intergenerational analysis
In this section we directly assess which factors help to explain why children from poor backgrounds tend to achieve lower test scores than those from better-off backgrounds. To do this we perform a ‘decomposition’ analysis similar to that shown in previous chapters. The unique contribution of this chapter is in understanding the importance of attributes and circumstances of parents, captured long before their children were born, in explaining this cognitive gap.
The starting point for this analysis is the ‘raw’ gap in cognitive test scores between children from the richest and poorest SEP groups, of 14.1 percentile points. Figure 6.3 shows how we decompose this raw gap into the relative contributions of different groups of factors in our model, which we divide very broadly into those related to the cohort member parent’s own childhood (in grey), and those relating to that of their children (in blue). These relative contributions Figure 6.2: Significant intergenerational transmissions: relative risks
1.75 1.39 1.49 1.71 1.43 1.15 1.25 1.32 2.00 1.00 1.50 2.00 2.50
Good cognitive skills Poor social skills Parent suffers from depression Lone parenthood Breastfeeding Parent thinks child is likely to go to university Parent reads to child daily Mother in full-time work In poverty (in poorest SEP quintile)
Age/gender controls only Additional parental ability/SEP controls
Transmissions in parenting from grandparents to parents Transmissions in child attributes from parent to child
Note: The relative risk is the ratio of the probability of an event occurring in one group to the probability of it occurring in another
group. These relative risks are estimated using Poisson regression models with robust error variances in which we control only for the child’s age and gender, and for whether the cohort member is the child’s mother or father (the results shown by the lighter bars are obtained by adding parental ability and the parent’s childhood SEP as additional controls).
are calculated by multiplying the difference in the proportions of rich and poor children with each characteristic by the coefficient estimates from a regression model of child ability that includes all characteristics simultaneously.17
This analysis shows that the direct effect of circumstances from parents’ childhoods explains nearly 40% of the gap in cognitive test scores between children growing up in rich and poor families. Of particular note here is the importance of the following:
• Parental cognitive ability: Nearly one fifth
(17%) of the gap is explained by parents’ cognitive ability at ages 5 and 10; moreover, this is the contribution that remains even after taking into account many of the channels through which cognitive ability might operate, such as parents’ subsequent educational attainment, adult SEP, attitudes to education, and so on. This relationship serves to highlight
the intergenerational cycle of poverty and low attainment: parents who grew up in poverty themselves performed poorly in cognitive tests as children relative to those from richer backgrounds, and this pattern is then repeated in the next generation.
• Educational attitudes and aspirations: 9% of the gap is explained by the apparent direct effects of grandparents’ and parents’ attitudes to education while the parent was a child. Of particular importance here are the expectations of the grandparents regarding the parents’ likelihood of staying on at school after 16, and the parents’ self-perceived maths ability at age 16. These contributions will be reinforced by any indirect effects, such as the influence of these attitudes on parents’ subsequent educational outcomes, or on educational attitudes and aspirations in the next generation. (As we showed in the last section, there are Figure 6.3: Explaining the gap in cognitive test scores between children who grow up in rich and poor families: decomposition analysis
Home learning environment 5% Family background, 9% Soclal skills, 0% Current Parent’s childhood Other 2% Other attitudes and behaviours 3% SEP during parent’s childhood 2% Social skills 12% Residual gap, 10% Educational attitudes/aspirations 9% Educational attitudes/ aspirations 17% Family background 11% Parental cognitive ability 17% Young person’s behaviours 10%
Note: The contributions of each set of factors are calculated by multiplying the difference in the proportions of rich and poor with each characteristic by the coefficient estimates from a regression model including all characteristics simultaneously. The way in which missing data is treated here differs slightly from the way in which it is treated in the other chapters in this report. For more details, see Crawford et al. (2010).
significant intergenerational links between some of these attitudes.)
It is interesting to note, however, that parents’ social skills, as captured in childhood at ages 5 and 10, do not appear to make any direct contribution to the gap in cognitive test scores between children from rich and poor backgrounds.
Among the circumstances in the current
generation that are important for explaining the gap in cognitive test scores between children from rich and poor backgrounds, we note that differences in educationalattitudes and aspirations, the
home learning environment, young people’s
risky and positive behaviours and young people’s social skills (measured by a Strengths and Difficulties score) between them account for 44% of the gap in cognitive test scores between children from rich and poor backgrounds. Factors within these groups of particular importance are the parent’s assessment of the likelihood that the child will attend university; whether or not the child wants to stay in education beyond age 16; whether or not the child regularly reads for enjoyment; and whether or not the child smokes.
Overall, attitudes and behaviours in the previous generation account for around 40% of the gap in test scores between children from rich and poor backgrounds. This suggests that while circumstances in the current generation are more important in explaining why children growing up in rich families tend to have higher cognitive test scores than children growing up in poor families, the importance of intergenerational influences
in perpetuating the cycle of poverty and low attainment should not be underestimated.