7. Discussion
- 7.1 The LSAC Outcome Index and children's developmental competencies
- 7.2 Families, children and social demography
- 7.3 The infant cohort
- 7.4 The child cohort
- 7.5 Implications for policy
There is consensus that the early years of life are particularly important to children's ongoing physical health, social–emotional development and early learning (McCain, Mustard & Shanker 2007; eds Shonkoff & Phillips 2000). The experiences of children during the early years are considered important because they affect biological pathways as well as providing the basis for subsequent learning. Development is a cumulative process and attainment at any age can have important consequences for later competencies and capacities (see, for example, McCain & Mustard 1999; Heckman 2000; and the National Scientific Council on the Developing Child 2007). The processes by which early experiences affect later patterns of development are inherently complex. However, research findings from LSAC, as a nationally representative sample of children, can help researchers, practitioners and policy makers understand some of these complexities about the nature of Australian children's experiences and their potential impact on later development.
This section provides a review of the findings from LSAC Wave 1 data presented in the previous sections. Its main focus is the relationships between key independent variables and the overall Outcome Index of children's development, as well as outcomes in the physical, social–emotional and learning domains. The implications of these findings for policy are discussed and the complexities in using information from the Outcome Index to inform policy are also considered.
Perspectives from several disciplines were encompassed in the design and measurement of LSAC, and thus shape the interpretation of findings in this report. As discussed in Section 1, ecological systems theory is proposed as a theoretical model to explain how children's development is influenced by the experiences available to them in society (Bronfenbrenner 1989, 1993). Bronfenbrenner argues that development occurs throughout life and is influenced by events and experiences across time, in various contexts, that directly and indirectly impact on the developing child (for example, interactions between children and parents in the home; the quality of parental work experiences; the quality of early education programs). Further, interactions across these contexts have important effects (for example, parents' relationships with school; the impact of work on family life). Whereas psychological and sociological perspectives emphasise intrinsic, environmental and cultural factors, economic theories place more emphasis on how societies and families make choices on the use of resources that affect outcomes for children. For example, Haveman and Wolfe (1995) proposed that three factors affect children's wellbeing: social investment determines the nature of options available to children or their parents in a society; parental investment is about the choices parents make about the quality and quantity of resources they devote to their children; and children's investment reflects the choices children will make themselves, a factor that begins to exert greater influence over time as children are able to make personal decisions. Thus, different theories may inform interpretations of the data analyses in understanding the experiences of young children and their families in Australia.
7.1 The LSAC outcome Index and children's developmental competencies
As discussed in Section 1, reports using social indicators provide an efficient way for policy makers to understand trends and present information about subgroup differences within the larger population (for example, by gender, race/ethnicity, and income status), as well as the influence of particular life experiences and opportunities available to groups within the population. The use of aggregated measures to reflect children's developmental competencies and wellbeing has evolved internationally during the last 25 years (Ben-Arieh 2006). However, there is still much to be learned about the substantive issues in the use of such measures (Land et al. 2007; Moore, Brown & Scarupa 2003). The LSAC Outcome Index is unusual in that it is a composite measure that reflects not national trends, but individual children's competencies across three domains: physical, social–emotional, and learning. The Outcome Index enables representation of developmental strengths as well as vulnerabilities.
The analyses in this report utilised Wave 1 data for the LSAC infant and child cohorts. As noted in Section 1, not all of the outcomes that were measured for the child cohort were measured for the infant cohort. This reflects the reality of development, with some skills either not observable at such a young age or not measurable within the LSAC design. Hence, the meaning of the Outcome Index in these analyses varies to some degree across cohorts and limits the comparisons that can be made between cohorts.
In Section 1, we reported the number of domains on which a child's score was in the top 15 per cent, and in the bottom 15 per cent, of the range of scores on any specific domain. The findings revealed that, as expected, development does not occur uniformly across all domains at these ages. Approximately two-thirds of children in both cohorts (infant and child) did not score below the negative cut-off (bottom 15 per cent) on any of the three domains. A similar pattern was revealed in relation to the positive cut-off (top 15 per cent). For infants, it was not possible to define a positive cut-off in the physical domain, so no infant could score above the cut-off on all three domains. These summary findings of the percentage of children who are identified with developmental strengths or difficulties need to be interpreted with caution. The selection of 15 per cent to indicate the percentage of children who are doing well or who are having difficulties, overall or in any domain, is essentially an arbitrary decision, but relates to the statistical concept of one standard deviation below the mean and is also in line with prevalence estimates from population surveys such as the National Mental Health Survey of Children and Youth (Sawyer et al. 2000).
