The measurement of students performance. The use of an extended Rasch model for the analysis of predictors of high educational performance

Characteristics of the higher educational programs (e.g. non systematic variability of course’s difficulty among and within programs and over times) make observed data (e.g. number of credits acquired) poorly informative indexes of the students’ performance. As alternative, it is proposed an extended version of Rasch model (the Three Facets Model, TFM). TFM conceptualizes student’s performance the expression of a three-component latent variable to be esteemed. In so doing, TFM is able to take into account the non-systematic sources of variation characterizing higher educatonal settings, thus avoiding limits entailed in the use of indexes based on observed data. An exemplificative longitudinal case study has been performed, aimed at detecting predictors of performance within an undergraduate program of psychology of an Italian university. Two regression models have been compared: one using a traditional index of performance based on observed data versus one using the TFM estimation.