Quantile regression estimates of the effect of student absences on academic achievement

Credible evidence from a variety of contexts suggests that student absences harm academic achievement. However, extant studies focus entirely on the average effects of student absences, and how those average effects vary by student, school, and absence type. This paper enhances our understanding of the nature of the causal relationship between absences and achievement by estimating quantile regressions that identify the impact of student absences on the full distribution of achievement, not just its mean. Somewhat surprisingly, the harmful effects of student absences are approximately constant across the achievement distribution. This suggests that cost-benefit analyses of interventions designed to improve attendance can use previously-estimated average effects to predict benefits. Moreover, it suggests that interventions that target all students would neither increase nor decrease the variance in test scores.