This research examined the hypothesized interaction between sex and interpersonal relationship by comparing the actual self-disclosures made to best friends and strangers by male and female subjects. It was hypothesized that subjects would disclose more to best friends than to strangers, that females would disclose more than males to best friends, and that males would disclose more than females to strangers. Eighty introductory psychology students from the University of Manitoba served as subjects to help satisfy a course requirement. Neither hypothesis was confirmed. There was a significant effect for greater female self-disclosure, but only with the measure of retrospective self reports. The self-disclosure literature, upon closer look, reveals the same pattern: Only retrospective self-report paradigms, as opposed to paradigms of predicted self-disclosure, actual self-disclosure, or evaluation of hypothetical situations, show the sex effect. This is discussed in terms of various methodological issues and of the sex-role stereotype for higher female self-disclosure. (Author)