A longitudinal analysis of the decision to drop out of post-compulsory education

In this paper we analyse the magnitude, timing, determinants and outcomes of dropping out of post-compulsory education over the period 1985-94. We use data from the Youth Cohort Surveys for England and Wales to estimate non-parametric single and competing risks duration models allowing for the effects of unobserved heterogeneity and time varying covariates. By the mid-1990s approximately 1 in 10 young people dropped out of post-compulsory education, especially in April and July. Dropouts were more likely to get jobs rather than become unemployed and the most important predictor of the likelihood of dropping out was prior attainment.