41(0.04)

2.50(0.02)

Interaction F(1,2819)=4.150, ns

Flexible hours No(SE)

2.56(0.07)

2.44(0.04)

Yes(SE)

2.40(0.04)

2.51(0.02)

Interaction F(1,2819)=6.51, ns

Notes: Analyses are within cohort. All analyses controlled for: fathers having a chronic medical condition; fathers’ age; number of children in household; age of study child; fathers’ work hours per week; age of youngest child in family; fathers’ self‑employment; sex of study child; regular use of child care. The main effect of socioeconomic position was not significant in any analysis. ns=not significant; SE=standard error.

Page 141: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

129

APPENdIx I

Appendix I: Parent and infant/child wellbeing by family type: summary regression modelsTable I1: Association between infants’ approachability and employed mothers’ work hours and conditions by

family type: adjusted means (standard errors) and interactions

Family type

Lone Couple

Work hours Part time(SE)

4.72(0.11)

4.81(0.02)

Full time(SE)

5.01(0.21)

4.72(0.04)

Interaction F(1,1852)=2.51, ns

Job security No(SE)

4.72(0.17)

4.80(0.04)

Yes(SE)

4.81(0.11)

4.78(0.02)

Interaction F(1,1851)=0.31, ns

Self‑employed No(SE)

4.78(0.10)

4.76(0.02)

Yes(SE)

4.57(0.36)

4.85(0.04)

Interaction F(1,1852)=0.69, ns

Job control No(SE)

4.67(0.18)

4.73(0.04)

Yes(SE)

4.81(0.11)

4.79(0.02)

Interaction F(1,1851)=0.14, ns

Flexible hours No(SE)

4.93(0.19)

4.71(0.04)

Yes(SE)

4.73(0.11)

4.80(0.02)

Interaction F(1,1851)=1.68, ns

Notes: Analyses are within cohort. All analyses controlled for: equivalised income; mothers’ years of education; mothers on leave from employment; mothers having a chronic medical condition; mothers’ age; number of children in household; age of study child; mothers’ work hours per week; age of youngest child in family; mothers’ self‑employment; sex of study child; regular use of child care.

ns=not significant; SE=standard error.

Page 142: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

130 Social Policy research Paper No. 44

Table I2: Association between infants’ irritability and employed mothers’ work hours and conditions by family type: adjusted means (standard errors) and interactions

Family type

Lone Couple

Work hours Part time(SE)

2.61(0.11)

2.44(0.02)

Full time(SE)

2.77(0.20)

2.51(0.04)

Interaction F(1,1852)=0.16, ns

Job security No(SE)

2.39(0.17)

2.44(0.04)

Yes(SE)

2.76(0.11)

2.46(0.02)

Interaction F(1,1851)=3.00, ns

Self‑employed No(SE)

2.61(0.10)

2.45(0.02)

Yes(SE)

3.07(0.35)

2.48(0.04)

Interaction F(1,1852)=1.45, ns

Job control No(SE)

2.36(0.18)

2.49(0.04)

Yes(SE)

2.75(0.11)

2.45(0.02)

Interaction F(1,1851)=3.85, ns

Flexible hours No(SE)

2.40(0.19)

2.52(0.04)

Yes(SE)

2.73(0.11)

2.44(0.02)

Interaction F(1,1851)=3.65, ns

Notes: Analyses are within cohort. All analyses controlled for: equivalised income; mothers’ years of education; mothers on leave from employment; mothers having a chronic medical condition; mothers’ age; number of children in household; age of study child; mothers’ work hours per week; age of youngest child in family; mothers’ self‑employment; sex of study child; regular use of child care.

ns=not significant; SE=standard error.

Page 143: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

131

APPENdIx I

Table I3: Association between preschoolers’ SDQ (total) scores and employed mothers’ work hours and conditions by family type: adjusted means (standard errors) and interactions

Family type

Lone Couple

Work hours Part time(SE)

8.48(0.42)

8.27(0.12)

Full time(SE)

9.59(0.62)

8.84(0.22)

Interaction F(1,2069)=0.50, ns

Job security No(SE)

9.20(0.58)

8.93(0.23)

Yes(SE)

8.64(0.43)

8.27(0.12)

Interaction F(1,2068)=0.02, ns

Self‑employed No(SE)

8.90(0.37)

8.49(0.12)

Yes(SE)

8.35(1.07)

8.17(0.22)

Interaction F(1,2068)=0.04, ns

Job control No(SE)

9.10(0.70)

8.88(0.26)

Yes(SE)

8.74(0.40)

8.31(0.12)

Interaction F(1,2069)=0.06, ns

Flexible hours No(SE)

9.70(0.79)

8.56(0.26)

Yes(SE)

8.61(0.39)

8.38(0.12)

Interaction F(1,2068)=1.02, ns

Notes: Analyses are within cohort. All analyses controlled for: equivalised income; mothers’ years of education; mothers on leave from employment; mothers having a chronic medical condition; mothers’ age; number of children in household; age of study child; mothers’ work hours per week; age of youngest child in family; mothers’ self‑employment; sex of study child; regular use of child care.

ns=not significant; SDQ=Strengths and Difficulties Questionnaire; SE=standard error.

