This study proposed that a principal's behavior is affected by the feedback he receives from teachers and by his own commitment to change his behavior. The behaviors of 206 elementary school principals (volunteers) were rated by teachers at the beginning and end of the study; the teachers also used the rating scale to describe an ideal principal. The principals were then placed in one of four feedback groups: the first to receive both "actual" feedback ratings of their own behaviors) and "ideal" feedback (ratings of an ideal principal); the second to receive only "ideal" feedback; the third, only "actual"; and the fourth, no feedback. Moreover, the principals in each group either had not been asked to commit themselves to change or had been asked to choose one of two areas in which to commit themselves--task assistance behaviors or personal support behaviors. It was hypothesized that group 1 would change more positively (approach the ideal) than the other groups and that group 2 would similarly surpass groups 3 and 4. It was also hypothesized that commitment would cause more positive change than no commitment, particularly in the behavioral area selected by the principal. Although covariance analysis of scores did not support the hypotheses, chi square analysis and change patterns suggest that feedback, especially "ideal" alone or "ideal" and "actual", promote positive change and that "actual" feedback alone and solicited commitments to change may inhibit such change. (LP)