Researchers, policymakers, educators, and journalists have focused copious amounts of attention on student absenteeism data in the years since a pandemic-era surge that is only now beginning to moderate. Of particular concern is chronic absenteeism, nearly universally defined as missing at least 10 percent of school (generally 18 days) for any reason, and the potential academic harm missing so much school could cause. Research protocols, early warning systems, accountability structures, and considerable resource allocations to identify and reduce chronic absenteeism have all been calibrated with this definition as their guiding principle. However, a trio of researchers from the University of Michigan hypothesize that other measures might be more accurate in predicting negative academic outcomes—and indeed might reset the definition of “chronic absenteeism” in that regard. They set out to test their theory in a new working paper.
The researchers use data on nearly 8,900 students who were enrolled in the Boston Public Schools (BPS) pre-K program starting in the 2007–08 to 2009–10 school years. Students were followed through their eighth grade year (which occurred between 2016–17 and 2018–19). Administrative records from district and state sources include demographics, details of attendance and absences, and English language arts and math standardized test scores. The analytic sample was 48 percent female and racially/ethnically diverse (45 percent Hispanic, 29 percent Black, 15 White, 9 percent Asian, 3 multiracial/other). Roughly half of students spoke English as their home language, 30 percent Spanish, and 21 percent a language other than those. Approximately 58 percent of students were eligible for free or reduced-price lunch.
The researchers used the administrative absences data provided by BPS—total days absent, total unexcused absences (like oversleeping or going to work), and total excused absences (like family vacations or a dental appointment)—as well as calculating a novel student absence rate. That is, dividing each student’s number of total absence days by the total days the student was enrolled during a given year. This helped account for variation in enrollment length for students transferring into the district at different times of the year.
The researchers then evaluated the diagnostic accuracy of six different absenteeism measures over time—total days excused, total days unexcused, total days for all reasons, excused absence rate, unexcused absence rate, total absence rate—in predicting whether a student would score at the lowest category of achievement (“not meeting expectations”) on Massachusetts’ state math and ELA assessments in eighth grade. They found minimal predictive value of absenteeism rates in the early grades, when rates are largely driven by excused absences…and a long way from eighth grade test day! But by the upper elementary grades (when unexcused absences began to predominate), the predictive strength steadily increased. Perhaps not surprisingly, eighth grade absenteeism data was most predictive of eighth grade test performance. Across all years, the researcher-created absence rates and the BPS total-days-absent measures proved the most predictive of academic outcomes.
And what of “chronic” absenteeism? Their analysis found that—rather than the traditional 10 percent of the school year/18-day cutoff that generally denotes chronic absenteeism today—academic harm from absences became highly likely closer to the 7 percent of time enrolled point for most students and ranged from six to 17 days, depending on grade level, test subject, length of time enrolled, and absence measure used. In other words, the default definition of “chronic” absenteeism is likely too generous. More importantly, the warning sign for schools to intervene to mitigate academic harm from chronic absenteeism comes about quite a bit sooner for many students than schools currently believe.
In an interview, research team leader Tiffany Wu says that “no single absence cutoff is likely to serve as a strong standalone predictor of academic risk.” But the team’s report recommends schools track both absence rates and total absences—if staff resources allow—and that leaders and policymakers seriously consider lowering the cutoffs that define “chronic” absenteeism no matter how it is measured. Because data show that early warning systems should be signaling that intervention is needed far sooner than they do now in order to prevent maximum academic harm.
SOURCE: Tiffany Wu, Christina Weiland, and Thomas Staines, “The Chronic(les) of Absenteeism Measurement: Unpacking the Many Measures of Attendance and Evidence for a Lower Chronic Absenteeism Threshold,” Annenberg Institute at Brown University Working Papers (January 2026).
