A multistep approach to detect and correct the cheating in Italian students data

We analyze students data carried out by INVALSI and propose a Multi Step Approach (MSA) for cheating detection and correction. This method integrates the "mechanistic" logic of fuzzy clustering with a statistical model based approach. The procedure aims to minimize the detection of false positives and to correct test scores at both class and student level. The results show a normalization of the scores and a stronger correction on Southern regions data where the propensity to cheating appears to be highest.