The purpose of this paper is to assess the measurement sensitivity of the estimated determinants of overeducation.
The paper analyses the determinants of overeducation among Flemish school leavers in their first job by means of probit regression analysis. Overeducation is measured on the basis of job analysis (JA), self‐assessments and realised matches.
The results demonstrate that the application of different overeducation measures sometimes leads to different outcomes. Only a few variables – for instance the student's academic grade in the final year – are consistently found to be important for the explanation of overeducation. Some outcomes are consistent with the supposition that several indicators actually measure other concepts.
Further research using JA measures that are based on alternative and more recent occupational classifications would be useful.
The application of different measures provides further insight into the overeducation measurement problem.
