Recently my friend John Halford posted on blog a link to an article entitled “Does Alcoholics Anonymous Work?” from the March 2011 issue of the Scientific American magazine. The article stimulated some re-thinking of the problems that scientific researchers face when they try to measure the effectiveness of Alcoholics Anonymous. In this post I will try to do the following: (1) outline the article from Scientific American, (2) sketch those features of Alcoholics Anonymous that stand in the way of scientific efforts to measure AA’s effectiveness, (3) show how the research relied upon in the article failed to deal with those factors and thereby failed to evaluate AA, and (4) conclude with a brief argument that AA cannot be scientifically evaluated.
I subscribed to the magazine for a while in the 1960’s when it appeared to be one of the leading journals of cutting edge research. While not necessarily any longer among the world’s elite science publications, the Scientific American has a good reputation with me as a venue for articles that help nonscientists to think scientifically. This particular article does not meet that standard. The authors advance what I believe is a nonscientific conclusion by use of still other unscientific studies. But I will also try to show why I believe scientific efforts to study AA are generally unsuccessful.
(1)OUTLINE OF ARTICLE “Does Alcoholics Anonymous Work?” begins with the tentative conclusion that “AA may help some people overcome alcoholism, especially if they also get some professional assistance,” then briefly sketches AA history from 1935 to 1939 before attempting to explain the authors’ answer mostly by reviewing two studies done by other researchers. After that the article summarizes with the tentative conclusion which began the piece: AA may help some problem drinkers, but professional assistance should be used in conjunction.
The article correctly observes that AA is difficult to evaluate for several reasons which include the fact that AA meetings vary widely. Nevertheless two studies are explained in some detail and the authors of this article then apparently rely upon the studies to reach the qualified conclusion above.
The first of these studies are of 900 subjects randomly selected by observers for inclusion in three therapy approaches to helping problem drinkers, one of which is said to be “AA based”. Apparently the approach which the authors label “AA based” included subjects who had not voluntarily gone to AA for the purpose of achieving total abstinence and their progress in doing AA’s recovery program, which is the 12 Steps, was guided by outside professionals. That study also measured the effectiveness of AA by data about the number of subjects whose drinking had diminished a short time later.
The second study showed that drinkers who attended 27 or more AA meetings during a first year of would-be abstinence reported a much higher rate of abstinence when surveyed 16 years later. This second survey also indicated that people who got professional help during the first year were more likely to be totally abstinent 16 years later, although the rate of increased
sobriety showed a much lower correlation for professional help than early AA attendance.
(2) ESSENTIAL FEATURES OF AA THAT MAKE RESEARCH DIFFICULT The article’s history lesson teaches that AA grew from one alcoholic’s work with another, that the solution was rooted in Christian philosophy, and that the book the first two wrote became the foundation of the immense fellowship that followed. That history might have been adequate if it had included the fact that AA is not only a 12 Step recovery program but an extremely loose confederation of autonomous groups, not isolated individuals, and that these groups are all asked to adhere to “12 Traditions” which are as integral to the AA recovery process as its 12 Steps. AA has no central governing board which enforces compliance, but AA literature has predicted for over 60 years that noncompliance with the Traditions will lead to diminished capacity for Alcoholics Anonymous to work.
AA’s Twelve Traditions were formally adopted by the first international convention of AA groups in 1953. These Traditions codified principles about how each and every AA group should conduct itself. Without them Alcoholics Anonymous would be as ineffective as a loose coalition of individuals or a conventional organization with a hierarchy and the usual organizational politics. Instead AA is composed of thousands of autonomous groups of alcoholics, each of which is devoted to the sole purpose of carrying the original message of AA. These traditions summarize the experience of early AA groups and the governing principles proved to be necessary for the survival of an AA which worked.
