Toward an equitable NAEP for English language learners: What contextual factors affect math performance

This study examines the validity of the national math assessment, known as the National Assessment of Educational Progress (NAEP), for English language learners or students with limited English proficiency (LEP). Secondary analyses of two existing CSE/CRESST datasets were conducted. Data samples were comprised of predominately Hispanic, 8th grade English language learners in southern California. Descriptive and hierarchical linear modeling analyses suggest that several variables influence NAEP math performance. Quantitative analyses were supplemented by informal interviews with a small sample of middle school math teachers.;In both studies, students' English reading proficiency and level of math class significantly predicted test performance, even after controlling for key student background and classroom-level variables. When these variables were held constant, differences in test performance between LEP and non-LEP students were no longer evident. Several proposed test accommodations for LEP students were also examined: Modified English, Spanish Translation, Glossary, Extra Time, and a combination of the Glossary and Extra Time. The data suggest that, after controlling for key confounding variables, selected test accommodation formats have potential for "leveling the playing field" between students with limited English proficiency and their more English fluent counterparts.;Analysis of cross-level interactions in the final analytic model suggest that specific test accommodations may benefit LEP students under certain testing conditions. Students tend to perform better on a NAEP math test where the language matches their language of math instruction. More specifically, LEP students who received math instruction in Spanish were more likely to score higher if the language of the math test was also Spanish, rather than English. This effect was more profound for students in lower-level math classes. Additionally, LEP students who received an Extra Time accommodation benefited more than with other test accommodations, especially students in higher-level math classes. These findings suggest that no single math test accommodation benefits all LEP students.;The data suggest that accommodations yield only nominal effects for LEP students, and that accommodation effects vary by level of math class and language of math instruction. Additional methodological complexities arise as the test formats are no longer standardized. Nonetheless, findings may be used to inform the development and interpretation of the NAEP for English language learners, as well as other large-scale, standardized content-based assessments.