|
Abstract:
|
English Language Learners are a growing share of U.S. public school students yet little is known about their academic achievement. Previous literature has focused primarily on the effects of different English language tests and test modifications on ELL performance. In contrast, this paper asks the following questions: What school-level characteristics are related to average math achievement test scores for students tested in English? Are the same relationships found for students who are tested in Spanish? Essentially, this paper investigates how school characteristics affect math test scores of English learners compared with how those same characteristics affect traditional students.
Using school-level data from Texas that includes both Spanish-language and English-language math assessment scores for 3rd, 4th and 5th grades, and employing both random effects and clustered models, this paper shows that these school characteristics explain very little of the variation in Spanish language test scores. It appears that none of the characteristics explored herein have the same affect on ELLs as they do on traditional English-fluent students. Differences are found in both standard errors and coefficient sizes. In addition, the R-squared values are much larger for the English models in all grades than for the Spanish language models. The largest differences in coefficients are found in the effect of percentage ELL students in a school, teacher experience, and the percent of economically disadvantaged students in a school. Finally, the event of a school meeting the state proficiency standard on the English language math assessment has a strong positive effect on the likelihood of meeting the state proficiency standard on the Spanish language math test. |