For the reader’s better understanding I now give you those traditions, word for word: (1) Our common welfare should come first; personal recovery depends upon AA unity. (2) For our group purpose there is but one ultimate authority—a loving God as he may express himself in our group conscience. Our leaders are but trusted servants; they do not govern. (3) The only requirement for AA membership is the desire to stop drinking. (4) Each group should be autonomous in matters affecting other groups or AA as a whole. (5) Each group has but one primary purpose—to carry its message to the alcoholic who still suffers (6) An AA group ought never endorse, finance, or lend the AA name to any related facility or outside enterprise, lest problems of money, property, and prestige divert us from our primary purpose. (7) Every AA group ought to be fully self-supporting, declining outside contributions. (8) Alcoholics Anonymous should remain forever nonprofessional, but our service centers may employ special workers. (9) AA, as such, ought never be organized; but we may create service boards or committees directly responsible to those they serve. (10) Alcoholics Anonymous has no opinion on outside issues, hence the AA name ought never be drawn into public controversy. (11) Our public relations policy is based on attraction rather than promotion; we need always maintain personal anonymity at the level of press, radio, and films. (12) Anonymity is the spiritual foundation of all our traditions, ever reminding us to place over personalities.
Two of the 12 Traditions strongly urge each AA group to respect the anonymity (i.e., privacy and public invisibility) of each of its members and participants. Anonymity makes difficult any effort to obtain verification of participation, but data obtained through violation of members’ anonymity would probably be needed in any research. But anyone familiar with AA would tell any researcher the truth that AA’s effectiveness depends in part on adherence to the principles of AA—including anonymity. An AA group which actively ignored the fellowship’s warning against affiliation (Tradition 6), as well as traditions against violating anonymity (see Traditions 11 and 12) would not deserve to be called an AA group. But accurate research depends upon the researcher being able to identify and categorize the alcoholic subjects in ways that would almost certainly conflict with these principles. The singleness of purpose stated in Tradition 5 seems incompatible with that of requiring subjects to account to researchers for their drinking, and the requirement that the member strive for total abstinence (Tradition 3) would be undermined by expectations that diminished consumption would be credited as progress.
(3) HOW THE SCIENTIFIC AMERICAN ARTICLE AND ITS CITED STUDIES FELL SHORT Some research designs have more potential for producing skewed results than others. Any research methodology which violates Traditions other than those concerning anonymity and nonaffiliation would likely produce more misleading conclusions. Unfortunately the research project most relied upon by the Scientific American article probably depended upon members and groups who were willing to violate traditions regarding (a) the requirement for AA membership (Tradition 3), and (b) the prohibition against professionalization of “12 Step work” (Tradition 8), and (c) and the prohibition against AA getting involved in “outside issues” (Tradition 10). Even if the proponents of the research within AA were right in arguing that this study would lead to no such violations, the inevitable debate within AA would lead to less unity within AA, which certainly conflicts with Tradition One.
Alcoholics Anonymous offers a method for overcoming alcoholism only to would-be members who have a “desire to STOP drinking” in the words of AA’s famous Third Tradition. So cooperation with studies which require “randomly assigned problem drinkers to attend 12 weeks of AA” would also violate this tradition since would-be controlled drinkers would presumably be included with those seeking abstinence. As the results of the study show, AA’s workability was measured by its success in achieving diminished drinking, an objective incompatible with AA membership and AA’s message. Those who assigned problem drinkers to the “AA based” neither measured success by total abstinence nor selected only from those who desired to attain total abstinence. Yet the authors of the Scientific American article cite such a study as “well designed”.
The same study used professionals “to help patients work the first few of AA’s 12 Step program”, and this professionally-assisted program is the chosen representative of what AA can do, although the authors of this article admit that such professionalization of AA is anathema to AA, as indeed it is (see Tradition 8). AA neither suggests that new members need professional help nor denies that professionals could help with some problems. But Tradition 8 is plainly against the idea of professionalizing the process of helping newcomers to do the steps.
