Youth Unemployment: Longitudinal Studies

Abstract

A recurrent theme in my discussion of the single-group studies and comparison-group studies described in the two preceding chapters has been recognition of the need to conduct longitudinal investigations of the effects of employment and unemployment. Without these kinds of studies results are ambiguous. It is always possible in cross-sectional studies to present alternative interpretations of the results. For example, if one finds that depressive symptoms are associated with unemployment in a correlational study one can either argue that this is because individuals who are depressed to begin with find it difficult to obtain jobs or because being unemployed leads to depressive symptoms. Warr (1987a) observes that the former type of interpretation is more likely to apply for cross-sectional studies when unemployment rates are very low. When there is an easy labor market it seems more reasonable to assume that those who fail to get jobs are likely to be those with personal attributes that impede them from becoming employed. When unemployment rates are high, however, and a lot of people miss out on obtaining employment, the interpretation in terms of prior characteristics becomes less plausible and it seems more reasonable to assume that differences are produced by the employment or unemployment experience (Spruit, Bastiaansen, Verkley, Van Niewenhuijzen, & Stolk, 1985; White, 1983).