The experiment was designed to test the effects of covert behavior upon the verbal expression of emotional words contained in a TAT-like story. Subjects were 63 college students. The first independent variable, "self-regulation," consisted of three levels: (a) a no self-monitoring condition, (b) a self-monitoring condition, and (c) a self-monitoring with self-evaluation condition. The second independent variable, visual imagery, involved the administration of positive, neutral, or aversive images contingent upon an actual or perceived increase in the target response. The data were analyzed using a 3x3 factorial design in a multivariate analysis of covariance and a repeated measures analysis of covariance. No significant differences were found on the major hypotheses, but a remarkable trend in the interaction graphs stimulates a number of specific suggestions to further researchers. (Author)