This article implements some techniques of multivariate statistics such as principal components analysis, mutidimensional scaling (MDS) and classification trees and integrates them with the techniques of impact valuation to construct a difference model to measure the impact of the program “Application of autonomous activities to improve the academic performance of Pre-calculus students at Corporacion Universitaria Minuto de Dios –CMUDCentro Regional Soacha”. The results evidence a positive and significant influence of the valuated program as well as an influence of variables such as School and Socio-economical Stratum on the student scores. This establishes the importance of student s initial formation and family socioeconomic conditions on the learning process. Keyword: Academic performance, Impact Valuation, Multivariate statistics.