Similar English Learner Students, Different Results: Why Do Some Schools Do Better?
May 2007
How elementary schools focus their time and energies, and what resources they have for doing it, can make a powerful difference in the academic achievement of English Learner students from low-income backgrounds, according to findings from this new analysis of data.
This new extended analysis was based upon extensive survey data from 4,700 K-5 classroom teachers (80% or more at each school) and all principals in 237 California elementary schools from 137 different school districts across the state. These schools were initially randomly selected from 550 schools in California’s 25-35% School Characteristics Index band. All schools from this band have high levels of student poverty and low parent education levels; for this analysis we further narrowed our original sample to eliminate any school that didn’t have enough English Learner students to have an EL Academic Performance Index score.
The research team analyzed the school practices covered by the teacher and principal surveys to see which most highly correlated with California’s new school level English Learner Academic Performance Index. In addition, the team analyzed the same practices against percent proficient on the California Standards Tests to see if the results were similar. Finally, the team ran an additional analysis to see if the results were similar for only schools in our sample with English Learner student populations that were 80% or more Spanish speakers. The results for all three analyses were essentially the same: there are four interrelated broad school practices – backed up by numerous examples of specific actionable practices – that most strongly differentiate the lower from the higher performing elementary schools with regard to English Learner API. These four practices are the same, although in a slightly different order of significance, as the team had found in October 2005 for the school-wide API.
This new analysis also included a small subset of questions about specific EL instructional practices and teacher qualifications that were not included in the previous Similar Students analysis. Not a comprehensive set, these questions served as a “pilot” to explore what practices and conditions are worth further exploration, and provided some interesting findings, some of which support oft cited recommendations for EL instruction and some of which don’t.
This extended analysis was based upon the original analysis and Initial Findings of the Similar Students, Different Results study. Those findings and their technical appendices can be found at the Similar Students page.
This extended analysis on English Learner academic outcomes in California was overseen by EdSource and conducted by researchers from Stanford University (Kenji Hakuta, Principal Investigator; Edward Haertel, Technical Consultant; Michael Kirst, Policy Consultant) and the American Institutes for Research (Jesse Levin, Principal Data Analyst). Robert Linquanti, Senior Research Associate at WestEd, served as Advisor. We thank the S.H. Cowell Foundation for their support of this extended analysis and its dissemination.
E-mail this page to your friends and colleagues.
Trouble reading this page? Our site is best viewed in IE 5, Safari 1, Firefox 1 or AOL 9 or higher.