We investigated language variation and score variation in the testing of English language learners, native Spanish speakers. We gave students the same set of National Assessment of Educational Progress mathematics items in both their first language and their second language. We examined the amount of score variation due to the main and interaction effect of student (s), item (i), rater (r), and language (l). We observed a considerable score variation due to the interaction, s x i x l, which indicates that each item has a unique set of linguistic challenges in each language and each student has a unique set of strengths and weaknesses in each language. Parallel results were obtained with samples of students tested across dialects of their native language. (Contains 7 tables, 1 figure and 5 footnotes.)