Reducing Student Absenteeism
Category: Student Learning
Chronic absenteeism has surged in recent years, drawing growing policy and research attention. However, a complicating factor often overlooked is that the measurement of absenteeism is inconsistent, with substantial researcher degrees of freedom. This study investigates how researchers’ measurement choices shape predictions of academic risk and how absenteeism can be more effectively operationalized as an early warning signal. Using a sample of 8,891 students followed from Pre-K through eighth grade in Boston Public Schools, we (1) describe developmental patterns in absenteeism; (2) apply ROC curve analyses to evaluate the diagnostic accuracy of multiple absence measures for predicting scores on state standardized tests; (3) use Youden’s J to derive empirical thresholds for chronic absenteeism; and (4) assess predictive validity over time. We find that measurement choices matter. By middle school, total and unexcused absences are more predictive of low academic performance than excused absences. Additionally, empirically derived thresholds for identifying students at risk consistently fall below the widely used chronic absenteeism benchmark of missing 10% of school days (≈18 days). We discuss implications for research, policy, and early warning systems.
Keywords
chronic absenteeism, student attendance, receiver operating characteristic (ROC) curves, signal detection theory, Youden’s J
Education level
Topics
Document Object Identifier (DOI)
10.26300/1zvw-qw93
