Statistical Analysis of State Tests Results for Admission to Higher Education in Colombia

Abstract

This study analyzes the Saber 11 common core scores obtained by the A-2016’s incoming students of the University of Santander. This diagnostic of the students’s background is proposed as a classification criterion to identify students with high risk of dropout based on their scores on each area evaluated. Besides the scores, variables such as type of high school, gender, selected program and region were considered. A multiple correspondence analysis was conducted to characterize the students. A factor analysis is proposed, which ended up being numerical, verbal and English. Finally, quantile regression is applied to measure the impact of some covariates on different levels of the factors. The results indicate that students from private schools and from Santander perform better in English and quantitative analysis, respectively.

Keywords

• Saber 11, deserción escolar, programas de acompañamiento académico, análisis de correspondencias, regresión cuantílica