English learners' standardized test performance: A longitudinal perspective

A longitudinal study was conducted analyzing the annual standardized test results for Spanish Limited-English-Proficient (LEP) students who had been continuously enrolled for 6 years (Grades 1--6) in Los Angeles Unified School District's (LAUSD) Master Plan for Bilingual Education. The LAUSD Student Longitudinal Database provided the quantitative data used in the analyses. The outcome variables of interest were reading and mathematics normal curve equivalent (NCE) scores from each of the 6 years. 1995--2000. Two predictor variables were used in all of the analyses over the 6-year period: (1) Spanish language proficiency, as measured by the Language Assessment Scales-Oral at grade 1; and (2) bilingual program placement in each of 6 successive years (either full bilingual or modified). The study spanned the years of 1995--2000, which included the changes in the delivery of bilingual education after the passage of California's English Only Initiative, Proposition 227 in 1998.;Oral language proficiency in the native language (L1) at the time of entry into Grade 1 was a significant factor in student achievement throughout the 6 years. In the first grade, all English learners, irrespective of L1 proficiency, appeared to benefit from a full bilingual program placement. Students designated as nonproficient in Spanish (i.e. L1) remained at the lowest levels of student achievement throughout the 6 years when compared with students in the other three categories of Spanish language proficiency. Those who were identified as being proficient in Spanish at the first-grade level continued to reflect superior achievement throughout the 6 years, compared to the other three levels of oral proficiency. Students who had completed redesignation from LEP status to mainstream English showed positive results on standardized tests and appeared to be closing the gap with English-only students over time.;Problems inherent in conducting research on bilingual education are discussed in relation to implementation of large-scale programs, with particular emphasis on the importance of master program placement fidelity. Topics for future research utilizing longitudinal student databases are proposed.