Page 144: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

132 Social Policy research Paper No. 44

Page 145: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

133

ENdNOTES

Endnotes

1. We also compared our results with analyses that adjusted for design and clustering effects in LSAC and analyses using weighted variables. An example of these analyses is presented at Appendix B and they are discussed in Section 2. Briefly, they indicate that except for already marginal findings (which we interpret cautiously) adjusting for design and clustering effects and using weighted variables does not change any substantive findings.

2. We also compared our results with analyses that adjusted for design and clustering effects in LSAC and analyses using weighted variables. An example of these analyses is presented in Appendix B. Except where noted, adjusting for design and clustering effects and using weighted variables does not change any substantive findings.

Page 146: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

134 Social Policy research Paper No. 44

Page 147: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

135

rEFErENCES

References

Almeida, D, Wethington, E & Chandler, A 1999, ‘Daily transmission of tensions between marital dyads and parent‑child dyads’, Journal of Marriage and the Family, vol. 61, pp. 49–61.

Australian Bureau of Statistics (ABS) 1985, Labour force historical summary: 1966–1984, ABS, Canberra.

——1997, National health survey: SF‑36 population norms, Australia, cat. no. 4399.0, ABS, Canberra.

——1999, Special article—casual employment, cat. no. 6203.0, ABS, Canberra.

——2000, Labour force status and other characteristics of families, cat. no. 6224.0, ABS, Canberra.

——2001a, Australian social trends 2001, cat. no. 4102.0, ABS, Canberra.

——2001b, Family characteristics survey, Census 2001, cat. no. 4442.0, ABS, Canberra.

——2001c, National Health Survey: mental health, cat. no. 4811.0, ABS, Canberra.

——2002, Australian social trends 2002, cat. no. 4102.0, ABS, Canberra.

——2003a, Australian social trends 2003, cat. no. 4102.0, ABS, Canberra.

——2003b, Family characteristics survey, Census 2003, cat. no. 4442.0, ABS, Canberra.

——2004, Australian social trends 2004, cat. no. 4102.0, ABS, Canberra.

——2005, Labour statistics: concepts, sources and methods, cat. no. 6102.0.55.001, ABS, Canberra.

——2006, Australian social trends 2006, cat. no. 4102.0, ABS, Canberra.

——2007, Australian social trends 2007, cat. no. 4102.0, ABS, Canberra.

Australian Institute of Family Studies (AIFS) 2006, Longitudinal Study of Australian Children: 2005–06 Annual Report, AIFS, Melbourne.

——2007, Longitudinal Study of Australian Children data user guide: Version 3.0, AIFS, Melbourne.

Australian Institute of Health and Welfare (AIHW) 2005, A picture of Australia’s children, AIHW cat. no. PHE 58, AIHW, Canberra.

Australian Centre for Industrial Relations Research and Training (ACIRRT) 2002, Examination of work and family measures in Queensland and Federal enterprise agreements (Adam Special Report), Prentice Hall, Sydney.

Baird, M 2003, ‘Paid maternity leave: the good, the bad, the ugly’, Australian Journal of Labour, vol. 29, no. 1, pp. 97–109.

Barling, J, Zacharatos, A & Hepburn, CG 1999, ‘Parents’ job insecurity affect children’s academic performance through cognitive difficulties’, Journal of Applied Psychology, vol. 84, no. 3, pp. 437–44.

Baxter, J, Gray, M, Alexander, M, Strazdins, L & Bittman, M 2007, Mothers and fathers with young children: paid employment, caring and wellbeing, Social Policy Research Paper no. 30, Department of Families, Community Services and Indigenous Affairs, Canberra.

Bernstein, J 2004, ‘The low‑wage labor market: trends and policy implications’, in AC Crouter & A Booth (eds), Work–family challenges for low‑income parents and their children, Lawrence Erlbaum Associates, Mahwah, New Jersey, pp. 3–34.

Page 148: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

136 Social Policy research Paper No. 44

Bittman, M & Craig L 2005, The effect of children on adults’ time‑use: an analysis of incremental time costs of children in Australia, SPRC Discussion Paper No. 143, Social Policy Research Centre, UNSW, Sydney.

Bittman, M, Hoffmann, S & Thompson, D 2004, Men’s uptake of family‑friendly employment provisions, Policy Research Paper no. 22, Department of Family and Community Services, Canberra.

Blakemore, T, Strazdins, L & Gibbings, J 2009, ‘Measuring family socioeconomic position’, Australian Social Policy, no. 8, pp. 121–68.