7.2 Families, children and social demography
Families have experienced rapid change over the last few decades. Understanding the nature of families and the complexities in their social demography imparts important information to researchers and policy makers. It can no longer be assumed that children live in a household with two biological parents. Households are complex structures and relationships of members of households are likely to vary considerably over time. Increased maternal employment, varying levels of unemployment, and changing work characteristics, as well as the changes in the racial and ethnic composition of Australian society, make it important to understand how such social variations, defined by demographic characteristics of families, impact on the lives of children (Sanson & Lewis 2001). Bronfenbrenner (1979) drew attention to the limitations of relying on a 'social address' to understand the experiences of children in families because demographic variables are distal representations of more proximal family processes that directly impact on the lives of children. However, sociodemographic characteristics remain important variables in research to inform policy about how different groups in society are faring, the nature of family diversity, and how vulnerable families with young children might be supported.
An emphasis in the analyses in this report was on multivariable associations between various sociodemographic characteristics on children's outcomes. Across the analyses in this report, nine sociodemographic variables were selected for inclusion because of their theoretical and policy interest and use in comparable research reports (for example, Hoddinott, Lethbridge & Phipps 2002; Jones et al. 2002; Lugaila 2003). This set of sociodemographic variables, as well as being analysed to explore their specific associations with infant and child outcomes (see Section 2), were also included in subsequent sections of this report as important variables in the analyses that investigated the influences of non-parental care, child health, maternal health, and family learning environment (child cohort only) on outcomes.
The selected sociodemographic variables represented four tiers of influence (child, mother, family and neighbourhood) as proposed by ecological models of development (Bronfenbrenner, 1979, 1989, 1993). Child level variables were gender, Aboriginal and Torres Strait Islander status, and speaking a language other than English. Maternal education was included because it is known from other research that it is an important influence on children's outcomes across domains of health, social–emotional development, and learning (see, for example, Hoddinott, Lethbridge & Phipps 2002). Family level variables included were family type, family income, financial stress and parental occupational status. Low family income and poverty, in particular, have been strongly associated with poorer outcomes for children internationally (see, for example, Logan et al. 2007; Daly 2006). However, a simple finding that children in low-income households, or those with any other specific sociodemographic characteristic, experience fewer opportunities and less optimal developmental experiences does not make it possible to establish whether such a characteristic is a causal factor, or if other factors are also operating that affect outcomes at the same time. For example, less well educated parents may have lower than average income and hence be less able to support optimal opportunities for their children. These reciprocal and indirect influences need to be considered in the interpretation of findings, requiring a multivariable analytical approach as used here.
Within the set of demographic variables used in the analyses, a measure of neighbourhood quality was also included because there is increasing evidence that neighbourhood variables are important to children's developmental outcomes (see, for example, Tremblay et al. 2001). The variables included measured social advantage/disadvantage of the neighbourhood in which a family lived, that is, the Socio-Economic Indexes for Areas [SEIFA] quintiles (ABS 2003b).
7.3 The infant cohort
Although few contemporary experts in child development would view the first two years of life as a 'critical' period during which the path of future development is fully determined, this stage of development is still regarded by many as being a 'sensitive' period during which developmental pathways are first established (Halfon & McLearn 2002; National Scientific Council on the Developing Child 2007). Because early developmental pathways have the potential to impact on subsequent development in important ways, the influences that shape early development need to be considered. However, for the infant cohort of LSAC (mean age of 9.5 months), overall there were relatively few predictors that exerted influence on outcomes from the cross-sectional analyses. These findings are discussed below within the various areas of the data analyses.
Sociodemographic characteristics and outcomes
For the infant cohort, the analyses of relationships between child, family and neighbourhood factors and the Outcome Index measures suggested they had only minor impact. This may reflect weaknesses in the measurement of the Outcome Index itself for this age group, since it is difficult to collect very sensitive outcome measures for infants using the methodology employed in LSAC. It may also reflect the fact that the impact of contextual factors on children's development is a cumulative process which occurs over time. Early measures of 'outcomes' may largely reflect infants' biological predispositions and the effects of environmental influences may only be seen at later stages. This hypothesis can be tested when data from future waves of LSAC become available.
