Conference Agenda

Student persistence in engineering programs represents a significant challenge for higher education institutions, particularly during the early semesters, where foundational courses exhibit high repetition and failure rates. In this context, the present study examines the relationship between academic performance, course load, and student persistence through the construction of longitudinal academic trajectories. The analysis is based on an institutional dataset comprising academic records of students enrolled in core courses of a Software Development program between the 2022-1 and 2025-1 academic terms. From these data, a synthetic performance indicator referred to as the approval rate was developed, along with a binary variable representing student persistence. The analytical approach combined descriptive statistics, exploratory visualizations, and a logistic regression model to assess the relationship between cumulative academic performance and the probability of persistence. The results reveal a strong association between approval rate and student persistence, with the proposed model demonstrating high predictive power. Additionally, the individual-level analysis of the relationship between course load and academic performance reveals substantial heterogeneity in students’ responses, enabling the identification of distinct academic profiles. These findings highlight the relevance of longitudinal performance indicators for understanding student trajectories and for supporting institutional strategies aimed at promoting academic persistence.