Examined the effects of offender sex, offender status, and consequence severity on accounts following an embarrassing predicament. Subjects were induced to believe they had committed a gaffe with either relatively mild or severe consequences for a confederate/victim of either higher or lower status than they, and their verbal and nonverbal behaviors captured on videotape served as the source of dependent variable measures. Verbal accounts were coded using Schönbach's (1980) account taxonomy. Nonverbal behaviors were also coded, as were measures of subjects'verbal and behavioral helping. Results showed a main effect for sex on account length (p < .001), number of concessionary elements (p < .001), and verbal helping scores (p = .001). Mitigating accounts were proffered more than aggravating accounts. Two-way interactions among sex, status, account type, and severity also were obtained.