This study examined the extent to which self-report of human immunodeficiency virus (HIV) risk behaviors might depend upon socially desirable response tendencies, and whether socially desirable responding might serve as a confounding variable in the study of risk behaviors and HIV serostatus. The subjects were 2885 intravenous drug users participating in the ALIVE study in Baltimore, Maryland. Participants completed an interview and were tested for HIV serostatus. The interview covered HIV risk behaviors, and included established scales to measure 'self-deception' and 'impression management', two separate dimensions of socially desirable responding. Seven items for each scale were scored true/false, with a summary score used for analysis. Scores on self-deception and impression management were inversely related to self-reports of sharing injection equipment, injecting at shooting galleries, and injecting more than once a day. Neither self-deception nor impression management was associated with cocaine use. Self-reported receptive anal intercourse was associated inversely with self-deception but not with impression management. HIV serostatus was not associated with social desirability, and statistically controlling for social desirability had a negligible impact on the magnitude of associations between risk behaviors and HIV serostatus. The results indicate that measures of social desirability may be used to ascertain sensitive areas of inquiry in interviews of intravenous drug users.