Test validity was studied in nine subtests of a kindergarten screening battery used to predict reading comprehension for children up to five years after entering kindergarten. The independent variables were kindergarteners' scores on the: (1) Otis-Lennon Mental Ability Test; (2) Bender Visual Motor Gestalt Test; (3) Detroit Tests of Learning Aptitude (DTLA) Visual Attention Span for Objects; (4) DTLA Auditory Attention Span for Unrelated Words; (5) Elenbogen Test of Auditory Discrimination; (6) Word Recognition (deHirsch et al.); (7) Word Reproduction (deHirsch); (8) Wide Range Achievement Test (WRAT) reading achievement; and (9) WRAT arithmetic achievement. Gender and intelligence were predictor variables; test battery scores were compared with the dependent variable, the Stanford Achievement Test (SAT) reading comprehension subtest for grades one through five. Subjects were 245 suburban students. The Reading Comprehension subtest of the SAT was administered every spring through the fifth grade. Multiple regression analysis was performed for each grade and sex, generating prediction equations for SAT scores. Gender did not generally correlate significantly with other variables, including WRAT reading and mathematics. Results indicated significant relationships between the independent variables and reading comprehension. Stepwise regression determined the most important variables at each grade level. A bibliography and numerous tables are included. (GDC)