A study was conducted to determine the effects of class size on the reading achievement of 517 representative Madison (Wisconsin) students in a three year longitudinal sample. Data included reading achievement, IQ, attitudes toward reading, parents' and teachers' ratings of student interest in reading, sex, age, socioeconomic status, and average class sizes. Scores on the Sequential Tests of Educational Progress (STEP) at the end of the third grade were used as a criterion of final reading achievement. Among the results were the findings that class size is virtually nonpredictive of reading achievement; that only one of the 517 sampled students was enrolled in classes of 20 or fewer students for three consecutive years; and that, when "small" was defined as less than the median class size for each of the three years, there was a slight trend toward lower STEP scores in the small classes, reflecting IQ differences associated with the effects of the special education programs in Madison. Because of confounding with attendance area and special education programs, it was not possible to determine if placing students in small classes, grades one to three, would have any effect on their reading achievement scores. (AA)