Are there changing socio-economic inequalities in childhood cognitive test performance? Methodological considerations from the analysis of three British birth cohort studies

Presenter: Roxanne Connelly, University of Edinburgh

Author: Roxanne Connelly

Co-author(s): Vernon Gayle

There is a large international literature that identifies socio-economic inequalities in childhood cognitive test performance. In this paper we investigate changes in this relationship in Britain from the mid-twentieth century, through a comparative analysis of existing longitudinal data.

We undertake analyses of three of the major British birth cohort studies, The National Child Development Study (1958), The British Cohort Study (1970) and The Millennium Cohort Study (2000/02). The paper addresses three distinctive methodological challenges. The first is the development of measures that are comparable across the three cohorts. The second is the provision of a sensitivity analysis of different measures of parental socio-economic position. The third is the development of a strategy for analysing data from the three cohorts within a unified multivariate framework which appropriately accounts for the variation in the design and structure of the datasets.

The results indicate that there is a persistent link between parental characteristics and children’s performance on verbal similarities tests however the effects have decreased between cohorts. The paper presents these original British findings, which are in line with existing results from international studies on childhood cognitive ability. The paper provides methodological reflections on the development of cross-cohort measures, and evaluating alternative socio-economic measures. The paper also provides clear prescriptions relating to combining data from surveys with different designs, which is an increasingly common problem in survey data analysis, but which is frequently overlooked.