This paper evaluates the appropriateness of the standard methodologies and the quality of the data frequently used to analyse labour market dynamics in Europe. Our results indicate that, due to recall error and heterogeneous survey design, the retrospective approach tends to result in a considerable number of spurious transitions being recorded. Whilst the use of quasi-longitudinal data should overcome such problems, sample attrition and more importantly, misclassification error, is shown to result in significant over-reporting of transitions. Studies which failure to allow for the error structure of the underlying data are therefore, likely to be subject to considerable bias.
Additional information:
Das Dokument wird vom Publikationsserver der Universitätsbibliothek Mannheim bereitgestellt.