Convergent validity and divergent validity were studied in several basic skill areas using the multitrait-multimethod procedure outlined by D. T. Campbell and D. W. Fiske (1959). The academic skill domains considered were: (1) reading; (2) mathematics; (3) writing and language arts; and (4) spelling. A total of 297 students in grades 1 through 5 served as subjects (65 in grade 1, 57 in grade 2, 59 in grade 3, 60 in grade 4, and 56 in grade 5). The multitrait-multimethod matrix was established using the Stanford Achievement Test (SAT) and some curriculum-based measures. The results do not support specific skill independence, even across such major domains as reading and mathematics using very different methods to measure proficiency. Although convergent validity was found, divergent validity was almost entirely lacking. The findings support a uniform conception of achievement that is globally defined in terms of components and methods. Six tables contain data from the study. (SLD)