Efficacy of programmes for the unemployed: discrete time modelling of duration data from a matched comparison study

An evaluation is described of two UK Government programmes for the long-ter m unemployed in Great Britain, Employment Training and Employment Action, u sing discrete time hazard modelling of event histories. The study design em ployed a closely matched comparison group and carefully chosen control vari ables to minimize the effect of selection bias on conclusions. The effect o f unobserved heterogeneity is investigated by using some standard random ef fect model formulations.