Bond, JT, Thompson, C, Galinsky, E & Prottas, D 2002, Highlights of the National Study of the Changing Workforce (No. 3), Families and Work Institute, New York.

Bradley, RH & Corwyn, RF 2002, ‘Socioeconomic status and child development’, Annual Review of Psychology, vol. 53, pp. 371–99.

Bray, RB 2001, Hardship in Australia: an analysis of financial stress indicators in the 1989–99 Australian Bureau of Statistics household expenditure survey, Occasional Paper no. 4, Department of Family and Community Services, Canberra.

Bronfenbrenner, U 1979, The ecology of human development, Harvard University Press, Cambridge, MA.

Broom, DH, D’Souza, RM, Strazdins, L, Butterworth, P, Parslow, R & Rodgers, B 2006, ‘The lesser evil: bad jobs or unemployment? A survey of mid‑aged Australians’, Social Science and Medicine, vol. 63, no. 3, pp. 575–86.

Cass, B 2002, ‘Employment time and family time: the intersections of labour market transformations and family responsibilities in Australia’, in R Callus & R Lansbury (eds), Working futures: the changing nature of work and employment relations in Australia, Federation Press, Leichhardt.

Centre for Mental Health Research (CMHR), 2005, PATH Through Life Project: PATH 40+, CMHR, Canberra.

Chen, E, Matthews, K & Boyce, WT 2002, ‘Socioeconomic differences in children’s health: how and why do these relationships change with age?’, Psychological Bulletin, vol. 128, no. 2, pp. 295–329.

Clarke‑Stewart, KA 1978, ‘And daddy makes three: the father’s impact on mother and young child’, Child Development, vol. 49, pp. 466–78.

Cole, D, Ibrahim, S, Shannon, HS, Scott, FJ & Eyles, J 2002, ‘Work and life stressors and psychological distress in the Canadian working population: a structural equation modelling approach to analysis of the 1994 National Population Health Survey’, Chronic Diseases in Canada, vol. 23, no. 3, pp. 91–99.

Connell, AM & Goodman, SH 2002, ‘The association between psychopathology in fathers versus mothers and children’s internalizing and externalizing behaviour problems’, Psychological Bulletin, vol. 128, no. 5,pp. 746–73.

Cooksey, EC, Menaghan, EG & Jekielek, SM 1997, ‘Life‑course effects of work and family circumstances on children’, Social Forces, vol. 76, no. 2, pp. 637–65.

Coyne, JC, Thompson, R & Palmer, SC 2002, ‘Marital quality, coping with conflict, marital complaints and affection in couples with a depressed wife’, Journal of Family Psychology, vol. 16, no. 1, pp. 26–37.

Craig, L 2002, The time cost of parenthood: an analysis of daily workload, SPRC Discussion Paper No. 117, Social Policy Research Centre, University of New South Wales, Sydney.

——2004, Time to care: a comparison of how couple and sole parent households allocate time to work and children, Social Policy Research Centre, University of New South Wales, Sydney.

——2007, ‘How employed mothers in Australia find time for both market work and childcare’, Journal of Family Economic Issues, vol. 28, pp. 69–87.

Page 149: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

137

rEFErENCES

Crosier, T, Butterworth, P & Rodgers, B 2007, ‘Mental health problems among single and partnered mothers’, Social Psychiatry and Psychiatric Epidemiology, vol. 42, pp. 6–13.

Cummings, EM 1994, ‘Marital conflict and children’s functioning’, Social Development, vol. 3, no. 1, pp. 16–36.

D’Souza, RM, Strazdins, L, Lim, LL‑Y, Broom, DH & Rodgers, B 2003, ‘Work and health in a contemporary society: demands, control, and insecurity’, Journal of Epidemiology and Community Health, vol. 57, no. 11,pp. 849–54.

De Witte, H & Näswell, K 2003, ‘Objective’ vs ‘subjective’ job insecurity: consequences of temporary work for job satisfaction and organizational commitment in four European countries’, Economic and Industrial Democracy, vol. 24, no. 2, pp. 149–88.

Dex, S 2003, Families and work in the twenty‑first century, Joseph Roundtree Foundation, York, UK.

Dickens, L 1999, ‘Beyond the business case: a three‑pronged approach to equality action’, Human Resource Management Journal, vol. 9, no. 1, pp. 9–19.

Downey, G & Coyne, JC 1990, ‘Children of depressed parents: an integrative review’, Psychological Bulletin,vol. 108, no. 1, pp. 50–76.

Drago, R, Tseng, Y‑P & Wooden, M 2004, Family structure, usual and preferred working hours, and Egalitarianism in Australia, Working Paper No. 1/04, Melbourne Institute of Applied Economic and Social Research, Melbourne.

Druss, BG, Rosenheck, RA & Sledge, WH 2000, ‘Health and disability costs of depressive illness in a major US corporation’, The American Journal of Psychiatry, vol. 157, no. 8, pp. 1274–87.

