Endnotes
1. See Table B1 for a summary of changes in employment patterns between 1984 and 2004.
2. A more detailed analysis of changes in family labour supply is provided by Renda (2003) and analysis of the changes in partnered mothers' employment by Baxter (2005c).
3. Examples include Breusch and Gray (2004); Brusentsev (2002); Gray et al. (2003); Kalb (2004); and Scutella (2001).
4. Almost all interviews were conducted between March and November 2004.
5. Children were selected from a random selection of 293 postcodes. The selection of children within postcodes was also random. A process of stratification was used to ensure that the numbers of children in each State and Territory and within and outside each capital city were roughly proportionate to the total numbers of children in these areas. The sampling strategy was deliberately designed to reduce the age differences of the child at the interview date and so families with children born between September and February of the relevant years were selected for interview slightly later than those born between March and August. Refer to Soloff, Lawrence and Johnstone (2005) for details of the sample selection process.
6. In order to investigate the process that may have led to the male parent being nominated as the primary carer in couple-parent families, the labour force status of these men was examined. Interestingly, 76.3 per cent of male primary carers were employed (23.2 per cent part-time, 53.1 per cent full-time). Of those who were not employed, 78.9 per cent had an employed partner. There is likely to be a range of reasons for nominating the father as the primary carer, including health or language barriers of the mother. Alternatively, it may simply be that it was more convenient for the father to respond to the survey and be nominated as the primary carer.
7. This response rate calculation assumes that the non-contacts would respond at the same rate as those who could be contacted by an interviewer and is probably the most reasonable assumption. Nonetheless, the fact that more geographically mobile families were more likely to be a non-contact may introduce some biases into the sample.
8. Only 53 per cent of carers based at long day care centres and 43 per cent of home-based carers (family day carers or informal carers such as grandparents) completed and returned their questionnaires. Information from teachers, long day care centres and home-based carers are not used in this report.
9. No adjustments, such as imputations, were made to account for non-response to particular items or particular components of the survey. For example, if the self-completion questionnaire was not returned, these non-responding parents were excluded from analyses that rely on items that are contained in this instrument, without any attempt to re-weight the remaining observations back to population benchmarks. For more details see LSAC User Guide (LSAC Project Operations Team 2006).
10. As the sampling frame was child based, all children from multiple births can be included. This means that for families with multiple-birth children, one child does not represent one family. Treating these child records as representative of families will therefore slightly overestimate the number of families (LSAC Project Operations Team 2006).
11. In Wave 1, detailed information is only collected about parents who live in the same household as the study child. Wave 2 will collect information from non-resident parents.
12. Only 11 of the 5,107 study children that make up the infant cohort had a younger sibling, although 165 study children were one of a multiple birth.
13. For the 48 children where the female carer was not the biological mother, they were an adoptive or foster mother, a grandmother or aunt or stepmother or had another relationship to the study child. Of the 181 cases in which the male carer was not the biological father, in 60 per cent of the cases the relationship to the study child was stepfather.
14. A small number of single parents were in a relationship with the study child's other parent, or with another partner, even though they did not live together. These families are classified as being single-parent families. Of 1,174 single-parent families in LSAC, 75.6 per cent did not have a partner, 11.5 per cent did not reside with study child's other parent but had an ongoing relationship with them and the remainder (12.9 percent) had a partner who was not the study child's father with whom they did not reside. Of the 8,916 couple-parent families, there were 147 families in which the partner was temporarily living away from home for work reasons, and 47 families in which the partner was temporarily living away from home for other reasons.
15. This refers to the definition used in the monthly Labour Force Survey collected by the ABS. For more detail refer to Appendix A.
16. Mothers on paid maternity leave at the time of the interview represented 1.0 per cent of the infant cohort, while those on unpaid maternity/parental-leave represented 8.4 per cent of this cohort.
