Un modelo predictivo de fracaso/éxito académico a partir de indicadores de ingreso, en estudiantes de una Universidad Estatal del Norte de Chile

Extensive evidence supports the social and economic benefits generated at the individual and social levels by the academic success in higher education, as well as the costs of academic failure for people and their impact on countries' growth and development. Therefore, it is essential to have models that allow early identification of potential risk profiles to develop focused prevention. The objective of this study was to develop an empirical proposal for the prediction of academic failure/success in university students, based on early entry indicators, academic and sociodemographic, considering the mediating effect of the first year approval rate. An expost-facto correlation study was conducted, with the cohorts enrolled in the first year between the years 2001 and 2016 in a state university in northern Chile. The final sample was made up of 4012 graduates and 3393 students who defected or failed their undergraduate programs. Route analyzes were performed, estimating direct and indirect effects, differentiated by fields of study. Although each area of knowledge presented differential behaviors in the generated models, the results show that the predictors of entry were better proximal predictors (i.e., approval rates in the first year) than distal ones (failure/success). Although the best predictor of failure/success corresponds to the first year approval rate, the grades obtained in middle school turned out to be the best transversal predictor of performance at the time of entry, except in the engineering area.