How Well Can We Predict Which English Learners Are Most at Risk of Dropping Out?

• In Event: 53.039 - Achieving the Promise of Equal Educational Opportunity for Subgroups of English Learners in Washington State

Abstract

Objectives
The tendency to see ELs as a homogeneous group and inattention to the fate of former ELs after reclassification have restricted educators’ ability to respond to the different needs of subgroups of current and former ELs (Olsen, 2010). As a result, researchers and policymakers have called for new analyses to help identify key EL subgroups and describe differences in their experiences so they can focus resources and target interventions (e.g. Gwynne, 2012). In this study, we examined how graduation and dropout rates vary for five groups of students, classified by their EL status at the start of grade 9; we also compared the accuracy of early warning indicators for each group.

Theoretical Framework
To reduce dropout rates, districts around the country are using early warning indicators to identify and provide supports for students at risk of dropping out, typically relying on some combination of attendance, course failures, grade point average, and suspensions or expulsions (Allensworth & Easton, 2007).

It is not clear whether these commonly used early warning indicators work equally well for a specific student population, English learner students. While national evidence suggests that English learner students drop out of high school at higher rates than other students (Kena et al., 2015; U.S. Department of Education, 2015), these statistics do not capture the variation in dropout and graduation rates for different subgroups of English learner students. Nor do we have sufficient evidence about the accuracy of indicators for these subgroups.

Data Sources & Methods
We drew on five years of data for the 9,595 students (28% current or former EL) who began grade 9 in 2008/09 in six Road Map districts. We classified current ELs as newcomers (entered grade 7 or later) or established ELs (entered grade 6 or earlier); we also broke former ELs into recently reclassified (proficient in grade 7 or later) or long-term reclassified (proficient grade 6 or earlier). For each group, we calculated four- and five-year graduation rates and dropout rates, using state definitions. We assessed the early warning indicators by the Road Map Project using regression and checking the statistical precision and sensitivity of the indicators.

Results
Graduation and dropout rates varied substantially across subgroups. Although everELs had lower graduation and higher dropout rates than neverELs, differences among EL subgroups than between ever- and neverELs. Long-term reclassified ELs had higher graduation rates than neverELs, while newcomers had the lowest rates of graduation.

The project’s early warning indicators were unable to accurately identify many future dropouts across the six districts and were especially poor for newcomers. This is especially problematic since newcomers had much lower graduation rates than other subgroups.

Significance
The sizable variation in graduation and dropout rates among subgroups of ELs underscores the utility of unpacking the category “English learner” and attending instead to crucial variations in students’ immigration and English proficiency development. While the indicators were weak for all students, results suggest a need for alternative indicators for some EL subgroups, particularly newcomers.

Authors

• Theresa Deussen, Education Northwest

• Havala Hanson, Education Northwest