Recent validiry studies of Jourard's Self-Disclosure Inventory (SDI) have failed to confirm the validity of this instrument (Himelstein & Lubin, 1965; Pederson & Breglio, 1968). These studies have suggested, however, that reported self-disclosure and actual self-disclosure may reflect behaviors sufficiently different to necessitate diffetent measurement operations. Furthermore, it appears that the measurement of actual disclosing behavior must account for at least rwo parameters of self-disclosure, amount and intimacy. The SDI confounds these two important parameters, which may partly account for the fact that no consistent relationship has been demonstrared bemeen the SDI and actual disclosure. Vondracek (1968) has reported the development of a behavioral measure of selfdisclosure, and its utilization in a study of interviewing techniques (1969). The present study was undertaken in part to explore the relationship of SDI scores to a behavioral measure of self-disclosure. Sixty-four male undergraduates participated in structured interviews which were presented to rwo judges as typewritten transcripts. The judges were trained to rate the transcripts on a scale ranging from 1 (low intimacy) to 7 (high intimacy). Amount of self-disclosure was measured by timing the actual verbalizations of each S. The SDI was administered after the interview. The correlation between the measures of amount and intimacy of self-disclosure was .42. Although this is a fairly high correlation, it accounts for less than 25% of the variance. This appears to justify the separate measurement of these two parameters of selfdisclosure. Scoring amount of self-disclosure, as described above, is highly reliable; repeated measurements on the same tape resulted in virtually 100% agreement. The judges of intimacy achieved interjudge reliabilities berween .71 and .76, which indicates that these judgments can be made quite reliably. Correlations between the SDI and the measures of amount and intimacy of self-disclosure were .20 and .14, respectively. The fact that these correlations are relatively low and not statistically significant appears to confirm the hypothesis that a self-report measure of self-disclosure and a behavioral measure assess different variables. In view of the importance of disclosing behaviors in psychotherapy, counseling, and interpersonal relations in general, a reliable behavioral measure of this variable is long overdue.
Read full abstract