17. The comparison of LSAC, Census and Labour Force Survey estimates is provided in Appendix A.
18. See Section 2.2 for details of the labour force status variable.
19. See Section 2.2 for a discussion of the issues involved in analysing LSAC by age of the youngest child.
20. The ABS Pregnancy and Work Transitions survey, November 2005 (cat. no. 4913.0) (ABS 2006b) provides information on maternal employment following pregnancy. The Parental Leave in Australia Survey (LSAC Wave 1.5), conducted on the infant cohort in June to September 2005, also provides extensive information on mothers' employment before and after the birth of the study child (see Whitehouse & Soloff 2005 for information) and also <http://www.uq.edu.au/polsis/parental-leave>.
21. According to the June 2004 Labour Force Survey, the rates of employment for men and women aged 20 to 44 years without dependant children were 80.3 and 81.9 per cent respectively. Men aged 20 to 44 years with a dependent child had an employment rate of 91.4 per cent, over 10 percentage points higher than for childless men. The employment rate of women with dependent children was considerably lower at 55.9 per cent.
22. Parents in permanent jobs or in long-term casual jobs are entitled to one year's unpaid leave, such that they can return to their same job when the child is 1 year old. Some women, as a condition of their employment, are also able to take some of this one year off work on paid maternity leave. Further, women can take other types of leave such as recreation leave, long service leave or leave without pay, to extend the period of paid leave. Not all women take official leave from work, with some having less formal arrangements with an employer about if and when they will return to work. The Parental Leave in Australia Survey (LSAC Wave 1.5), collected in 2005 and released in late 2006, was designed to update and extend Glezer's study and will greatly enhance our understanding in this area (see <http://www.uq.edu.au/polsis/parental-leave>)
23. The question about work during their pregnancy did not distinguish between those who worked for some of or all of their pregnancy.
24. For example, 29 per cent of couple mothers and 9 per cent of single mothers had bachelor degrees or higher, while 21 per cent of couple mothers and 40 per cent of single mothers had incomplete secondary education.
25. Estimates from ABS (2006a).
26. Jobless families are those in which no parent is employed, including those in which the parent/s were classified as being not employed because they were on maternity/parental leave.
27. Given the relatively low number of employed single mothers with infants, the age of the youngest child is not taken into account in this analysis.
28. Employed parents were asked if they ever worked after 6 p.m. or overnight and if so, how frequently. The information on frequency of working after 6 p.m. or overnight is used to identify permanent night shift workers (see Table 3.20).
29. Employed parents were asked if they ever work on Saturdays or Sundays. Those parents that answered 'yes' were defined as sometimes working weekends.
30. The description of the Australian child care system draws heavily on Harrison and Ungerer (2005, p. 26).
31. This information was collected in the face-to-face interview.
32. In some non-employed families, child care was said to be used for work or study reasons and in two-thirds of these families, the primary carer was studying. In a small number of families the primary carer was not employed but seeking work.
33. In these families, 92.4 per cent of those whose main type of care was formal care and 90.4 per cent of those whose main type of care was informal care did so for parent's work or study reasons.
34. These analyses were restricted to working families: employed single-parent families and dual-employed couple-parent families. First, a logistic regression was estimated on the likelihood of using parental care only (compared to using some form of child care). Second, a multinomial logistic regression was estimated that fitted the different child care types, and the results of this model were used to assess whether there were characteristics associated with the use of formal care only versus informal care only. For both methods, the models estimated contained the family and job characteristics shown in Table 4.4 and Table 4.5, with the full model results presented in Table C1 and Table C2.
35. From Table 4.4 and Table 4.5 the differences were not so easy to identify, but by using the multinomial logistic regression to test for the effect of variables on the odds of using informal care only versus formal care only, a few effects were revealed (see Table C2 for details).
36. To answer questions about the association between employment and child care, there would need to be more information on the employment preferences of parents (that is, to ask those who were not employed whether they would prefer to be, and why they were not), and on parents' child care preferences beyond those already collected in LSAC. Specifically, parents would need to be asked whether they preferred more care or a different type of care, and whether there were barriers to being able to access these arrangements.
