The Effect of Changes in Maternal Employment and Family Composition on Children's Behaviour - May 2002

Revisiting the research problems and discussion from the literature review, we see now that the relationship between children's behaviour, maternal employment and family composition, is not uncomplicated. The evidence in Tables 2 to 7 shows that over the period in question (1994 to 1998) it was common for children's mothers to be employed and to be so for a majority of the year. However, for many mothers it is also quite common to experience periods of unemployment interspersed with periods of employment, or not to have experienced any employment at all. Similarly, most of the children were in stable family situations over the period from 1994 to 1998 (Tables 8 and 9). However, a moderate proportion of children did experience changes in the family composition during the same period.

Analysis of the cross-sectional relationship between maternal employment, family composition and children's behavioural scale scores yields some evidence to suggest that two biological parent family structures are associated with significantly lower behavioural scale scores in both 1994 and 1998 (Table 11). However, there are important exceptions to this. Children in lone parent mother families, for example, had significantly lower conduct disorder-physical aggression scale scores in 1998 compared to children in two biological parent families (see Table 11).

The evidence from 1994 and 1998 shows that maternal employment variables in 1994 tended to be more strongly related to children's behavioural scale scores, while family composition in 1998 appeared to be much more important for affecting children's behaviour scale scores (Table 11). In 1994 many of the associations between maternal employment and behaviour scale scores were significant, but had disappeared in 1998. For example, in 1994 there were nine significant associations between various measures of maternal employment and the three behavioural outcomes, but only one relationship was statistically significant in 1998. Similarly, many family composition associations that were significant in 1998 were not significant in 1994. This occurred quite often in the cross-sectional analysis. Many relationships did remain significant after applying control variables, but in some cases relationships became non-significant after the control variables were applied.

What the cross-sectional evidence does not provide is an answer as to whether changes in family composition or in maternal employment over time are associated with changes in children's behavioural scale scores. Social stressors, including changes and transitions in employment and family composition, as mentioned previously, have been implicated in affecting children's behaviour. They do so by reducing parenting time and also increasing the amount of stress on parents as they seek employment. We find in the cross-sectional results some evidence to support the contention that social stressors adversely affect children's behaviour. However, it is not possible to know what happened to children's mothers or families over the period from 1994 to 1998 using the data in Table 11. Instead it is necessary to observe how changes in maternal employment and family composition are associated with children's behaviour.

Evidence of change in maternal employment and family composition was observed in Table 13. It was found that children of mothers who were employed compared to children of mothers who were not employed, seemed to have significantly higher indirect aggression scores and in some circumstances higher conduct disorder-physical aggression, and emotional disorder-anxiety. Being a currently employed mother may be an important factor affecting children's behavioural outcomes, but the evidence also suggests that any experience of employment transition may itself affect children's behavioural outcomes. That is, many of the mothers that were not employed in 1998 had still experienced some employment transition over the period from 1994 to 1998. This might explain why we observed that mothers who were not employed in 1998 but had "some" to "high" employment transition experiences, still had significantly worse conduct disorder-physical aggression scores than mothers not employed and with no transition experience. Such evidence provides support for the view that social stress from less stable maternal employment, in the form of higher levels of transitions in employment (particularly if it is more frequent unemployment) can affect children's propensity for poorer behavioural outcomes. However, mother's occupational prestige (as measured by the Pineo scale) can attenuate some of the poor behavioural outcomes for children's emotional disorder-anxiety scores, since children of mothers with occupations rated higher on the Pineo scale tended to have lower emotional disorder-anxiety scale scores.

Changes in family composition have also been identified by Menaghan, Mott and Cooksey (1997) as increasing the social stress of families, which can affect children's behavioural outcomes. Compared to children living in stable two biological parent families, children living in biological mother and step-father families had relatively lower scores for each of the three behavioural scales, while those in stable lone parent mother families had relatively lower scores for emotional disorder-anxiety and indirect aggression. In contrast, children living in families which entered lone parent mother situations had higher scores on each of the three behavioural scales. Where the relationship was significantly related with our behavioural outcomes, children living in families that experienced changes in composition over the period 1994 to 1998 tended to have significantly poorer behavioural outcomes than children in stable family compositional situations.

