Comparing accountability models: Outcomes for English language learners by English language proficiency level

The No Child Left Behind Act has led to increased federal requirements and sanctions related to student achievement outcomes. As reauthorization of the legislation approaches, researchers and policymakers have increasingly raised concerns about the appropriateness of current accountability models for all students and in particular for students labeled English language learners (ELLs). Differentiated status models and growth models have been suggested as a less biased and more stable way to account for ELLs' learning on standardized assessments. This study compares growth models and status models to analyze the percentages of students that are making average growth or meeting proficiency for ELLs at different English language proficiency levels compared to other student subgroups. The data analyzed include longitudinal results from a large-scale mathematics assessment for students in 3rd -- 5 th grades from one Midwestern state. Outcomes from the different models are compared to investigate the consistency of conclusions across models for (a) students with matched and unmatched data, (b) student subgroups, and (c) school effects. Results indicated that ELLs who are redesignated as English proficient relatively outperform all other subgroups across all models tested. Also, growth models seem to have relatively more similar outcomes for ELLs and other subgroups compared to status models. Implications for research, policy, and practice are discussed..