An early warning system uses student data and predictive analytics to identify students who may go on to drop out or not graduate on time. Administrators employ early warning systems (EWSs) to help interventionists (e.g., social workers, counselors, specialists) proactively support students to an on-time graduation. While EWSs differ significantly from institution to institution, EWS models include the following core components:
EWSs predict either dropout or on-time graduation.
An EWS flags a student as at-risk when their performance in a variable (e.g., behavior) reaches a certain threshold. For example, an EWS may flag a student as at risk after three discipline referrals.
Some EWSs do not use uniform thresholds. For example, the EWS at Department of Education A uses multivariable regression.
Administrators often use attendance, behavior, and course grades—known as the ABCs—to predict outcomes.
Administrators use student data from past years at their district to determine which variables best predict dropout or delayed graduation.
EWSs assign each student a risk status based on the system’s variables and thresholds/analytics.
EWSs usually use three risk levels: high risk, medium risk, or low risk—often coded as red, yellow, and green.
EAB interviewed contacts at four school districts, two state departments of education, and a non-profit to understand how institutions in various administrative positions operate EWSs. The profiled nonprofit, Nonprofit A, operates an EWS in 19 schools.
Predictive mechanisms identify at-risk students
Most districts build early warning systems that use attendance, behavior (as measured by discipline incidents), and course grades to estimate the likelihood that a student will go on to dropout. Based on that likelihood, administrators assign a risk status to students. Administrators target intervention resources toward high-risk status students.
At one district, administrators built an early warning system that helped lower district dropout rate from 5.4% to 3.5% over ten years.
Qualities of a perfect EWS
Perfectly sensitive
The EWS identifies all the students who will go on to drop out or not graduate on time. The model does not fail to identify any student who goes on to drop out or graduate late.
Perfectly specific
The EWS only identifies students who will go on to drop out or not graduate on time if no one intervenes. This model does not flag students who will not go on to drop out (given no intervention).
Early warning systems inform school-wide intervention
While administrators design early warning systems (EWSs) primarily to identify and support at-risk students, EWSs can also inform broader strategy. EWSs can help identify how schools fail to keep their students on track for graduation. In response, district administrators can provide extra support to those schools and deliver targeted school-wide interventions aimed at supporting students to graduation.
Intervention effectiveness with students
at high risk at District B
Data from the 2015-2016 school year
No intervention
21.1%
Dropout rate
Intervention
10.8%
Dropout rate
24+ Intervention contacts
7%
Dropout rate
4-5
Early warning systems promote resource efficiency
An early warning system can effectively identify which students and schools would benefit most from intervention. As a result, administrators can optimize the use of district resources to achieve maximum impact.
Interventionists who work with students at high risk can prevent four to five times as many dropouts as those who work with low-risk status students.
This resource requires EAB partnership access to view.
Access the research report
Learn how you can get access to this resource as well as hands-on support from our experts through District Leadership Forum.
Learn More