Dwyer, SB, Nicholson, JM & Battistutta, D 2003, ‘Population level assessment of the family risk factors related to the onset or persistence of children’s mental health problems’, Journal of Child Psychology and Psychiatry, vol. 44, no. 5, pp. 699–711.

Earle, J 2002, ‘Family‑friendly workplaces: a tale of two sectors’, Family Matters, vol. 61, no. 3, pp. 12–17.

Edwards, ME 2001, ‘Uncertainty and the rise of the work–family dilemma’, Journal of Marriage and the Family, vol. 1, no. 63, pp. 183–96.

Elder, GH 1985, ‘Perspectives on life course’, in GH Elder (ed.), Life course dynamics: trajectories and transitions, 1968–1980, Cornell University Press, New York.

——1995, ‘The life course paradigm: social change and individual development’, in P Moen, GH Elder & K Luscher (eds), Examining lives in context: perspectives on the ecology of human development, American Psychological Association, Washington, pp. 101–40.

Fagan, C 2001, ‘The temporal reorganization of employment and the household rhythm of work schedules’, American Behavioural Scientist, vol. 44, no. 7, pp. 1199–212.

Fenwick, R & Tausig, M 2001, ‘Scheduling stress: family and health outcomes of shift work and schedule control’, American Behavioural Scientist, vol. 44, no. 7, pp. 1179–98.

Ferrie, JE 1999, ‘Health consequences of job insecurity’, in JE Ferrie, MG Marmot, J Griffiths & E Ziglio (eds), Labour market changes and job insecurity, 81 edn, pp. 59–99, World Health Organization, Geneva.

Ferrie, JE, Shipley, MJ, Marmot, MG, Martikainen, P, Stansfeld, SA & Smith, GD 2001, ‘Job insecurity in white collar workers: toward an explanation of associations with health’, Journal of Occupational Health Psychology, vol. 6, no. 1, pp. 26–42.

Fisher, J, Rowe, H & Feekery, C 2004, ‘Temperament and behaviour of infants aged 4–12 months on admission to a private mother‑baby unit and at 1‑ and 6‑month follow‑up’, Clinical Psychologist, vol. 8, no. 1, pp. 15–21.

Page 150: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

138 Social Policy research Paper No. 44

Galinsky, E 1999, Ask the children, William Morrow, New York.

Glass, J 2000, ‘Envisioning the integration of family and work: toward a kinder, gentler, workplace’, Contemporary Sociology, vol. 29, no. 1, pp. 129–43.

Goode, A & Watson, N (eds) 2007, HILDA User Manual—Release 5.0, Melbourne Institute of Applied Economic and Social Research, Melbourne.

Goodman, R 1997, ‘The Strengths and Difficulties Questionnaire: a research note’, Journal of Child Psychology and Psychiatry, vol. 38, no. 5, pp. 581–86.

Gottfried, AE & Gottfried, A 2006, ‘A long‑term investigation of the role of maternal and dual‑earner employment in children’s development: the Fullerton Longitudinal Study’, American Behavioral Scientist,vol. 49, no. 10, pp. 1310–27.

Goward, P, Mihailuk, T, Moyle, S, O’Connell, K, de Silva, N, Squire, S & Tilley, J 2005, Striking the balance: women, men, work and family, Discussion paper, Human Rights and Equal Opportunity Commission, Sydney.

Gray, H 2002, Family‑friendly working: what a performance! An analysis of the relationship between the availability of family‑friendly policies and establishment performance, Discussion paper, Centre for Economic Performance, London School of Economics and Political Science, London.

Gray, M, Qu, L, Stanton, D & Weston, R 2004, ‘Long work hours and the wellbeing of fathers and their families’, Australian Journal of Labour Economics, vol. 7, no. 2, pp. 255–73.

Gray, M & Tudball, J 2002, ‘Access to family‑friendly work practices: differences within and between Australian workplaces’, Family Matters, vol. 61, pp. 30–35.

Green, F 2006, Demanding work: the paradox of job quality in the affluent economy, Princeton University Press, Princeton, New Jersey.

Greenberg, PE, Kessler, RC, Nells, TL, Finkelstein, SN & Berndt, FR 1996, ‘Depression in the workplace: an economic perspective’, in JP Feighner & WF Boyer (eds), Selective serotonin uptake inhibitors: advances in basic research and clinical practice, John Wiley and Sons, New York, pp. 327–63.

Greenfield, E & Marks, N 2006, ‘Linked lives: adult children’s problems and their parents’ psychological and relational well‑being’, Journal of Marriage and Family, vol. 68, no. 2, pp. 442–54.

Grimm‑Thomas, K & Perry‑Jenkins, M 1994, ‘All in a day’s work: job experiences, self esteem, and fathering in working‑class families’, Family Relations, vol. 43 (April), pp. 174–81.

