We analyze the factors influencing dropout, defined as not enrolling in the second year, in the bachelor programs at the University of Trieste.
A total of 23,333 undergraduate students enrolled from 2003 to 2010 are considered.
The probability of dropout is modeled as a function of the characteristics of the students (available from Univ. of Trieste records), of the Faculty they are enrolled to and of their work placement (available from Ergon@t, the administrative archive of FVG region). We consider a relatively large set of covariates, including many interactions. Three different variable selection methods are chosen: GLM stepwise selection standard, Lasso and random Lasso.