Reducing Dimensionality for Policymaking: Can We Determine an Optimal Number of Student Absenteeism Measures?

• In Event: Reimagining the Use of Standard Education Data to Signal Economic Mobility

Abstract

Despite the urgency of the absenteeism crises in many countries across the world, policymakers are currently unable to provide concrete, research-based answers for how best to address this. Part of this reason is that the data on reasons for why students are absent have not been fully understood. The purpose of this study is to conduct an analysis around student metrics for absenteeism. We asked: (1) What is the distribution of reasons for absenteeism? (2) Is there an optimal number of reasons for absenteeism to use to predict student performance in school?

This study relies on a partnership with the state of West Virginia in the United States to access their statewide administrative data. Our dataset includes all K-12 students from 2017-18 through 2023-24. Incredibly unique for West Virginia are their metrics of absenteeism. Namely, West Virginia requires all schools to code each absence as one of 20 reasons. These range from discipline to family (emergencies, health, death) to truancy (suspension, weapons brought to school) to illness (parent authorized, doctor authorized, dental), and so forth.

The first research question we answered with descriptive statistics. We were able to tabulate absences across all categories. The second research question was analytic and the focus of our study. We first assessed whether all 20 reasons were equally as predictive on student achievement. Many predicted lower student test scores, but not all 20 did. As a result, we sought to aggregate reasons into fewer comprehensive measures. We adopted both a quantitative approach by conducting principal component analysis and a qualitative approach using thematic analysis to determine the most salient groupings. Our preliminary findings revealed the most efficient way to measure the effect of absences on student test scores was through 4 categories, namely absences that a student accrued in health, family, school events, and unexcused.

Author

• Colby Woods, University of Pennsylvania
  Presenting Author