Family structure, child outcomes and environmental mediators

An overview of the Development in Diverse Families Study
Research Paper No. 30 – February 2003

You are in an archived section of the AIFS website. Archived publications may be of interest for historical reasons. Because of their age, they do not reflect current research data or AIFS' current research methodologies.

Implications of family diversity for children

Although there is scant evidence that 'traditional' family types are normative historically or cross-culturally (Lamb, Sternberg and Thompson 1999; Gilding 2001), reservations have been expressed about whether 'new' household arrangements - single-parent households, stepfamilies, blended families, extended families, gay and lesbian families - are adaptive for children.

Public attitudes toward family diversity

Some make the assumption that children can only be brought up successfully in a two-parent family structure involving a heterosexual relationship. Others take the perspective that children can function well in any family structure, provided certain basic conditions are met.

Researchers have attempted to provide an evidence base to inform this debate by examining the developmental trajectories of children growing up in various family structures. Some findings from this field of research appear to have validated concerns about the wellbeing of such children.

Particular attention has focused on the wellbeing of children growing up in single-parent families. The general conclusion from a large body of data is that children from single-parent families overall fare less well than children from intact two-parent families. Studies have shown that children in singleparent families are apt to have more health problems, poorer social and motor development and more academic problems, and higher probability of both internalising and externalising problems (Underwood and Kamein 1984; Amato and Keith 1991; Dawson 1991; Ferri 1984; Zill 1988 1994; Dunn, Deater-Decard, Pickering and O'Connor 1998: 1083; Human Resources Development Canada 1998: 5). Indeed, there is some evidence that children from single-parent homes continue to experience academic and social adjustment problems even many years after their parents' divorce (Linder, Hagan and Brown 1992).

However, it needs to be noted that most of these data are based on divorced single-parent families, or families involving teenage mothers, rather than women who make a decision to raise a child without a parenting partner. Differences between children growing up in single-parent households and intact couple families also tend to be modest. Many children in single-parent families do just as well as the average child in a two-parent family.

Australian and New Zealand studies that provide data on outcomes for children following divorce or long-term separation of parents tend to mirror findings from overseas research that have shown parental divorce to be a risk factor for a wide range of social and psychological problems in adolescence and adulthood (Rodgers 1998; Fergusson, Diamond, and Horwood 1986; Feehan, McGee, Williams, and Nada-Rada 1995).

The adjustment of children in divorced single-parent families and in stepfamilies is rather similar (Hetherington and Stanley-Hagan 2002: 290). Membership of a stepfamily is also a robust correlate of higher rates of psychopathology in children (Hetherington, Bridges and Insabella 1998; Amato and Keith 1991; Amato 1993).

In contrast to research on single-parent and stepfamilies, research to date has shown that children in same-sex couple families are no more likely to experience psychological disorder than children growing up in more traditional family types (Golombok and Tasker 1994; Green, Mandel, Hotvedt, Gray and Smth 1986; Golombok 1999). However, gay and lesbian-parent families have been little researched. More work is needed in Australia to understand the dynamics of gay and lesbian families and the developmental trajectories of their children.

A family structure effect?

Although family structure is consistently reported as contributing to children's outcomes, there are some key caveats about this body of research that deserves noting.

First, the 'effect size' (the standardised measures indicating the size of differences found between two groups) of differences between family types is generally small in magnitude. In their recent book, Jan Pryor and Bryan Rodgers (2001) summarised findings from individual studies that compared children from intact families and children who do not live with their two biological parents (that is, stepfamilies and single-parent families) in terms of the size of differences found across a broad range of outcomes areas. They concluded that 'children from separated families typically have between one-and-a-half times to double the risk of an adverse outcome compared with children from intact original families' (p. 66). The effect sizes found in Australian divorce studies are similar to those of (predominantly) United States studies (see Rodgers 1998). These effect sizes are considered small because the rates of problem outcomes among children in intact families are generally low (approximately 10 per cent of children experiencing social and emotional problems, and 8 per cent of children demonstrating aggressive or antisocial behaviour are typical findings) (Pryor and Rodgers 2001: 59-61).

The conclusion that effect sizes of differences between different family types are small concurs with findings from a meta-analysis conducted by Amato and Keith (1991), which compared divorced and non-divorced families in the United States (see also Amato 2001). When divorced and non-divorced groups were compared according to eight outcome categories, in three-quarters of cases there were no differences, in 23 per cent there was a significant difference favouring the non-divorced, and in 2 per cent there was a significant difference favouring the divorced. Where significant differences were found, effect sizes were generally small (Burns, Dunlop and Taylor 1997: 136). However, as Burns et al. (1997: 136) note, 'small and often non-significant effects should not be disregarded, because even small disadvantages are serious when they apply to a large group'. Yet, while family transitions place children at slightly increased risk of problem outcomes, 'studies of children's adjustment to divorce and remarriage have shown that relatively few children and adolescents experience enduring problems' (Hetherington and Stanley-Hagan 2002: 309).

Another issue is the striking variability within groups - that is, differences found within groups are often as strong as differences between groups. Many single parents, unmarried couples, stepparents and same-sex parents raise their children successfully, and the majority of children growing up in 'nontraditional' family forms emerge as reasonably competent and well-functioning individuals. As Richards and Schmiege (1993: 277) state in relation to singleparent families: 'For the most part, studies compare single-parent families to two-parent families, thereby ignoring the enormous variability in this increasingly common family form. This lack of attention to family diversity within single-parent families contributes to negative judgements and inhibits the targeting of services to parents and children that most need assistance.'