37. There were some difficulties in accurately making the distinction between preschool and pre-Year 1 of school, especially taking into account State and Territory differences in the naming of different levels of early education or school. However, detailed analyses of these data suggest that most children have been classified correctly (LSAC Project Operations Team 2006).
38. A logistic regression was estimated on the likelihood of being in school or preschool only, as opposed to being in some other type of care (including those in school or preschool plus some other care). Those in parental care only were excluded from the multivariate analysis. The multivariate results are presented in Table C3.
39. The evidence suggests that maternal employment 'protects' children against the damaging effects of maternal depression and poverty. Research suggests that parental employment improves children's cognitive and social developmental, educational achievement and their likely earnings as adults (Brooks-Gunn, Leventhal & Duncan 2000; Han, Waldfogel & Brooks-Gunn 2001; Brooks-Gunn, Han & Waldfogel 2002). However, some researchers have also been concerned about the possible negative effects of maternal employment during the child's infancy on cognitive development and attachment (Belsky 2001; Brooks-Gunn, Han & Waldfogel 2002). There has been far less exploration of the mechanisms which connect child development and maternal employment. The existing literature on successful child development concentrates on characteristics of the parent–child relationship (Solchany & Barnard 2001; Shonkoff & Phillips 2000).
40. Single mothers sometimes recorded that the child was with the father, but this information was not used, since it could not be related to the father's characteristics.
41. For fathers, there is considerable overlap between their time and the mother's time with children, such that fathers are less likely to be present with the child with the mother not also there (Australian Institute of Family Studies 2005; Craig 2006).
42. The measures of time spent with children using time-use data are described in Budig and Folbre (2004) and Craig (2006). Estimates of time spent with children based on primary child care, primary plus secondary child care and time with children are given in Craig (2006).
43. A similar, but more sophisticated approach, further classifying activities according to the likely degree of parental involvement, was used on a different data source, by Bittman, Craig and Folbre (2004).
44. For a discussion of the problems related to this release of the Time Use Data, refer to the LSAC User Guide (LSAC Project Operations Team 2006). A second release of these data was issued in October 2006, which addressed some of the data quality issues. A more detailed analysis of the missing data is being undertaken by the Australian Institute of Family Studies, the findings of which will be released as an upcoming LSAC Discussion Paper.
45. The analyses were based on unweighted data, and statistical tests were not used to test for significant associations or differences between groups.
46. To be more definitive, a more detailed analysis of the contextual information is required, to determine whether fathers with employed partners spend more time with the child without their partners also present than fathers with not-employed partners.
47. Most of this increased time, however, can be accounted for by additional hours of sleep where the mother was present (see Table 5.6).
48. This discussion of wellbeing draws heavily upon Gray et al. (2004).
49. An overview of the history of the literature on material welfare and perceived poverty is provided by Saunders (2003). Bradshaw (2003) provides an excellent overview of the related concept of social exclusion.
50. Throughout this report the measure of parental income used is that collected in ranges. Wave 1 of LSAC also collected income as a continuous measure from each parent. These continuous measures were used to derive an approximation for the top range in the grouped parental income data item. This approximation was set to the median of the continuous measure parental income, from those whose combined parental income was reported to be in this top range.
51. That is, income before tax, superannuation or health insurance is deducted.
52. Of the 5,107 families in the infant cohort, 556 were excluded because of the presence of adults other than the parents, and for the 4–5 year-old cohort, 551 of the 4,983 families were excluded for the same reason. There were further exclusions because income was missing for some households (245 of the infant cohort and 274 of the 4–5 year-old cohort) or was negative (eight families).
53. See Citro and Michael (1995) for a discussion of different equivalence scales and their use, and ABS (2005a) for a discussion of the modified OECD equivalence scale and application to Australian data.
