Past compliance-gaining message selection experiments were reviewed, and three strong predictors of persons' compliance-gaining message choices emerged: (1) benefit to other, (2) negativism, and (3) dogmatism. Two major procedural limitations of these studies were identified: (1) a lack of realism, and (2) a lack of attention to change in the use of compliance-gaining messages. Thus an experiment without these limitations was designed. Subjects (Ss)senl messages to a confederate trying to persuade the confederate to change point allocations on an anagram task. These messages were sent under conditions of either high or low benefit to the confederate (other) and high or low benefit to the subject (self). Individual differences also were assessed. Results indicated that persons are concerned with equity, but that they are more concerned with others being treated fairly than they are with self being treated fairly. While dogmatism did not prove to be a strong predictor of compliance-gaining messages directed toward other, other predictors gleaned from past research correlated as predicted. The benefit self treatment proved to be a stronger predictor of compliance gaining message selection than it has in past studies.