This then raises the question: is family structure itself the prime determinant of adverse outcomes, or do these problems flow from other sources that tend to cooccur with particular family structures? If (as most research that focuses on the roles of different environmental mediators suggest) other risk factors such as poor parental care, financial strain and parental mental health are closely linked to different types of family structures, then it becomes important to understand how these factors occur and operate in different family structures, and whether family structure has a direct effect on adverse outcomes.

The adequacy of family structure models

The absence of a parent or the loss of a parent are two frameworks for explaining differences in child outcomes among children and young people growing up in non-traditional family structures compared with children living in intact twoparent families.

Parental loss

Although loss of a parent through separation, divorce or death has been associated with child distress in the short term, it is not considered to have deleterious longterm effects (see Rodgers 1998; Pryor and Rodgers 2001). Partial support for this position is provided by evidence that depression in adulthood is just as common in those whose parents divorced while they were young as it is in those whose parents stayed together during their childhood but divorced subsequently (Rodgers 1998).

Comparisons of children who experience parental death compared with parental separation show that parental death does not have the same degree of adverse social and psychological outcomes as divorce (Amato 1993), which suggests that other factors may be more significant in explaining children's poor development. Further, several studies have shown that children suffer significant disadvantage before separation (Elliot and Richards 1991), and that some children may be better off after family change if relationships within the family of origin were hostile, conflictual or abusive (see Amato 1993; Burns 1981).

Parental absence

The absence of a parent is often advanced to explain difficulties in adjustment and functioning among children growing up in single-parent families compared with children who grow up with both biological parents. However, the evidence for poorer outcomes is not consistent among children in single-parent families compared with children in stepfamilies (for example, Adcock and Demo 1994; Amato and Keith 1991), suggesting that the absence of a second parental figure is not the crucial factor.

Parental absence theories often refer to the lack of paternal authority. Although information on the role of fathers as child care providers has increased in recent years, and despite a social shift whereby men are expected to assume more direct responsibility for child care, studies of father-child interactions and their consequences are sparse. However, research is beginning to indicate benefits of positive father involvement, such as engagement, accessibility and responsibility (Lamb, Pleck, Charnov and Levine 1987), for specific types of child outcomes. These effects can be direct, arising from children's experiences and activities with fathers, or may be mediated through the psychological or instrumental support provided to the mother, or through provision of financial assistance. However, more information is required about the nature and consequences of father involvement for children, particularly among traditionally understudied family arrangements.

Given that the sequelae of family change is not adequately explained by parental absence or parental loss frameworks, it is necessary to explore what alternative explanations have to offer.

Family process models

Studies that suggest increases in the risk for child adjustment problems as the result of family membership cannot easily explain why children experience these outcomes. If we are to clarify the mechanisms behind increased rates of adjustment problems reported in single-parent and stepfamilies, these sources of risk must be examined. In order to understand how family structure influences outcomes in Australian children we need to look beyond structural characteristics of families and consider the resources and contexts within which relationships are established and the effective nature of those relationships.

As noted above, family structure itself does not automatically result in negative impacts on child wellbeing. Research that adopts a main effects model, or examines family structure as a direct causal influence on aspects of child functioning, fails to control for crucial intra-familial processes (such as parenting style and monitoring) and extra-familial contexts (such as economic stress and discrimination). Studies that have attempted to disentangle family structure from other factors tend to suggest that there are no simple causal relationships between family structure and child wellbeing.

As Michael Rutter (2002: 324) notes: 'The statistical associations may not derive from any direct reflection of a risk mechanism but rather an association between the postulated risk variable and with some other factor that truly causes the risk . . . For example, for many years the experience of a child being separated from its parents was seen as a major risk for mental health. The findings that this did not apply to happy separations and that the risks associated with parental divorce far exceeded those associated with parental death called that assumption into question (Rutter 1971). Since then, much research with both children (Fergusson, Horwood and Lynsky 1992) and adults (Harris, Brown and Bilfulco 1986) has shown that the major risk mediation derives from poor parental care or family discord rather than from the separation as such.'

Rutter's argument is that family type is a proxy for exposure to psychosocial risks (O'Connor, Dunn, Jenkins, Pickering and Rasbash 2001). That is, certain family types may encounter stressors associated with their family situation, such as compromised quality of parent-child relationships, parental depression and socio-economic adversity. These commonly occurring characteristics of different family environments may chiefly account for the risks to children's wellbeing (Lansford, Ceballo, Abbey and Stewart 2001; Dunn, Deater-Decard and Klebanov 1998; Smith, Brooks-Gunn and O'Connor 1997; Demo and Acock 1996; Forgatch, Patterson and Ray 1995; Amato 1994; Hetherington 1993; Hetherington and Clingempeel 1992; Amato and Keith 1991).

To unpack the relative influence of intra-familial processes and family structure on child wellbeing requires an approach that considers aspects of family functioning in addition to family structure. Attention must also be given to characteristics of the broader environment that indirectly affects child outcomes through their influence on families generally. For example, differences in family processes across family types may well be explained by factors such as social support or financial strain. Care must also be taken to evaluate the effects of individual differences between children and the timing of transitions and reorganisations characteristic of parental divorce and separation.

Hence, the challenge of understanding 'non-traditional' families is best tackled from a multi-dimensional framework, which considers multiple influences on children, including individual child factors, intra-familial processes, and broader familial environments.