Non-parental care experiences and outcomes
The impact on development of children's participation in formal and informal child care and early education settings is not independent of other family and child factors. For the infant cohort, fewer children were in a non-parental care/education arrangement if their families were economically disadvantaged, if they were of Aboriginal or Torres Strait Islander background, if their mothers had not completed high school, and if the main language spoken at home was other than English. These and other related indices of family socioeconomic status are all variables associated with poorer outcomes in the broader developmental literature, and it is important to take these factors into account when trying to assess the relative importance for children of their experience in care settings.
No child care variables were strong predictors of the overall Outcome Index score for the infant cohort. However, child care variables contributed to the prediction of the physical and learning domain scores. Infants not in care had better physical domain scores than infants in formal or mixed formal/informal care, infants in more than 20 hours of care per week, and infants in one or more care arrangements per week. These findings indicate that children in larger group settings are at most risk for perceived poorer general health and additional health care needs in the first year. These findings are consistent with those reported by the NICHD Study of Early Child Care and Youth Development (NICHD Early Child Care Research Network 2005a) in the United States. For the learning domain, there was evidence of higher scores for infants who experienced only informal care compared to infants not in care, while the outcomes for infants in formal care or mixed formal/informal care did not differ from those of infants not in care. Because most infants in informal care only were being cared for by grandparents, this finding should be explored in further analyses about the influences that extended family as care givers may have on infants' development.
Infant health and outcomes
Most of the children in the infant cohort were reported to be in very good or excellent health. However, there were significant numbers with physical health problems including low birth weight or preterm birth, special health care needs, rapid weight gain since birth, and a very high prevalence of infant wheeze. Additionally, while most mothers had breastfed their children, only half of all mothers had met the current NHMRC recommendations for exclusive breastfeeding for at least six months. Almost all mothers who had not breastfed for at least six months had either never established breastfeeding or breastfed for less than three months.
For the infant cohort, using a multivariable model, it appeared much of the influence of these early risk factors was yet to become apparent. Of strong policy relevance, however, was evidence that very preterm birth was strongly associated with poorer Outcome Index scores; prolonged breastfeeding was strongly associated with better physical and social–emotional outcomes; and infant wheeze was strongly associated with poorer physical and social–emotional outcomes. Given that preterm birth, breastfeeding and infant wheeze are potentially modifiable through social, welfare and health policies, these are important findings. Paradoxically, given its emerging associations with later overweight and obesity (Stettler 2007), rapid weight gain since birth was associated with better outcome scores in this cohort. This finding could prompt a re-evaluation of growth trajectories perceived as 'healthy' by parents and promoted by early-years health professionals.
Maternal health and outcomes
Across both cohorts, the LSAC mothers were broadly representative of all Australian mothers on parameters such as smoking and alcohol in pregnancy, mental health status, and a high prevalence (45 per cent) of overweight and obesity. In the multivariable analyses, there was no evidence that diabetes, hypertension, or cigarette smoking during pregnancy were associated with overall Outcome Index or domain scores for the infants, suggesting that the correlations in the bivariate analyses between heavy prenatal smoking and infant Outcome Index scores may be explained by sociodemographic factors. Two factors stood out as being strongly associated with better infant outcomes: good maternal general health (especially the physical and social–emotional domains) and greater maternal enjoyment of physical activity (especially the social–emotional and learning domains).
The findings confirmed just how strongly maternal health and lifestyle are intertwined with child outcomes even at this young age, and support the need for holistic health care and policies targeting both these areas. An important corollary is to consider fathers' health and lifestyle in future analyses. Unlike many other studies, LSAC is collecting the data to address the influence of not just mothers but fathers on their children's health and other outcomes.
7.4 The child cohort
While the family remains the key influence on children's experiences and outcomes throughout childhood, children at age 4 to 5 years are also more likely to spend more time in contexts outside the immediate family and participating in early education programs, including child care programs. By 4 years old, children are developing the skills and competencies that will hold them in good stead for a successful transition to school. Features of children's home environment, such as reading to children and participation in educational activities, can have substantial effects on children's future language and literacy skills, as can participation in early education programs. Family characteristics, such as household income and mothers' level of education, are also related to these competencies. At this age, social and emotional competencies, including self-regulation of behaviour, also become important. Variations in these competencies have been found to be influenced by parental mental health problems, and associated with family socioeconomic status (Tremblay et al. 2001). For the child cohort of LSAC (mean age of 4.7 years) there were a number of findings comparable to previous research conducted in other national contexts.
