A fourth grade class (49 students) with a mean grade equivalent score of 2.8 on the Stanford Diagnostic Arithmetic Test was placed in a Sequential Computational Skills Math Program. The program consisted of 12 units (from basic addition and subtraction facts to fractions computation), and individual Ss' entry levels were determined by computation placement tests. The first and third phases of the study were experimental conditions, during which problems were completed using a computational approach; the second phase was the baseline condition, during which problems were completed according to a conceptual approach. From September to April, the number of Ss working on fourth grade material increased from 13 to 48. Additionally, rate of problems completed was significantly lower during the conceptual phase, the mean increase in math achievement was 2.5 years, and students' attitudes toward math noticeably improved. (LS)