Eighty four female undergraduate students participated in a psychotherapy analog study to determine the effects of imagery ability, guided imagery therapy treatments, and personal versus supplied constructs upon self-disclosure variables in a 2 x 3 x 2 Anova design, with repeated measures on the final factor. Dependent variables were measured by reaction time, total talk time, speech duration, silence quotient, and Doster's (1971) Self-Disclosure Rating Scale. Subjects were divided into two imagery ability levels on the basis of local mean scores on Sheehan's (1967) modification of Betts' (1909) Questionnaire upon Mental Imagery. Three treatment procedures were employed: a guided focal imagery treatment, which encouraged imagery involving the interpersonal topics to be discussed, a guided relaxation imagery treatment which used standard sensory relaxation scenes, and a treatment which imparted ambiguous instructions. The final factor was repeated measures of the eight negative topics the subjects were asked to discuss. Four were chosen from the subjects' Role Construct Repertory Test grid (Kelly, 1955; Landfield, 1971), and four were selected from the Semantic Differential (Snider & Osgood, 1969).