This study evaluated: (1) the reporting of sociodemographic characteristics of research subjects in published outpatient studies of cocaine abuse pharmacotherapy; (2) the association of study characteristics with such reporting and with the distribution of characteristics; and (3) the comparison of sociodemographic characteristics in the research subjects with those of a community-based sample of cocaine abusers who had sought treatment. Medline search identified 68 articles on cocaine abuse outpatient pharmacotherapy published from 1983 to 1993 in an English language, peer-reviewed journal. Sociodemographic characteristics of research subjects (n = 1802) were compared with those of respondents (weighted n = 135) to the National Comorbidity Survey (1990-1992), who reported at least one cocaine-related problem and had sought substance abuse treatment. Only three (4.4%) articles reported all six of the following sociodemographic characteristics of their subjects: 82.4%, reported mean age; 58.8%, race/ethnicity; 85.3%, sex; 22.1%, employment status; 13.2%, educational status; and 5.9%, socioeconomic status/income. Compared to survey respondents, research subjects were significantly more likely to be African-American and live in the Northeast region of the US and marginally more likely to be male and currently unemployed. These findings indicate that many published articles do not follow currently recommended guidelines for describing sociodemographic characteristics of research subjects and that, aside from race/ethnicity and geographic location, research subjects are fairly comparable in basic sociodemographic characteristics to the larger population of treatment-seeking individuals with cocaine-related problems.