The increasing use of media advertisements to recruit subjects in alcohol treatment research has raised concerns about the comparability of such solicited subjects and subjects drawn from more traditional clinic populations. This paper, based on a study that involved both solicited and clinic subjects, identifies differences between the two types of subjects in terms of sociodemographic variables; pre-treatment drinking and marital adjustment; several dimensions of 'self-definition'; and three outcome measures. These findings suggest that there may be significant differences between solicited and clinic subjects and that these differences should be taken into account in the design of studies utilizing solicited subjects in order to determine the generalizability of the results.