The study examined whether the California Achievement Test (CAT) results would be affected if the test instructions were administered in the language the children best understood. In a pre-post-test sequence, 77 Mexican American second graders between the ages of 7 and 8 of a low socioeconomic background attending the same elementary school in west San Antonio were administered the CAT Reading Section, Level I, Form A. In the pre-test, the individual classroom teachers gave the instructions only in English. In the posttest, 28 children received the instructions in Spanish, 25 in English, and 21 in the local Spanish dialect. Language dominance of the posttest subjects was informally rated by their teachers. Pre- and posttest scores were subjected to Pearson correlations and to analysis of variance. Findings included: there was a significant difference of .008 in favor of the posttest group receiving instructions in English on the vocabulary subtest; the group receiving instructions in the local dialect omitted fewer items than the English or Spanish group on the reading subtest; significant differences were obtained in favor of the English group in the vocabulary and reading subtests; and an analysis of the gains for each group revealed that, in general, higher gains were obtained by the group receiving instructions in the local dialect. The instructions in Spanish and in the local dialect and the language dominance rating instrument are appended. (NQ)