When we look at associations between physical health and household income (Figure 1), there is an increase in the mean PCS score for women earning less than $500 per week as their partner's income increases. Above this income, there is no significant change in the mean PCS across the partner's income range.

In the first category shown in Figure 1, for example, where women are earning less than $300 a week, their health improves with each increase in their partner’s income. This trend continues until what appears to be a potentially self-sufficient income is reached, of $500 per week, when increases in partner’s income seem to make little difference. The implication is that a woman’s good health is linked to her ability to earn a reasonable income in her own right.

Similar observations can be drawn from the results for mental health (Figure 2).

Again, in looking at income and health gradients, as shown in Figure 2, it appears that mental health improves as household income improves. A further analysis using CESD, a standard measure of depression, was also carried out and the results followed the same trend with income group as the Mental Health scale.

In regard to the strength of association between health and income, there are several points which can be made. Firstly, in relation to associations between income and physical health, links between poverty and poor health have been well documented and are linked to health behaviours, and factors such as nutritional status and quality of housing, as well as the accessing and availability of health services.

Secondly, particularly in relation to mental health, there are at least two mechanisms which may be implicated. Firstly, the ‘control’ factor identified in the Whitehall study of British civil servants (Marmot et al, 1991) and more recently in the Boston Nurses’ Study (Cheng et al, 2000) may be relevant. Autonomy at work, which increases as one ascends the occupational status ladder, is associated with feelings of control over one’s environment, and this sense of mastery has positive effects such as improvements in confidence and self esteem. There are likely to be similar effects associated with increasing income, which affirms one’s performance in the job. Secondly, the greater a woman’s income, the greater her bargaining power in domestic decision making and the general distribution of household labour (Arber and Ginn, 1995; Rogers, 1999; Rosenfield, 1992), and the more likely it is that domestic services will be purchased and that the family will eat out.

It has been shown that if the sharing of housework is perceived to be unfair it is likely to lead to feelings of anxiety and decreased self-worth. It is therefore of interest to look at WHA respondents’ perceptions of the fairness of the distribution of domestic work in their households and associations with their income and health.

American research into links between depression and the amount and distribution of household responsibilities, shows that it is the not the amount of housework carried out but the distribution of that work between partners, which is relevant to levels of depression (Bird, 1999). Women in the United States reported doing 68 per cent of the housework and this finding is paralleled in Michael Bittman’s time use research in Australia (Bittman, 1995).

The distribution of housework is also characterised by a persistence in inequity. For example, in the 1980s men increased the amount of time they spent cooking, but this trend did not continue into the 90s. Janeen Baxter (1998) has also found a persistent acceptance by Australian women of this inequitable distribution.

In the baseline questionnaire for the mid age WHA cohort in 1996, respondents were asked how they felt about their share of various activities, including household work. The response options were as follows: happy as it is; would like other family members to do more; prefer another arrangement; not applicable.

Figure 3 shows women’s satisfaction with the distribution of housework in the household according to their income. For the purposes of this particular analysis, weekly income was categorised as follows: less than $300; $300-499; $500-699; $700-999; more than $1,000.

It seems from the results shown in Figure 3 that women in lower income categories are more likely to accept that the housework is their responsibility, while women in the middle income groups are more likely to feel that other family members are not pulling their weight. Women earning this level of income ($500-699 per week) are likely to be in full time employment, possibly doing routinised work, or alternatively they may be holding down multiple part-time jobs. It may be assumed that women in this group are shouldering the double burden, without a great deal of control in the sharing of housework. Having such significant family responsibilities yet little power has been found to be associated with anxiety and depression.

Figure 4 shows associations between mental health score and women’s responses to the question about domestic work. Those who were happy with their share had a higher mental health score than women who said they would prefer other family members to do more.

Figure 5 shows that women who are happy with their share of the housework are also likely to report better physical health.

The data in Figure 6 (below) are for the group of women who are happy with their share of housework, across all income groups. This figure shows the effects of concordance in earnings, i.e., in each income group, the majority of women who are happy with the distribution of domestic work have partners earning a similar amount. Among women who are earning less than $300 per week and who are happy with their share of housework, the greatest proportion are married to men earning a similar amount. Similarly, women earning between $300 and $499 per week who are happy with the distribution of housework have partners with a similar income, and so on.

