May 2007

What was the purpose of this study? The study sought to determine why California elementary schools serving similarly-challenged student populations vary in their ability to help English learner (EL) students meet California's academic standards. According to California testing data, schools with similar proportions of low-income and Spanish speaking EL students vary by more than 250 points (on a scale of 200 to 1,000) on their EL-API scores. This study sought to explain that dramatic variation or gap in EL-API.

What is an EL student and how many are enrolled in California's public schools?

The California Department of Education (CDE) defines EL students as those who "have not developed listening, speaking, reading, and writing proficiencies in English sufficient for participation in the regular school program." There are approximately 1.6 million EL students in California's public schools (K-12). An estimated 85 percent of California’s EL students are Spanish speaking. California serves one third of all EL students in the nation.

What is the Academic Performance Index (API) and English Learner Academic Performance Index (EL-API)?

An API score is a one-number summary of multiple test scores, with different tests receiving various weights in the index. For elementary schools, students’ results on the California Standards Tests in the core academic subjects of English language arts and math are a major part of the index. All California schools receive an API score between 200 and 1000.

California's new EL-API scores include test results from English learner students only. EL-APIs make it possible to assess schools' progress helping EL students meet state academic standards. EL-API is a new measure, released for the first time by the California Department of Education in the spring of 2006 based on tests taken in 2005.

What makes this study different?

Many studies have examined high-performing schools as a group, in an effort to understand their methods. And most studies of this type take a case study approach, examining a small group of outperforming schools in great detail. This study took a different approach. Rather than studying only schools considered high performers, the researchers surveyed a random selection of schools within a specific student demographic band, and then used a regression analysis to determine which policies and practices were more common at high performing schools than at low performing schools. In addition, this study’s large sample of schools and unusually high participation rates from principals and K-5 teachers sets it apart. This methodology succeeded in identifying practices associated with high performance among EL students. 

How does this study differ from 2005’s Similar Students, Different Results: Why Do Some Schools Do Better?

This new study analyzed EL-API (English learners only), while the previous study analyzed school-wide API (all students). Both studies were based on the same survey data file. This study also analyzed some survey questions about specific EL practices that were not studied originally, and also analyzed all the responses against a measure of student English language proficiency. 

What were the study’s key findings?

The study identified four broad practices associated with higher EL-API scores: using student assessment data; ensuring access to good teachers and instructional resources, closely aligning the curriculum with state academic standards, and setting measurable and ambitious goals for student achievement. 

One of the key practices identified was using student assessment data to improve student achievement and teacher instruction. What is this?

Principals from better performing schools more often reported that they and the district use assessment data from multiple sources to evaluate and improve teacher instructional practices. Principals who reported frequently and personally using assessment data to address the academic needs of students in their schools led, on average, higher performing schools. They report using these data to develop strategies to help selected students reach goals and to follow up on the progress of selected students. The principals report that their districts expect that all schools in the district will improve student achievement and evaluate principals based upon student achievement.

Were there any findings related to specific EL instructional practices?

The study found that extensive exposure to the school’s core academic curriculum, when accompanied by assistance to ensure that the curriculum was accessible to EL students, was associated with higher performance. "Pull out" programs, in which students are briefly removed from the classroom to receive explicit English Language Development instruction from resource specialists, were also associated with better student outcomes. The increased presence of teachers in a school with a CLAD or BCLAD certificate was not associated with higher EL-API.

Both this study and the 2005 study found that parental involvement was not as significantly correlated with higher achievement as the four practices previously identified. Does this mean parental involvement isn't important?

No. A school’s outreach to parents, encouragement of teacher collaboration, and enforcement of positive student behaviors (like attendance and tolerance) have long been recognized as important contributors to the student and professional culture at a school and to community engagement. Our analyses indicate that while these practices contribute to student achievement, they are not the most critical features that differentiate higher- from lower-performing schools.

In summary, it is not correct to conclude from this study that parental involvement does not contribute to EL-API performance. However, the study did conclude that other school practices more directly related to academics are more important.

Did this study evaluate the effectiveness of bilingual education programs?

No. Bilingual education has been significantly curtailed in California since the passage of Proposition 227 in 1998. Although parents wanting such an instructional setting may request it; only 8% of the state's English learner students are currently taught in bilingual instructional settings (compared to 29% before the passage of Proposition 227).

Which schools were included in the study?

The study selected elementary schools in the 25th to 35th percentile of the California School Characteristics Index (SCI), a measure of student demographics. Schools in this SCI "band" generally serve a high percentage of low-income students. The study’s sample for this analysis included 237 elementary schools from 137 different school districts across California. The EL populations at the schools were primarily Spanish speaking, and ranged from 17% to 80% with the median being 42%.

Only traditional public schools were included in the survey. Charter and private schools were not included.

Can the survey's conclusions be applied to students in schools outside of the 25th to 35th percentile of the School Characteristics Index?

Because the researchers focused solely on schools within the 25th to 35th SCI band, results cannot be generalized to schools outside that band. Educators working in schools with similar student populations may, however, want to consider whether the practices this study found to be correlated with higher EL-API might also work for their schools and students.

What is EdSource?

EdSource was established in 1977 as an independent, impartial, not-for-profit organization whose sole mission is to clarify complex California education issues and to promote thoughtful decisions about public school improvement. EdSource does not advocate or lobby and has developed a solid reputation as a credible and respected source of K–14 education information.

Who was on the research team that conducted the study?

The research team consisted of a group of experts from EdSource, Stanford University, the American Institutes for Research, and WestEd. Team members from EdSource included Trish Williams (Study Project Director), Mary Perry, Isabel Oregon, and Noli Brazil. Team members from Stanford University were professors Kenji Hakuta (Principal Investigator), Edward Haertel (Senior Technical Consultant), and Michael Kirst (Policy Consultant). American Institutes for Research’s Jesse Levin served as principal data analyst, and WestEd’s Robert Linquanti served as an advisor to the research team. The complete biographies of research team members are available. Electronic copies of Similar English Learner Students, Different Results: Why Do Some Schools Do Better? are available for download as a PDF file. Additional materials, including technical and other appendices and media materials, are also available for download.

For more information:

• Study findings      
• Press release