Attendance trajectories as early predictors of school dropout: Evidence from a National Cohort in Chile

School dropout is associated with long-term economic, social, and health consequences. Evidence also suggests that preventive interventions and early identification strategies are more effective than post-dropout reintegration efforts. This study examines predictors of dropout at the transition to high school in Chile using longitudinal administrative records for a full national cohort (N = 248,169) followed from 2020 to 2024. We estimate a series of logistic regression models with school-clustered robust standard errors, sequentially adding attendance and academic achievement trajectories. Results show that attendance—especially in the years immediately preceding the transition—is the most powerful predictor of dropout, exceeding the explanatory contribution of achievement. Misalignment between age and grade and socioeconomic disadvantage are also strongly associated with dropout. Importantly, students in public schools who are not formally classified as socioeconomically vulnerable exhibit the highest predicted dropout probabilities, suggesting potential gaps in targeting and identification mechanisms. Findings support strengthening the early identification process, improving vulnerability identification, and intervening before disengagement becomes entrenched.