Figure 6 (above) shows that, consistent with results for the other income groups, a majority of women earning more than $1000 per week and who are happy with their share of housework also have partners on a high income. Although not shown here, our data also indicate that high earning women with low income partners tend to feel that the share of housework in their household is not fair. While unfairness in the distribution of household work is likely to reduce feelings of control over one’s life and therefore increase psychological distress, it has also been shown that there are ‘spillovers’ from work to home. People who feel they lack control in the workplace may attempt to dominate their spouse in order to feel that they have control in at least one area of their life. This is likely to exacerbate the feelings of powerlessness experienced by those who carry the greatest share of responsibility for household work.

It also appears that in lower income couples, income disparity is related to spousal abuse. American research has shown that in such couples, women whose income is higher than their husband’s are more likely to be subject to domestic violence (McCloskey, 1996). This finding may be of some relevance as a contributory factor to lower self reported health scores among some women in relationships where there are differentials in partners’ incomes.

Given that it is not the total amount of time spent on housework but the relative distribution of time spent on housework between partners that is relevant to mental health status, it is not surprisingly that equitable sharing of marital and work roles is associated with happier marriages which in turn are associated with better mental health for women (Rogers, 1999).

Our analysis demonstrates that women’s self assessed health is positively associated not only with their own income, but also with that of their partner. For women on relatively low incomes, it appears that there is an improvement in both physical and mental health with increases in the amount of income contributed by their partner. However, once a woman begins to earn about $500 or more per week, this association disappears, suggesting that there are factors associated with earning a potentially self-sufficient income which are associated with better health. We also find that women who are happiest with the sharing of housework are likely to report better mental and physical health, and that these perceptions are associated with the income of the woman and of her partner. Women are likely to be more satisfied with the distribution of domestic labour in partnerships where there is equality of earnings.

However, with current patterns of social organisation, employment remains more liberating for men in general, because they earn more, do less housework and contribute a greater proportion of the household income. This points to the need to redress both continuing inequalities in earnings and occupational gender segregation, so that women are not constrained to the secondary labour market and can begin to achieve rewards for their work which are comparable with those of men.

While considerable progress has been made in the developed world in terms of equality of employment opportunity, equal pay for comparable work has proved more elusive and substantial earnings inequality between women and men remains. Income inequality is related to decision making power within the household, which in turn affects the distribution of domestic labour. Within the household, gender inequalities in earnings are greater than inequality of occupational level (Arber and Ginn, 1995) and this is likely to continue, if not increase, in Australia, given the trend to enterprise bargaining agreements which are inimical to gender equalities in pay.

Not only is this trend likely to jeopardise improvements in women’s health, but, given the importance of women’s income to the household, it will have wider repercussions in perpetuating the socioeconomic gap, with decreases in the income of poorer households having implications for the health of all members of those households.

Currently, there is some evidence that women are moving out of the workforce, due to neoliberal social policies that attempt to reward women for staying at home. It is more likely to be working class women who are reproducing traditional gendered household patterns while middle class women combine paid work and parenting (Sharp and Broomhill, 1999). In the light of these developments, it is imperative that social policies which facilitate education and labour force attachment for women be strengthened. Equally, those which perpetuate conservative gender ideologies and constrain women’s options, should be revised.

Future directions

This work shows that there is a need for more ‘fine-grained’ research in this area. These results need to be pulled apart to look at some of the specific groups identified here, and to determine the relative weights of factors which we have addressed in a general way. For example, we have examined women’s perceptions of the fairness of the distribution of domestic work in their household, but we have not yet related these perceptions to the amount of time that our respondents spent in paid work.

This is cross sectional research, but the longitudinal nature of the WHA project will enable us to look at these issues over time. In particular we will be interested in following the young cohort as they now enter the peak years of family formation.

Finally, in this study we have discussed addressed women’s health and not men’s. Interactive associations between partner’s income, and the health of Australian men requires investigation.

The Commonwealth Department of Health and Aged Care funds the WHA project, which was conceived and developed by groups of interdisciplinary researchers at the Universities of Newcastle and Queensland. The contribution of members of the WHA team, including investigators and support staff, is gratefully acknowledged.