CASE LSE RSS Email Twitter Facebook

Abstract for:

An exploration of childhood antecedents of female adult malaise in two British birth cohorts: Combining Bayesian model averaging and recursive partitioning

John Hobcraft,  Wendy Sigle-Rushton,  March 2005
Paper No' CASE 095: Full paper (pdf)
Tags: wealth and social mobility; intergenerational and social mobility; children, families and education; children and child poverty; poverty, exclusion and equalities; poverty and social exclusion; health and social care; health; methodology, concepts and measures; employment and income; measurement, concepts and measures; well-being; cohort; bayesian model averaging; recursive trees

Abstract:

We use information from two prospective British birth cohort studies to explore the antecedents of adult malaise, an indicator of incipient depression. These studies include a wealth of information on childhood circumstances, behaviour, test scores and family background, measured several times during childhood. We are concerned both with incorporating model uncertainty and using a person-centred approach. We explore associations in both cohorts using two separate approaches: Bayesian model averaging (BMA) and recursive trees. The first approach permits us to assess model uncertainty, necessary because many childhood antecedents are highly correlated. BMA also aims to produce more robust results for extrapolation to other data sets through averaging over the range of plausible models. The second approach is concerned with partitioning the sample, through a series of binary splits, into groups of people who are as alike as possible. One advantage is that the approach is person-centred in that it retains real groups of respondents. We compare and contrast the insights obtained from the two approaches and use the results from each to inform the other and thus refine our understanding further. Moreover, we explore the claimed added robustness for extrapolation by using a split-sample for the 1970 cohort. The consistency of results across methods and cohorts is discussed throughout.