Grzywacz, JG, Almeida, D & McDonald, D 2002, ‘Work‑family spillover and daily reports of work and family stress in the adult labour force’, Family Relations, vol. 51, pp. 28–36.

Grzywacz, JG & Dooley, D 2003, ‘“Good jobs” to “bad jobs”: replicated evidence of an employment continuum from two large surveys’, Social Science and Medicine, vol. 56, pp. 1749–60.

Hancock, K 2002, ‘Work in an ungolden age’, in R Callus & RD Lansbury (eds), Working futures: the changing nature of work and employment relations in Australia, The Federation Press, Sydney, pp. 6–26.

Hawes, DJ & Dadds, MR 2004, ‘Australian data and psychometric properties of the Strengths and Difficulties Questionnaire’, Australian and New Zealand Journal of Psychiatry, vol. 38, no. 8, pp. 644–51.

Hesselink, DJK & van Vuuren, T 1999, ‘Job flexibility and job insecurity: the Dutch case’, European Journal of Work and Organizational Psychology, vol. 8, no. 2, pp. 273–93.

Heymann, SJ & Earle, A 2001, ‘The impact of parental working conditions on school‑age children: the case of evening work’, Community, Work and Family, vol. 4, no. 3, pp. 305–25.

Page 151: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

139

rEFErENCES

Higgins, C, Duxbury, L & Lee, C 1994, ‘Impact of life‑cycle stage and gender on the ability to balance work and family responsibilities’, Family Relations, vol. 43, no. 2, pp. 144–50.

Houtman, I 2005, Work‑related stress, European Foundation for Improvement of Living and Working Conditions, Ireland.

Jacobs, J & Gerson, K 2004, The time divide, Harvard University Press, Cambridge, MA.

Jones, F & Fletcher, B 1996, ‘Taking work home: a study of daily mood fluctuations in work stressors, effects on mood and impacts on marital partners’, Journal of Occupational and Organizational Psychology, vol. 69,pp. 89–106.

Jones, P & Dickerson, A 2007, Poor returns: winners and losers in the job market, Warwick Institute for Employment Research, Manchester.

Keating, DP & Hertzmann, C 1999, Developmental health and the wealth of nations, Guildford Press, New York.

Kessler, RC, Barker, PR, Colpe, LJ, Epstein, JF, Gforerer, JC, Hiripi, E, Howes, MJ, Normand, SLT, Manderscheid, RW, Walters, EE & Zaslavsky, AM 2003, ‘Screening for serious mental illness in the general population’, Archives of General Psychiatry, vol. 60, no. 2, pp. 184–89.

Kinnunen, U & Mauno, S 1998, ‘Antecedents and outcomes of work‑family conflict among employed women and men in Finland’, Human Relations, vol. 51, no. 2, pp. 157–77.

Kivimaki, M, Vahtera, J, Pentt, J & Ferrie, J 2000, ‘Factors underlying the effect of organisational downsizing on health of employee: longitudinal cohort study’, British Medical Journal, vol. 320, no. 7240, pp. 971–75.

La Valle, I, Arthur, S, Millward, C, Scott, J & Clayden, M 2002, Happy families? Atypical work and its influence on family life, Policy Press, Bristol, UK.

Larson, JH, Wilson, SM & Beley, R 1994, ‘The impact of job insecurity on marital and family relationships’, Family Relations, vol. 43, pp. 138–43.

Little, RJA & Su, HL 1989, ‘Item non‑response in panel surveys’, in D Kasprzyk, GL Duncan & MP Singh (eds), Panel Surveys, John Wiley and Sons, New York.

Lovejoy, MC, Graczyk, PA, O’Hare, E & Neuman, G 2000, ‘Maternal depression and parenting behaviour: a meta‑analytic review’, Clinical Psychology Review, vol. 20, no. 5, pp. 561–92.

MacEwen, KE & Barling, J 1994, ‘Daily consequences of work interference with family and family interference with work’, Work and Stress, vol. 8, no. 3, pp. 244–54.

Marmot, M, Siegrist, J, Theorell, T & Feeney, A 1999, ‘Health and the psychosocial environment at work’, in MM & RG Wilkinson (eds), Social determinants of health, Oxford University Press, Oxford, pp 105–31.

Marmot, MM & Wilkinson, RG (eds) 1999, Social determinants of health, Oxford University Press.

Marshall, K 1993, ‘Dual earners: who’s responsible for housework?’, Canadian Social Trends, Winter, pp. 11–14.

Masterman‑Smith, H, May, R & Pocock, B 2006, Living low paid: some experiences of Australian child care workers and cleaners, CWL Discussion Paper No. 1/06, Centre for Work and Life, University of South Australia, Adelaide.

Mathers, CD & Loncar, D 2006, ‘Projections of global mortality and burden of disease from 2002 to 2030’, PLoS Med, vol. 3, no. 11, e442, doi:10.1371/journal.pmed.0030442.

Meiers, M 1988, ‘Parental leave and the bottom line’, Personnel Journal, vol. 67, no. 9, pp. 108–13.

