Most 12-step research recruits participants who are seeking treatment (cf. Kaskutas, Turk, Bond, & Weisner, 2003; Robinson, Cranford, Webb, & Brower, 2007), leaving open the question of how non-treatment-seeking individuals respond to 12-step involvement. The current study examined whether participants recruited from community-based Alcoholics Anonymous (AA) clubs or via advertisement/word-of-mouth differed in their 12-step attendance rates, substance use, and the association between these constructs compared to participants recruited from substance abuse treatment. Multilevel analyses showed that while associations between 12-step attendance and improved substance use did not differ by recruitment group, participants recruited from community-based AA clubs had higher 12-step attendance rates and a higher proportion of alcohol abstinent days than did all other participants. Results indicate that using diverse recruitment strategies may be necessary to obtain samples that are representative of the actual population of 12-step affiliates and to achieve results that estimate the true magnitude of 12-step effects.