Sociodemographic characteristics and outcomes
In contrast to the infant cohort, sociodemographic factors accounted for substantial variation in the Outcome Index measures for the child cohort, indicating a role for child, maternal and family characteristics. Girls consistently showed more positive outcomes than boys. Aboriginal and Torres Strait Islander children had poorer outcomes in all but the physical domain. Similarly, children in families where a language other than English was spoken tended to have poorer outcomes. There was strong evidence for more positive outcomes with higher maternal education, with higher family income, and in the absence of financial stress. Children whose parents had a skilled or professional occupation also had more positive outcomes. Family type (two parents versus single-parent family) and neighbourhood disadvantage, while showing some bivariate relationships with outcomes, did not make a unique contribution to overall child outcomes, suggesting that their influence may be mediated through family variables such as income, financial stress and family occupational status.
Non-parental care experiences and child outcomes
As noted for the infant cohort, children's participation in formal and informal child care and early education settings is not independent of family and child factors that influence children's development. For the child cohort, fewer children were in a non-parental care/education arrangement if their families were economically disadvantaged, if they were of Aboriginal or Torres Strait Islander background, if their mothers had not completed high school, and if the main language spoken at home was something other than English. These and other related indices of family socioeconomic status are all variables associated with poorer outcomes in the broader developmental literature and, as noted above, were also shown here to be related to Outcome Index scores. Hence it is important to take these factors into account when trying to assess the relative importance for children of their experiences in formal (or informal) early education and care programs.
In the multivariable analyses for the child cohort, children who were attending a pre-Year 1 program had higher overall Outcome Index and learning domain scores than children who were attending informal care arrangements only. There were no significant differences between children attending a pre-Year 1 program and children who attended other group programs such as preschool or day care, although a higher proportion of children attending pre-Year 1 programs had higher Outcome Index scores than children in preschool or day care. This relative advantage for children attending pre-Year 1 programs may be explained by the nature of the program (for example, full-time with a strong focus on the development of academic skills). These children were also likely to be slightly older than the children with other care arrangements because age eligibility criteria for participation in pre-Year 1 programs varies across states and territories, so that only children whose birth date was later in the year were able to be enrolled. Additionally, children in pre-Year 1 programs also may have been older because of parental choices to delay entry even if the child was age eligible in the previous year, so the child was more mature and able to cope with the demands of a full-time school program.
Child health and outcomes
Most of the children in the child cohort were reported to be in very good or excellent health, but many also experienced physical health problems. Such problems included a high prevalence of each of special health care needs, overweight/obesity, and asthma requiring medication. As for the infant cohort, while most mothers had breastfed their children, few mothers had met the NHMRC recommendations for exclusive breastfeeding for at least six months. Less than one-quarter of children had diets that adequately met nutritional guidelines and many preferred less physical activities.
In a multivariable model, relationships were more evident than for the infant cohort, suggesting that impacts may be cumulative and that adverse outcomes may take time to become apparent. Low birth weight, preterm birth, and asthma were all strongly predictive of poorer outcomes, while longer breastfeeding, enjoyment of physical activity and healthful nutritional behaviours were all strongly associated with better outcomes.
Maternal health and outcomes
As noted for the infant cohort findings, the LSAC mothers were broadly representative of all Australian mothers on parameters such as smoking and alcohol in pregnancy, mental health status, and a high prevalence of overweight and obesity.
More and stronger associations were seen between maternal factors and child outcomes in the child than in the infant cohort multivariable analyses. Prenatal factors (diabetes, hypertension and cigarette smoking) were far outweighed by the current issues of poor maternal health, psychological distress, and mothers' own enjoyment of physical activity. These effects were often greater in the social–emotional and/or learning domains than in the physical domain, emphasising the relevance of mothers' health to all aspects of their children's functioning and wellbeing. Overall, these findings underscore the importance of policy initiatives targeting maternal (and therefore, probably, paternal) health and lifestyle. These perhaps surprising findings raise challenging policy issues, since healthy lifestyles are often considered the province of individual responsibility. Lack of perceived impacts on their children may be another reason for little motivation for individual behaviour and lifestyle change in young parents. This may well require substantial policy as well as research realignment, since neither maternal nor paternal factors figured prominently in a recent Australian study of stakeholder priorities for population child health and wellbeing (Davis et al. 2005). However, 'lifestyle' indicators such as heavy current alcohol intake, smoking, mother's own weight status, and maternal fruit and vegetable consumption were not predictive of children's outcomes at this young age.