Menaghan, EG 1991, ‘Work experiences and family interaction processes: the long reach of the job?’, Annual Review of Sociology, vol. 17, pp. 419–44.

Page 152: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

140 Social Policy research Paper No. 44

Merllié, D & Paoli, P 2001, Ten years of working conditions in the European Union, European Foundation for the Improvement of Living and Working Conditions, Dublin.

Mistry, RS, Vandewater, EA, Huston, AC & McLoyd, VC 2002, ‘Economic well‑being and children’s social adjustment: the role of family process in an ethnically diverse low‑income sample’, Child Development, vol. 73, no. 3, pp. 935–51.

Murphy, J 2002, ‘Breadwinning: accounts of work and family life in the 1950s’, Labour and Industry, vol. 12,no. 3, pp. 59–75.

Newnham, EA, Harwood, KE & Page, AC 2007, ‘Evaluating the clinical significance of responses by psychiatric inpatients to the mental health subscales of the SF‑36’, Journal of Affective Disorders, vol. 98, no. 1–2,pp. 91–97.

Organization of Economic Cooperation and Development (OECD) 2002, Babies and bosses—reconciling work and family life, vol. 1, Australia, Denmark, The Netherlands, OECD, Paris.

——2004, OECD employment outlook 2004, OECD, Paris.

——2005, ‘LMF7: Usual working hours per week by gender (aged 15–64)’, OECD Family database,viewed 3 March 2008, <www.oecd.org/els/social/family/database>.

——2007, OECD employment outlook 2007, OECD, Paris.

Phares, V & Compass, BE 1992, ‘The role of fathers in child and adolescent psychopathology: make room for daddy’, Psychological Bulletin, vol. 111, no. 3, pp. 387–412.

Phares, V, Duhig, A & Watkins, M 2002, ‘Family context: fathers and other supports’, in SH Goodman & IH Gotlib (eds), Children of depressed parents: mechanisms of risk and implications for treatment, American Psychological Association, Washington DC.

Propper, C, Rigg, J & Burgess, S 2007, ‘Child health: evidence on the roles of family income and maternal mental health from a UK birth cohort’, Health Economics, vol. 16, no. 11, pp. 1245–69.

Quiggin, J 1999, ‘Globalisation, neoliberalism and inequality in Australia’, The Economic and Labour Relations Review, vol. 10, no. 2, pp. 240–59.

Repetti, RL 1989, ‘Effects of daily workload on subsequent behaviour during marital interaction: the roles of social withdrawal and spouse support’, Journal of Personality and Social Psychology, vol. 37, no. 4,pp. 651–59.

——1994, ‘Short‑term and long‑term processes linking job stressors to father–child interaction’, Social Development, vol. 3, no. 1, pp. 1–15.

Repetti, R, Taylor, S & Seeman, T 2002, ‘Risky families: family social environments and the mental and physical health of offspring’, Psychological Bulletin, vol. 128, no. 2, pp. 330–66.

Rodgers, B 1996, ‘Reported parental behaviour and adult affective symptoms. 1. Associations and moderating factors’, Psychological Medicine, vol. 26, no. 1, pp. 51–61.

Rosenman, S & Rodgers, B 2006, ‘Childhood adversity and adult personality’, Australian and New Zealand Journal of Psychiatry, vol. 40, no. 5, pp. 482–90.

Roxburgh, S, 2004, ‘There just aren’t enough hours in the day’: the mental health consequences of time pressure’, Journal of Health and Social Behaviour, vol. 45, no. 6, pp. 115–31.

Saltzman, A & Barry, P 1988, ‘The superdad juggling act: trying to find a way to rise on the job and be a good father too’, U.S. News and World Report, vol. 104, pp. 67–70.

Page 153: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

141

rEFErENCES

Sanders, MR, Gooley, S & Nicholson, J 2000, ‘Early intervention in conduct problems in children’, in R Kosky, A O’Hanlon, G Martin & C Davis (eds), Clinical approaches to early intervention in child and adolescent mental health, vol. 3, Australian Early Intervention Network for Mental Health in Young People, Adelaide.

Sanderson, K, Tilse, E, Nicholson, JM & Oldenburg, B 2005, Impact of work conditions on the mental health and wellbeing of call centre employees: report for participating organisations, Queensland University of Technology.

Sanson, A, Prior, M, Oberklaid, F & Northam, E 1987, ‘Measurement of temperament in 1 to 3 year old children’, International Journal of Behavioral Development, vol. 10, pp. 121–32.

Sanson, A, Prior, M, Garino, E, Oberklaid, F & Sewell, J 1987, ‘The structure of infant temperament: factor analysis of the revised infant temperament questionnaire’, Infant Behavior and Development, vol. 10, no. 1,pp. 97–104.

