A discriminant analysis of the relationships between 335 Florida students' first and second grade scores on the reading subtests of the Metropolitan Achievement Test (MAT) and their third grade scores on the Florida Statewide Assessment Test (FSAT) was undertaken, using race, sex, and intelligence as independent variables. Three null hypotheses were tested, suggesting that no significant relationships would be found between first or second grade MAT scores and FSAT results. Discriminant function analysis was found to be an inappropriate statistic for one of the nine FSAT standards (basic vocabulary), because only two children in the entire sample did not achieve mastery of it. A number of significant relationships were found between FSAT standards and student achievement scores after grade one, after grade two, and after both grades; therefore, all three null hypotheses were rejected. Since MAT subtests for word knowledge and word analysis significantly predicted meeting the FSAT standards, it was advised that instruction in the various subskills of these predictor variables be included in Florida schools. (RL)