The employment results observed here differ somewhat from those found by Menaghan, Mott and Cooksey (1997). Maternal employment characteristics in their study were associated with lower behavioural scores for children's oppositional action. Family composition results in this analysis, though, are much more similar to those observed in their study in that the specific form of stable family composition or change would seem to be important when looking at children's behavioural scores.

Despite observing quite large changes in maternal employment patterns and the high proportion of mothers who experienced a number of spells of unemployment in our descriptive data, the effect of these on children's behavioural outcomes from 1994 to 1998 as measured in the regression analysis is statistically significant but substantively relatively weak. Similarly, the literature on family composition and change indicates the deleterious effect for children's behaviour of family changes. The results here support this notion when observing selected information from the cross-sectional regression findings, and are particularly supportive when observing the longitudinal analysis (Table 13). Nevertheless, these results are still relatively weak when attempting to connect family composition, and changes in the composition, with poorer child behavioural outcomes.

The children as well as their families in the research cohort were being exposed to a number of experiences from 1994 to 1998, such as new school experiences and challenges, an improving economy and many other social processes. As detailed in the analysis above, these affected many children in terms of changing family composition and maternal employment. What effect did all of these changes have, particularly with respect to family composition and maternal employment, on children's behaviour? Results show that there does not seem to be any one condition of employment or family circumstance that can confer absolutely better behavioural outcomes for children. Nonetheless, some final comments can be put forward.

The results of the analysis here cannot provide any perfect avenue for bettering the behavioural development for middle childhood children and their families. Certainly we see that many of the variables considered here affect children's indirect aggression, most particularly the maternal employment variables. The fact that mothers are employed (whether they experience any transitions or not) seems most important for children's indirect aggression. However, the level of maternal employment transitions (that is, high levels) seems important for affecting children's emotional disorder-anxiety, while the mother's experience of some employment transitions (whether they are currently employed or not) seems most important for affecting children's conduct disorder-physical aggression behaviour. Stable family situations can aid children in receiving relatively better behavioural variable scores when compared to children in families experiencing a change in composition. In particular, children in families experiencing an entry into a lone parent mother situations seem prone to worse behavioural outcomes.

Cross-sectional and multivariate results indicate that children in families with high maternal labour market transitions have relatively poorer behavioural development. That is, children in these situations have a relatively higher propensity for child behavioural problems. At a policy level, therefore, it would seem prudent to investigate families with high maternal transition rates into and out of the labour force, since stability seems to confer benefits for children's behavioural development. Also, evidence suggests that the type of occupation and socioeconomic prestige of the occupation (as seen from the Pineo scale) of the mother can act to decrease behavioural problems, and should also be investigated. Another potential avenue of investigation is to research family compositional change. This might include looking at broader legal policies involving areas where laws may be affecting family's circumstances including divorce and custody arrangements. It is also necessary to pursue research that would investigate children who are prone to particular behavioural problems. The research presented here observed that the behavioural effects may differ by child's gender as well as the health of the children involved. A propensity for behavioural problems in children is also associated with particular characteristics of mothers, including mother's age and level of education. We have seen in particular that indirect aggression and conduct disorder-physical aggression are affected by these characteristics.

If we want to be able to aid in providing children with a good start to life it makes sense to investigate how they are affected by changes in the early part of their lives. Changes in children's lives including those attached to maternal employment and new family compositions have been identified by researchers as possible social stressors for children and their parents. These changes can affect the relationship and time parents have for children with a consequent effect on the children's behavioural well-being and development. Continuing to track and understand the effect of these changes on children's development and the longer term consequences for their later life is important.