This study sought to determine the convergent and discriminant validity of the PIAT when administered to separate samples of Anglo and Mexican-American children. Thirty-one Mexican-American and 31 Anglo-American children were matched on the basis of sex, age, SES, educational placement, and reason referred for psychological assessment. All children were administered the PIAT, Wide Range Achievement Test (WRAT), and Wechsler Intelliegnce Scale for Children-Revised (WISC-R). Partial intercorrelations (CA held constant) among the PIAT subtest raw scores and subtests of the WRAT and WISC-R purporting to measure the same construct were performed. Adequate convergent validity was found for the PIAT Mathematics, General Information, Spelling, and Reading Recognition subtests for both Anglo and Mexican-American children. Poor discriminant validity was obtained for all PIAT subtests evaluated. It was concluded that the PIAT may be of value when used in conjunction with the WISC-R to differentiate between specific learning difficulties and an English language deficiency with Mexican-American populations. (Author)