Evasão acadêmica em cursos de tecnologia da informação no IFPE: uma análise de sobrevivência

Abstract

This article investigates student dropout in Information Technology programs at the Federal Institute of Pernambuco (IFPE). This is a quantitative study based on longitudinal data from a cohort of 16,264 students enrolled between 2001 and 2025. Survival Analysis was applied, with emphasis on the Cox Proportional Hazards Regression Model, to model the time until dropout and identify the effects of academic and demographic covariates. The independent variables analyzed were: age, gender, grade point average, education level, study shift, and race/ethnicity. The model was fitted with 14,587 observations and showed good discriminative ability (c-index = 0.81). Results indicated that grade point average was the main protective factor, reducing the risk of dropout by approximately 30% for each additional point. Students in the morning shift had a 21% higher risk of dropout compared to those in the evening shift, while the night shift was associated with a 15% lower risk. It is concluded that strategies aimed at improving academic performance and adjusting course schedules can significantly reduce dropout rates, providing support for more effective educational policies at IFPE and similar institutions.