Lo studio dei successi e degli insuccessi universitari tramite analisi di segmentazione

Using the characteristics of a sample of 4.652 students of the university in Bari, graduates or drop-out (known outcomes), the segmentation analysis has been applied, identifying a four levels tree with 0,39% classification error. The identified criteria of segmentation have been validated at first through a further sample (independent from the first one), getting an error of classification equal to 0,37%, and then also applied to the universe of the students with unknown outcomes (students still affiliates or not more enrolled that have not formalized their renounce to the studies), discriminating students drop-out-like from those whose remarkable characteristics are more similar to those of the graduates.