Approximately 1,700 English language learners (ELLs) and former ELL students, in Grades 4 and 10, were tested using both an English-language (Stanford Achievement Test, 9th ed.) and a Spanish-language (Aprenda, 2nd ed.) achievement test. Their performances on the two tests were contrasted. The results showed that ELL students, for the most part, answered more items correctly on a home-language mathematics test, compared to a similar English- language math test, regardless of their level of home-language literacy. Additionally, former ELL students are often unable to exhibit their content-area knowledge on English-language achievement tests, possibly due to language and cultural barriers. In summary, the results show that the achievement test results of ELL students, when tested in English, are not always valid measures of their content-area knowledge. (Contains 6 tables and 3 figures.)