Sayer, L, Bianchi, S & Robinson, J 2004, ‘Are parents investing less in children? Trends in mothers’ and fathers’ time with children’, The American Journal of Sociology, vol. 110, no. 1, p. 1.

Schoon, I, Sacker, A & Bartley, M 2003, ‘Socio‑economic adversity and psychosocial adjustment: a developmental‑contextual perspective’, Social Science and Medicine, vol. 57, pp. 1001–15.

Senate Community Affairs References Committee (SCARC) 2004, A hand up not a hand out: renewing the fight against poverty, SCARC, Commonwealth of Australia, Canberra.

Soloff, C, Lawrence, D & Johnston, R 2005, Sample design, LSAC Technical Paper no. 1, Australian Institute of Family Studies, Melbourne.

Soloff, C, Lawrence, D, Misson, S & Johnson, R 2006, Wave 1 weighting and non‑response, LSAC Technical Paper no. 3, Australian Institute of Family Studies, Melbourne.

Stansfeld, SA, North, FM & Marmot, MG 1995, ‘Work characteristics and psychiatric disorder in civil servants in London’, Journal of Epidemiology and Community Health, vol. 49, pp. 48–53.

Statistics Canada 1995, National Longitudinal Survey of Children and Youth (NLSCY) Wave 1 survey instruments: parent questionnaire, Ottawa, Canada.

Strazdins, L, Clements, M, Korda, RJ, Broom, DH & D’Souza, RM 2006, ‘Unsociable work? Nonstandard work schedules, family relationships and children’s well‑being’, Journal of Marriage and the Family, vol. 68, no. 2, pp. 394–410.

Strazdins, L, D’Souza, RM, Lim, LL‑Y, Broom, DH & Rodgers, B 2004, ‘Job strain, job insecurity and health: rethinking the relationship’, Journal of Occupational and Health Psychology, vol. 9, no. 4, pp. 296–321.

Turrell, G, Oldenburg, B, McGuffog, I & Dent, R 1999, Socioeconomic status and health: towards a national research program and a policy and intervention agenda, Commonwealth Department of Health and Aged Care, Canberra.

US Department of Labor 1999, Futurework: trends and challenges for work in the 21st century, Author, Office of the Secretary, Washington, DC.

UNICEF (United Nations International Children’s Emergency Fund) 2007, Child poverty in perspective: an overview of child well‑being in rich countries, Report Card 7, UNICEF Innocenti Research Centre, Florence.

van der Hulst, M 2003, ‘Long work hours and health’, Scandinavian Journal of Work, Environment and Health, vol. 29, no. 3, pp. 171–88.

Ware, JE, Snow, KK, Kosinski, M & Gandek, B 2000, SF‑36 Mental Health Survey: Manual and Interpretation Guide, QualityMetric Inc., Lincoln, RI.

Page 154: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

PArENT ANd ChIld WEllbEINg ANd ThE INFluENCE OF WOrk ANd FAMIly ArrANgEMENTS

142 Social Policy research Paper No. 44

Watson, I, Buchanan, J, Campbell, I & Briggs, C 2003, Fragmented futures: new challenges in working life, The Federation Press, Sydney.

Watson, N & Wooden, M 2004, HILDA Project Technical Paper Series no. 1/04: Wave 2 Survey Methodology, Melbourne Institute of Applied Economic and Social Research, Melbourne.

Weiler, A 2005, Annual review of working conditions in the EU: 2004–2005, European Foundation for the Improvement of Living and Working Conditions, Dublin.

Whitehouse, G 2001, ‘Industrial agreements and work/family provisions: trends and prospects under enterprise bargaining’, Labour and Industry, vol. 12, no. 1, pp. 109–30.

WHO (World Health Organization) 2007, ‘Mental health: strengthening mental health promotion’, viewed 2 March 2008, <www.who.int/mediacentre/factsheets/fs220/en>.

Wooden, M 2002, ‘The changing labour market and its impact on work and employment relations’, in R Callus & RD Lansbury (eds), Working futures: the changing nature of work and employment relations in Australia, The Federation Press, Sydney, pp. 51–69.

Zubrick, SR, Williams, AA, Silburn, SR & Vimpani, G 2000, Indicators of social and family functioning, Department of Family and Community Services, Canberra.