54. For the infant cohort, most families had equivalence scales of 1.8 (representing two adults and one child, 35 per cent) or 2.1 (representing two adults and two children, 34 per cent). For the 4–5 year-old cohort, the most common values of the equivalence scale were 2.1 (two adults and two children, 44 per cent) and 2.4 (two adults and three children, 24 per cent).
55. Respondents were also asked about having financial limitations on the type of food they could buy. However, the extent to which this question provides an indicator of hardship is unclear and it is therefore excluded from the analysis in this section.
56. A useful validation check of the hardship and prosperity measures is the extent to which they are correlated. The hardship score is highly correlated (correlation coefficient=-0.48, p<0.001) with perceived prosperity, with higher levels of hardship found among those who see themselves as poorer, providing some confidence in these measures for this survey.
57. This finding is consistent with those of Bray (2001) who analysed the incidence of hardship of different family types using the ABS 1998–99 Household Expenditure Survey. Bray found that the incidence of hardship and the experience of multiple hardships was higher among single-parent families than couple-parent families.
58. The analysis presented in this section was repeated excluding households with other adults, and it was found that comparisons by family type and employment status gave the same findings in terms of relative differences between the groups.
59. Recent changes to the conditions of receipt of Parenting Payment (Single) and the income testing of this benefit is likely to have had an effect on the proportion of employed single mothers in receipt of an income support payment. These families would, however, continue to be eligible for FTB Part B.
60. There were 77 couple-parent families where neither parent was employed and neither received an income support payment (as reported). In 15 of these families, a parent was on leave from work and in another 11 cases the reported gross weekly income of the couple was $1,000 or more. Of the rest, while not reporting to receive an income support payment, 34 reported receiving FTB Part A or B.
61. The measures in LSAC are a subset of those used in the HILDA survey.
62. This result is consistent with the findings of Alexander and Baxter (2005) who used LSAC data to examine the work-to-family measure and found that most differences between cohorts were accounted for by differences in family and job characteristics.
63. Both items were drawn from the Australian Temperament Study—see <http://www.aifs.gov.au/atp> and Prior et al. (2000).
64. The majority of these variables were explored in Section 3. Further tables for the latter two variables are in Appendix B.
65. Where a result is noted to be 'significant', it is at least statistically significant at the 5 per cent level.
66. For example, parents were less likely to be employed if they had a long-term medical condition (Table 3.8), which means the not employed are more likely to have a long-term medical condition than are other parents. This in part, then, could explain the poorer health outcomes of not-employed parents.
67. Wellbeing was also associated with the age of the youngest child in the family. As both cohorts were combined for this analysis, the age of the youngest child ranged from less than 1 to 5 years old. In the multivariate models, the 2 and 3 year olds were combined into a single category, as were the 4 and 5 year olds.
68. This measure is the combined parental income, collected in ranges. Refer also to endnote 50.
69. Which was more than nine out of 10 of the employed fathers in LSAC.
70. For feeling time pressured, results for fathers working part-time hours were not significantly different from those working short full-time hours; for work–family strains, fathers working part-time hours had the most favourable outcome.
71. Other models were explored, interacting hours of parent with hours of partner, but very few significant effects were found, and interpretation was complicated unjustifiably.
72. Data are extracted from the 'Labour Force Status and Other Characteristics of Families', which is based on the Labour Force Surveys as at June of each year. This is the only month in which sufficient family data are collected to enable a comparison with LSAC.
73. There are a number of other differences between Labour Force Survey and the 2001 Census, which are outlined in ABS (2002).
74. The ABS would define a person on unpaid leave as not in the labour force if they had not received any paid leave in the previous four weeks, something which is highly likely with the LSAC respondents but which is not able to be accurately confirmed.
75. For LSAC, this does not produce a representative sample (see discussion in Section 2.2) but it is useful as it enables a comparison of employment estimates between sources.