Nevertheless it may help many readers to become aware of the fact that there is a wide divergence of opinion among would-be experts. The article’s conclusion, which follows, strikes me as both saying very little and suggesting too much: “Taken as a whole, the data suggest that AA may be helpful, especially in conjunction with professional treatment, for many people who are addicted to alcohol.” No clear answer to the article’s title question is actually offered, but the suggestions are that AA is always better when accompanied by professional counseling and that no one really knows whether AA is helpful or not. But no research is offered by the article to show that professional treatment enhances the results of doing AA without professional help.
At least one of the studies which the article sites surely makes a strong case that AA works—if the conclusions of that study are accepted, and if by “works” one simply means that going to AA predicts success in staying sober. The authors point to a 16 year study which claims that problem drinkers who attended at least 27 meetings of AA during the first year were twice as likely to be sober 16 years later than those who did not go to AA (67 percent for the attenders versus 34 percent for those who did not go.) That study did not contain a breakdown between those who had professional counseling and those who didn’t. In fact the authors of the article never present such comparisons to support their claim that AA and treatment work better than AA alone.
The authors of the Scientific American article waffle on the subject of whether AA forbids or encourages its members to seek out professional help for their alcoholism. Then they mix professional guidance in the design of what they label as an “AA based” approach, which then becomes their choice for comparing AA with other therapies for groups of problem drinkers. Those sent to AA are then evaluated by criteria which conflict with AA’s stated goal—total abstinence—and they are evaluated without regard to AA’s traditions of anonymity for members and singleness of purpose for each group. In short this principal leg on which the article stands is in multiple conflicts with AA itself—in purpose, method and permissible context. The article winds up evaluating something that does not abide by AA’s traditions, and so does not deserve to be called AA. I am not arguing that AA is superior to what the chosen study is evaluating. But I am pointing out that AA is different from that which the study calls “AA based”. Likewise the 16 year study only keeps records of differences between those who did and did not go to 27 AA meetings during the first of those years. Such records do not even attempt to evaluate the impact of actually doing the AA program (the 12 Steps) in groups which observe the Traditions.
(4) AA CONNOT BE EVALUATED BY COMPARATIVE SCIENTIFIC STUDIES Oddly enough many of the best AA meetings are about evaluating AA and about very little else. The usual formula for each AA speech within each meeting is that every speaker is to reflect out loud on himself as a problem drinker, what happened in his experience with AA and what he is like now that he is a sober member of AA. While there are AA meetings that tend to focus instead on the existential vicissitudes of the most vocal “sharers”, these things usually survive in the special greenhouses of professional programs that use the AA name for some features of therapy. AA does not have a litigation team to suppress such practices, but even a casual study of AA history will surely convince the student that the fellowship and program of AA is about recovery from alcoholism through teaching and practicing the Steps in groups which observe the Traditions. In that context the AA member is constantly hearing about the effectiveness of doing the program. For the professional researcher the discouraging fact is that this process, which is extremely effective in helping those who adhere to it, is not open to the kinds of behavior which scientific research needs to be scientific.
I have tried to find some analogies to illustrate how AA differs from the researchable. How about poetry as a means of fighting illiteracy? Can scientists show that compulsory poetry readings are more effective in combatting illiteracy than phonetic spelling lessons or cartoons with bouncing ball subtitles? Or how about evaluating love making as a means to promote social contentment. Love making between consenting adults is an activity which appears to lead to some spectacular personal changes and satisfactions, to say nothing of some needed practical results. Familiarity breeds—as the saying goes. Yet I doubt if any scientific study could be designed which could collect candid data on the subject, much less data which would allow a scientific study of the effectiveness of love making compared to more open and controllable behavior such as canoeing or tennis. AA is no more a method for controlling problem drinking than love making is a method for promoting bilateral harmony. Both may lead to an enormous amount of health and happiness which would only be impaired beyond recognition by scientific disclosures of the participants.