The first step to solving a problem is identifying that you have one. For schools using early-warning systems, though, the problem is identification isn鈥檛 enough.
Early-warning systems use student data to predict which students are at risk academically and intervening to keep them on track. A new study published by by the Center for Education Policy Research at Harvard University looked at one large, unnamed school district鈥檚 data system and found that it worked to improve attendance鈥攂ut not for low-income students and not before students developed chronic absenteeism.
Student-data systems can be to install and manage, and they require significant staff time and security to protect the student data they collect. But more than a decade of research shows changes in attendance, along with behavior problems and course grades, can together predict a student鈥檚 risk of dropping out of school as early as elementary school.
That鈥檚 why a large majority of school districts nationwide have adopted these systems鈥攎ost recently with an eye to reducing chronic absenteeism, which has risen nearly 75 percent nationwide since the pandemic.
Yusuf Canbolat, a strategic-data fellow at CEPR, tracked the effects of one such early-warning system in a large urban district for the study, published this month in the journal Educational Evaluation and Policy Analysis. He found that the system was able to identify students with severe attendance problems鈥攊n other words, those who are missing more than 10 percent of class in a school year. But the system only led to improved attendance for high-income students.
鈥淥ne of the important assumptions of the early-warning systems is to create this organizational culture around well-identified problems and tailor solutions based on that data,鈥� Canbolat said. 鈥淢y study indicates that it may not be true always when it comes to some persistent issues like chronic absenteeism. Data is important, but data alone does not solve the problem.鈥�
For example, while the data system flagged students who missed 4 percent to 10 percent of class time, it only led to attendance improvements for the students with the highest rates of absenteeism, suggesting that schools did not intervene effectively鈥攐r enough鈥攖o make these students attend more consistently. While the study did not look at why the warnings led to improvements for only some students (a follow-up study is examining this), Canbolat thinks in part, the sheer number of chronically absent students may have led school staff to triage interventions to the most frequently absent students instead of resolving more minor absenteeism problems before they escalated.
This can be a common, but less effective approach to using the student data, said Marcia Davis, who was not part of the Harvard study but who has conducted several evaluations of early-warning systems. Students who have multiple warning flags often are already on educators鈥� and administrators鈥� radar, she said, but using student data to find the roots of attendance problems can allow educators to help larger numbers of students.
Davis, an associate education professor at Johns Hopkins and co-director of the Center for Social Organization of Schools, said such systems can make a difference in student performance and engagement but only when coaches actively help teachers understand and use the student data to find more 鈥渜uietly struggling鈥� students. In Davis鈥檚 own studies of early-warning systems in several states, effective coaches worked with a teaching team to analyze student data and plan interventions but also checked in with students directly. 鈥淚 followed some of the coaches around,鈥� Davis said. 鈥淭hey knew all the kids鈥� names. Students would be coming up to them for help.鈥�
Similarly, Canbolat鈥檚 study found high-income, chronically absent students flagged by the early-warning system had significant improvements in attendance, but there was no benefit for lower-income students.
鈥淒isadvantaged students have higher levels of absence, [and] the reasons for their absence are more difficult to influence,鈥� Canbolat said, noting that low-income students are more likely to work or care for siblings, less likely to have access to health care, and more likely to have unstable housing or transportation. 鈥淎 system may identify well, but if the solution does not fit the problem or that identification, we may not observe the expected outcome.鈥�
Fine-tuning systems
The current study looked only at one large school district, but the results align with other research in the United States and .
A consortium of eight early-warning research groups, including Johns Hopkins and the UChicago Network for College Success, have been developing new standards for early-warning systems since 2022. The EWS 2.0 consortium advised schools to couple early-warning alerts with ongoing analyses of root causes of common student problems. For example, the warning system in the current study logged missed instruction on a broad level but not whether absences were excused or unexcused nor underlying reasons, such as family vacation, health conditions, or other situations, Canbolat said.
Among the consortium鈥檚 recommendations:
- Identify how school policies and practices can stop or speed up the decline of students who are starting to slide off track.
- Identify the classes, grades, or times when interventions could help the most off-track students.
- Focus on interventions that build supportive relationships between adults and students who have become disengaged.
Some researchers are also exploring how data systems could better fill in the gaps. One of a Brazilian early-warning system has identified data on students鈥� relationships with peers, family, and school staff that could help educators target support.
