Estimating the added value of critical reading competence in college students using statistical modeling

This study develops a model to predict students' scores and added value in the critical reading competency at Luis Amigó Catholic University, Medellín campus (Colombia). The institutional added value is assessed by comparing results from the Saber 11 test with an institutional exam conducted in the fifth semester. A linear correlation matrix was used to identify statistically significant variables at a 95% confidence level, with the accumulated average emerging as the most representative variable. Assumptions of linearity, normality, independence, collinearity, and homoscedasticity were verified. Outliers were identified using the William diagram, showing that most data adequately describe the model. Quintile analysis reveals that 35% of students exceeded the average in both tests, 23% showed improvement in the institutional exam, 25.2% had consistently low scores, and 16.8% performed better on the Saber 11 test.