To test the effectiveness of the minicourse (an instructional microteaching package) in changing specific teacher behaviors, 20-minute pre- and postminicourse video-taped recordings of each of the 48 participating teachers' classroom lessons were made and were scored by trained raters. Further, to insure rater objectivity, delayed postcourse video tapes were mixed with pre- and postminicourse video tapes from another study and scored. Results of analyses of pre- and postminicourse scores showed that teachers made significant gains after the minicourse on 10 of 12 behavior scores and demonstrated a reduction to half the precourse level of teacher talk. Additional analyses of these data showed that when the sample was divided according to teacher grade level and compared on four behaviors relating to teacher talk and pupil response, teachers in all grade levels increased their use of higher cognitive questions, and students increased the length of their responses; when the sample was divided according to middle and lower class school setting, teachers serving lower class areas made greater gains on most of the skills; and when the sample was divided by sex, there were found to be no significant differences in each group's learning of teacher skills. Two months after completion of the course, a refresher course was given to one third of the group. Results of a posttest administered two months later showed no significant differences between this group and the rest of the sample, indicating that the teachers had retained most of the skills acquired in the minicourse without a refresher course. (SM)