Page 155: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

Social Policy Research Papers1. The Australian system of social protection—an overview

Peter Whiteford (2000)

2. Parents, the labour force and social security Karen Wilson, Jocelyn Pech and Kylee Bates (1999)

3. Estimates of the costs of children in Australian families, 1993–94 National Centre for Social and Economic Modelling (1999)

4. Social policy directions across the OECD region: reflections on a decadeDavid W Kalisch (2000)

5. Structural ageing, labour market adjustment and the tax/transfer systemDavid Ingles (2000)

6. Trends in the incomes and living standards of older people in AustraliaPeter Whiteford and Kim Bond (2000)

7. Updating Australian budget standards costs of children estimatesPaul Henman and Macquarie University (2001)

8. Social indicators for regional AustraliaJ Rob Bray (2001)

9. Means‑tested benefits, incentives and earnings distributions John Creedy and Rosanna Scutella (2001)

10. The duration of unemployment benefit spells: a comparison of Indigenous and non‑Indigenous persons Thorsten Stromback and Mike Dockery (2001)

11. A meta‑analysis of the impact of community‑based prevention and early intervention action Erin Gauntlett, Richard Hugman, Peter Kenyon and Pauline Logan (2001)

12. How do income support recipients engage with the labour market?Paul Flatau and Mike Dockery (2001)

13. The policy‑maker’s guide to population ageing: key concepts and issuesNatalie Jackson (2001)

14. The dynamics of participating in Parenting Payment (Single) and the Sole Parent Pension Garry Barrett (2001)

15. Jobs in a new labour market: changes in type and distributionAlan Jordan (2001)

16. Cost–benefit analysis of portability policy Kruno Kukoc and Norbert Zmijewski (2001)

17. Some issues in home ownership William Mudd, Habtemariam Tesfaghiorghis and J Rob Bray (2001)

18. The impact of social policy initiatives on labour supply incentives: a review of the literature Guyonne Kalb (2003)

19. Patterns of economic and social participation among FaCS customersPeter Saunders, Judith Brown and Tony Eardley (2003)

Page 156: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

20. Child poverty: a review Bruce Bradbury (2003)

21. Estimating the prevalence of mental disorders among income support recipients: approach, validity and findings Peter Butterworth (2003)

22. Men’s uptake of family‑friendly employment provisionsMichael Bittman, Sonia Hoffman and Denise Thompson (2004)

23. Household monies and decision‑making Saba Waseem (2004)

24. Understanding and improving data quality relating to low‑income householdsDavid Johnson and Rosanna Scutella (2005)

25. Effects of child care demands and policies on household labour supply in Australia Denise Doiron and Guyonne Kalb (2005)

26. Communities, social capital and public policy: literature reviewDavid Johnson, Bruce Headey and Ben Jensen (2005)

27. The causes of changes in the distribution of family income in Australia, 1982 to 1997–98 David Johnson and Roger Wilkins (2006)

28. Exploring the economic and social value of present patterns of volunteering in Australia Michael Bittman and Kimberly Fisher (2006)

29. Income poverty, subjective poverty and financial stressGary N Marks (2007)

30. Mothers and fathers with young children: paid employment, caring and wellbeingJennifer Baxter, Matthew Gray, Michael Alexander, Lyndall Strazdins and Michael Bittman (2007)

31. Intergenerational reliance on income support: psychosocial factors and their measurement Helen Berry, Emma George, Peter Butterworth, Bryan Rodgers and Tanya Caldwell (2007)

32. Contemporary Australian archetypes: different people, different needs Helen Berry, Peter Butterworth, Tanya Caldwell and Bryan Rodgers (2008)

33. The structure and distribution of household wealth in Australia: cohort differences and retirement issues Bruce Headey, Diana Warren and Mark Wooden (2008)

34. Parenting and families in AustraliaStephen R Zubrick, Grant J Smith, Jan M Nicholson, Ann V Sanson, Tanyana A Jackiewicz and the LSAC Research Consortium (2008)

35. Marriage breakdown in Australia: social correlates, gender and initiator statusBelinda Hewitt (2008)

36. How well are Australian infants and children aged 4 to 5 years doing? Melissa Wake, Ann Sanson, Donna Berthelsen, Pollyanna Hardy, Sebastian Misson, Katherine Smith, Judy Ungerer and the LSAC Research Consortium (2008)

37. Factors associated with relations dissolution of Australian families with children Peter Butterworth, Tamar Oz, Bryan Rodgers and Helen Berry (2008)

38. Young carers in Australia: understanding the advantage and disadvantages of their care giving Bettina Cass, Ciara Smyth, Trish Hill, Megan Blaxland and Myra Hamilton (2009)

Page 157: Parent and child wellbeing and the influence of work and family ... · Improving the lives of Australians Social Policy Research Paper No. 44 Parent and child wellbeing and the influence

39. Longing to belong: personal social capital and psychological distress in an Australian coastal region Helen L Berry and Megan Shipley (2010)

40. Child care and early education in AustraliaLinda J Harrison, Judy A Ungerer, Grant J Smith, Stephen R Zubrick and Sarah Wise with Frances Press, Manjula Waniganayake and the LSAC Research Consortium (2010)

41. Asset rich, but income poor: Australian housing wealth and retirement in an international contextBruce Bradbury (2010)

42. Parental divorce and adult family, social and psychological outcomes: the contribution of childhood adversityBryan Rodgers, Patricia Gray, Tanya Davidson and Peter Butterworth (2011)

43. The costs of caring and the living standards of carers Trish Hill, Cathy Thomson and Bettina Cass (2011)


Recommended