The effects of two components of formative evaluation, frequency of measurement and use of data, were compared. Fifty-two learning disabled and educable mentally retarded students in grades 2-6, enrolled in regular class programs and receiving reading instruction in a special education resource room, were randomly assigned to either (1) a pre-posttest non-data-based change group; (2) a daily measurement non-data-based change group; (3) a daily measurement data-based change group; or (4) an untreated control group. Four types of data were used: oral reading rate correct, oral reading rate incorrect, word meaning, and comprehension. The first three measures were derived from the Power Builder Kits (Science Research Associates); the last was taken from the Stanford Diagnostic Reading Test. Group 3 had the highest oral reading correct rate, after the four week program. There were no significant group differences in word meaning or comprehension. Superior achievement occurred when teachers measured oral reading daily relative to daily goals and altered both goals and consequences, based upon student performance. The effectiveness of traditional pretests and posttests for formative evaluation was questioned. (Author/CP)