A recent trial (n=508) of "Making Alcoholics Anonymous Easier" (MAAEZ), a group-format 12-step facilitation program, showed that MAAEZ participants had increased odds of abstinence (OR=1.58; p=0.063). Effects were especially marked in several subgroups, including those with more prior AA/NA/CA exposure, and those with severe psychiatric problems. This paper examines whether the effects of MAAEZ were explained by higher engagement in particular Alcoholics Anonymous (AA) and other 12-step organization activities. Mediation analyses were performed, estimating MAAEZ effects attributable to AA/NA/CA meeting attendance, overall AA/NA/CA involvement, having a sponsor, and engaging in 12-step service. The only variable that appeared to mediate MAAEZ effects in the sample overall was doing service at either 6 months or at both 6 and 12 months. Among those with high prior AA/NA/CA attendance, both having a sponsor and doing service emerged as mediators, with having a sponsor explaining approximately 25% of the MAAEZ effect. Doing service also partially explained the MAAEZ effect among those with high psychiatric severity. Results highlight the value of studying specific AA/NA/CA activities as mediators of treatment effects, and demonstrate the importance of exploring subgroups when trying to understand treatment mechanisms. Findings also suggest that treatment programs should emphasize specific activities for particular clients. For example, individuals may have attended many AA/NA/CA meetings, but never known how to ask for a sponsor or get involved in service; doing service may represent a non-threatening way of connecting with 12-step members for clients with psychiatric problems such as social anxiety.