Recent reliability studies in which self reports of alcohol consumption taken from diaries have been compared against retrospective recall of alcohol consumption have yielded contradictory results. While the popular wisdom, supported by some excellent studies (Lemmens et al., 1988a; Poikolainen & Kärkkäinen, 1983) is that diaries yield the more accurate account, some studies have shown better results with a retrospective recall (Midanik et al., 1989; Simpura, 1988). The present study of 83 volunteer subjects compares a diary with two retrospective recall procedures--a graduated quantity frequency approach developed at the Alcohol Research Group and a beverage-specific, usual amount procedure used in the U.S. National Health Interview Survey series. Subjects were asked to complete all three procedures. Results show that both of the recall techniques tested achieved an acceptable level of reliability when compared against the results of prospective diaries. The implication of this is that the more easily administered recall techniques may be acceptable for many research purposes.