Family learning environments and child outcomes
There was strong evidence that a number of opportunities and resources available in the family contributed to higher overall Outcome Index scores. These findings were in line with those from research in other national contexts (for example, Foster et al. 2005). These opportunities and experiences included that the child was read to by a family member on three or more days per week; there were 10 or more children's books in the home; the child enjoyed being read to for more than 10 minutes at a time; and the child had access to a computer at home. Positive child outcomes were also associated with a medium or high level of engagement by the child in out-of-home activities with family members.
Evidence was also strong for an association between the overall Outcome Index and hours spent television viewing on weekdays, with the Outcome Index score tending to decrease as time spent watching television increased. There was weaker evidence for a similar trend for weekend television viewing. Similar results were found when using the learning domain score as the outcome rather than the Overall Index, except there was no evidence of an association between the numbers of hours that children spent watching television on weekdays or on the weekend and learning outcomes. This suggests that the impact of time spent watching television upon child outcomes is attributable to either or both of the physical and social–emotional domains.
7.5 Implications for policy
The Wave 1 data provide a wealth of information on children's physical, social–emotional and learning competencies. The lives of the two cohorts are being assessed across multiple contexts, including home, school and community. The pattern of results for the child cohort supports an ecological model of child development in which the child's own attributes, along with their family and community context, exert influence on developmental trajectories. While the effects were weaker for the infant cohort, this could be interpreted in terms of the likelihood that there is a greater impact from cumulative external influences (for example, disadvantage) over time. For infants, less time had elapsed for such influences to impact on development. The findings underscore the importance of using a broad conceptual framework when trying to understand the complex nature of children's development. It also draws attention to the importance of longitudinal analyses. Limitations of the analyses provided in this report include the fact that the findings are based only on the first wave of data collection of LSAC and are, therefore, cross-sectional in nature. However, these findings do provide an initial basis for understanding a range of influences on developmental outcomes. Information on important influences on outcomes will be enhanced through the addition of subsequent waves of data.
The results of these analyses of the LSAC Wave 1 data showed a number of strong relationships between child and family factors and child outcomes. Important influences on outcomes were suggested within every one of the five areas studied (sociodemographic circumstances, non-parental care, child health, maternal health, and family learning environments). However, it is necessary to qualify these findings; since all the analyses reported are cross-sectional, cause and effect relationships cannot be identified and, for some of the associations, the effects on outcomes were modest in size. Given the current understanding of the importance of early child development for later development (eds Shonkoff & Phillips 2000), it will be important to continue to explore how the data from this stage of LSAC predict later child outcomes across childhood.
The persistent public and policy attention to indicators of negative outcomes for children has led to comment that there is an unbalanced depiction of children and families. Many of the findings reported in these analyses indicate that most children are doing well. Such positive findings deserve more attention. However, a considerable number of children do show decreased competence for overall outcomes and outcomes in specific domains. These children were not evenly represented across the Australian population in terms of the sociodemographic characteristics of children and families. While there was evidence that outcomes for significant numbers of children could be improved right across the socioeconomic spectrum, it was clear that there was also a socioeconomic gradient with poorer outcomes in the context of greater family disadvantage. Policy initiatives should ensure that there are effective systems to support the development of social and psychological capital within families through parenting programs and family support services in communities. Given the wide range of environmental risk factors that can influence family functioning and therefore children's development, a multi-service and whole-of-government approach to policy, crossing the health, education and community sectors, remains important. Universal and primary care services across the health, education and community sectors need to be coordinated to address multiple environmental risk factors and respond to the needs of families with complex needs. This direction has been important in current policy initiatives in the Stronger Families and Communities Strategy (Department of Families and Community Services 2004a) and the National Agenda for Early Childhood (Department of Families and Community Services 2004b). It is important that services for families are flexible so that they can respond to the emerging needs and problems of children and families rather than waiting until problems become established.
Intervening early in the life course has the greatest potential to prevent or significantly ameliorate some of the health and wellbeing problems seen in adult life (National Scientific Council on the Developing Child 2007). From an economic perspective, responsible investments in services for young children and their families focus on benefits relative to cost (Heckman 2000). Because risk factors frequently cluster together and are cumulative, interventions that focus on single issues are unlikely to lead to lasting effective change. The central issue is that a myriad of influences (individual characteristics, family circumstances and social and community resources) are likely to influence children's developmental pathways through childhood and into adulthood, and the relative importance of various influences may vary over time. Addressing inequalities for children and families is a critical social investment. A direct way to improving outcomes in childhood is to provide support to ensure that family environments in the early years meet the health and developmental needs of young children. Interventions need to focus on supporting efforts to provide healthy, nurturing and stimulating family environments for children.