The determinants of university dropout in Italy: a bivariate probability model with sample selection

Abstract

This study analyses the determinants of university dropouts in Italy using a bivariate probit model with sample selection. The empirical findings provide some support for the appropriateness of this technique and hence suggest that the estimates obtained from a univariate decision framework are likely to be inefficient.

Acknowledgements

The author thanks Professor Madhu Mohanty and Dr Dimitris Kyriakou for their valuable comments.

Notes

¹ Yet, it is important to note that in Italy the large majority of individuals who enrol at university do so straight after high school graduation. In our sample 85% of those individuals who enrolled at university did so immediately after leaving high school. Our attention is here focused on students who enrolled on four to five year university degree programmes (corsi di laurea).

² Unfortunately, it has not been possible to include subject of study among the determinants of dropouts as the survey reports it only for those individuals who are still enrolled at university.