Alcohol use disorders represent a major public health problem in the United States. Approximately 7.4% of the U.S. population meet the diagnostic criteria for alcohol abuse or dependence (Grant et al., 1994), a proportion that results in substantial costs to the individual and society, among them elevated morbidity and mortality rates, traffic accidents, injuries, crime, broken families, domestic violence and fetal alcohol syndrome (Gmel and Rehm, 2003; Greenfield, 1998; Harwood, 2000; Hingson and Howland, 1993; Hingson and Winter, 2003; Theobald et al., 2001). Recent years have witnessed an explosion of new treatment approaches for alcoholism. Pharmacotherapies such as naltrexone and acamprosate appear to reduce cravings, urges to drink and the reinforcing qualities of alcohol (Miller et al., 2003). Many new behavioral interventions have also been developed: brief counseling, behavioral self-management, relationship therapies, coping skill training, insight and exploratory therapies, confrontational approaches, milieu therapy, educational programs and 12-step facilitation (Miller et al., 2003). Pharmacological and behavioral interventions can be effective as long as individuals with alcohol use disorders remain engaged in treatment (Miller et al., 2003). By the end of the first year after treatment, however, the majority of clients have resumed pretreatment levels of abuse and dependence, with relapse rates as high as 90% (Brownell et al., 1986; Miller and Sanchez-Craig, 1996). Given these high relapse rates, there is a critical need for identifying affordable and accessible strategies to help clients maintain their abstinence from alcohol. Most alcohol treatment programs in the United States encourage clients to participate in Alcoholics Anonymous (AA) and associated 12-step programs, such as Narcotics Anonymous (NA), both during and after treatment. Originating in 1933 in Akron, Ohio, AA has grown to become the largest and most popular mutual-help program in the U.S. for individuals with alcohol problems (Emrick, 1999). The frequency at which AA meetings occur on any given day in the majority of American cities and the absence of membership fees contribute to the popularity of this community-based resource. The limited but advancing empirical literature examining the impact of AA, most commonly measured as frequency counts of AA meeting attendance, has generally concluded that participation in AA may be modestly associated with improved psychosocial functioning and drinking outcomes, possibly in a dose-response relationship Cross et al., 1990; Emrick, 1999; Humphreys et al., 1996; Montgomery et al., 1995; Pisani et al., 1993; Tonigan, 1995; Tonigan et al., 1996a; Zywiak et al., 1999). Notable biases exist, however, including study methodological quality and sample selection (Kownacki and Shadish, 1999). There is still little understanding, moreover, of what involvement in AA specifically entails. This involvement has been determined to be a multidimensional construct, one component of which is AA meeting attendance (Allen, 2000; Caldwell and Cutter, 1998; Gilbert, 1991; Morgenstern et al., 1997; Owen et al., 2003; Snow et al., 1994; Tonigan et al., 1995). Other identified components are faith development (Horstmann and Tonigan, 2002), social networks (Kaskutas et al., 2002; Longabaugh et al., 1998) and several AA-related practices, such as declaring a Home Group and engaging in conversations with AA members both before and after meetings (Caldwell and Cutter, 1998). There may, however, be such other important “active ingredients” of AA involvement as sponsorship and twelfth-step work. There has been little exploration of the relationship between the specific activities in which AA members participate as part of the organization and long-term changes in drinking behaviors. Additional investigations are therefore needed to help identify which aspects of AA are most related to maintaining abstinence. An important activity that AA promotes is alcoholics helping other alcoholics to stay sober through involvement in service work (Nealon-Woods et al., 1995). Although the benefits of having a sponsor are apparent to the recipient (Caldwell and Cutter, 1998; Kingree, 1997; Morgenstern et al., 1996; Sheeren, 1988), the AA literature encourages alcoholics to help other alcoholics primarily as a method of strengthening their own sobriety (i.e.,“Helping others is the foundation stone of recovery” [AA, 2001, p. 97]). A continuing theme throughout AA principles is the critical importance of recovering alcoholics shifting their focus from self to others (Carroll, 1993; McGrady, 2003). Through sponsorship and twelfth-step work, service involvement comprises a broad range of activities that involve directly helping other struggling alcoholics. Although there is no formal definition of sponsorship, sponsors typically have some regular contact with their sponsees to provide guidance and encouragement, to discuss what challenges sponsees may be facing and to reinforce the AA teachings and principles (Humphreys and Noke, 1997). Twelfth-step work also involves helping alcoholics more generally, although the manner of help is not concretely defined. It is described as taking on “the unspectacular but important tasks that make good twelfth-step work possible, perhaps arranging for the coffee and cake after the meeting, where so many skeptical, suspicious newcomers have found confidence and comfort in the laugher and talk. This is twelfth-step work in the best sense of the word” (AA, 1981, p. 110). Important research has begun to shed light on whether providing help to others increases the chances of sustained recovery from substance use disorders. In a prospective investigation of 503 inner-city drug users recruited from the community, Crape et al. (2002) report that being a sponsor was related to not using drugs at baseline (odds ratio [OR] ≈ 7) and to abstinence at 1-year follow-up (OR = 3.2), after adjusting for various demographic and clinical variables, including AA/NA attendance. They also suggest that being a sponsor may have increased abstinence likelihood because it was associated with greater treatment involvement and with motivation for lifestyle change. In a cross-sectional investigation of 200 alcoholics in recovery, Zemore and Kaskutas (2003) found the total amount of time spent in community-related helping was significantly related to length of sobriety (r = 0.40). No study to date has prospectively investigated whether helping others through sponsorship or twelfth-step work is helpful or deleterious to sustaining one’s own alcohol sobriety. It is unclear whether empirical evidence supports the sentiment often heard in AA meetings: “You can’t keep it unless you give it away.” It is possible that helping other alcoholics may reinforce the positive benefits of sobriety and the negative consequences of drinking, the recollections of which tend to fade as time since the last drink increases. It is also possible, however, that being a sponsor can result in worse outcomes through the influence of recurrent drinking episodes of those being sponsored. Analyzing the relationship between helping other alcoholics and drinking outcomes using complex statistical methodologies, such as time-to-event analyses, can help determine whether alcoholics who are helping other alcoholics following treatment have better longer term alcohol outcomes than those who are not. These advanced methods also enable exploration of the unique contribution of helping other alcoholics on drinking outcomes aside from the frequency of AA meeting attendance. Using prospectively collected data from Project MATCH, one of the largest clinical trials of alcoholism treatment ever undertaken, this study addresses three basic questions: (1) Do participants who become involved in helping other alcoholics differ in their personal characteristics (i.e., demographic information and severity of drinking behavior) in comparison with those who are not involved in helping others? (2) Do alcoholics who become involved with helping other alcoholics during treatment have better long-term alcohol outcomes than those who do not? (3) What is the relationship between helping other alcoholics and long-term drinking outcomes when the number of AA meetings attended is considered?