Dropping out of University of Salerno: a survival approach

Abstract

The aim of this paper is to analyse and model the interval of time between the first enrolment at University and the first occurrence of non-enrolment, so that the event of interest is the dropping out of University. The interest is focused on computing the probability of surviving at the University and analysing which personal, familiar and social characteristics may influence the non-completion of students’ academic carrier, using Survival Analysis techniques. The dataset analysed is collected from the Central Administrative Office of the University of Salerno, and includes all full-time students enrolled at the 2002-2003 academic year, followed for five years, until the 2006-2007 academic year. Those students can either complete their study and receive their degree or leave the university, that is the event of our interest. The estimation of the probability of surviving in University is made by the Kaplan-Meier estimator. Then, to test if there is a significant difference among the survival curves, Log-rank test is considered. Since it does not allow more than one explanatory variable to be taken into account, Cox Proportional Hazards model is used to analyse the interrelations between the covariates, and explored the influence of covariates on failure times. This study shows that there is a steep decline for students in Political Science and in Educational Studies, at the first year. Moreover, female students, students who attended "Licei" and those who completed their high school study with the highest mark (i.e. 110) have the highest probability of surviving at the University, i.e. the highest rate of taking the degree. In all these cases, it can be noted that there is a Faculty effect. These differences, tested by logrank test, are significant. Finally, Cox PH model for each Faculty confirms previous results.