The fate of this second book-length Rand report, The Course of Alcoholism: Four Years After Treatment, was more astonishing than that of the first. The directors of the NIAAA, which had funded the study, praised the research but offered their own very different version of its results. The NIAAA directors claimed that "abstinence is the appropriate goal in the treatment of alcoholism" and that attendance at Alcoholics Anonymous meetings offers the best prognosis. Yet the report itself found that those who had attended AA sessions were as likely to relapse as those who had not.

The head of the alcohol institute and his chief assistant were not researchers themselves. They were former administrators of alcoholism treatment centers and programs. This reflects a unique aspect of alcoholism policy in the United States—the degree to which people who are closely tied to grass roots programs like AA dominate views on the subject. The result is a national policy that sees alcoholism as the leaders of AA do, in black and white, without grays. By contrast, in most other Western countries, social workers and psychologists, who are inclined by training to notice social and cultural influences on behavior, and who regard moderate drinking as one possible goal for alcoholics, are influential in setting treatment policy.

Since the commitment that most Americans feel toward the philosophy of abstinence is rooted in strong feelings and cultural biases, not in hard information, they tend to downplay troublesome research. If the results of a study run counter to the prevailing notion that alcoholism is a disease, the study is rejected. It is not uncommon, for example, for alcoholics to maintain that they cannot taste alcohol in any form or in any quantity without endangering their sobriety—to believe that the drug sets off an allergic reaction in their bodies. But investigations of alcoholics' responses to liquor have consistently found that the opposite is true: not only are alcoholics capable of regulating their drinking, they actually do regulate it in many circumstances. One well known demonstration of this fact was performed by psychologist Alan Marlatt at the University of Washington. Marlatt gave alcoholics flavored beverages, some of which contained alcohol. If the drinker believed he was consuming alcohol, Marlatt found, he was likely to feel the need to drink more—even if he was actually drinking tonic water with a dash of lime. If he was drinking liquor but did not know it, he did not have this reaction. Clearly, the psychological factors in such an experiment are more important than the physiological. The "allergic reaction" is simply a medical myth, part of the general mythology of the disease theory of alcoholism.

Janet Fish, Tanqueray Bottles, 1973

The more sophisticated versions of the disease theory acknowledge that social situations and attitudes do affect a drinker, but still hold that there are unavoidable, inbred factors that point certain susceptible people toward alcoholism. In this view, there are three objections to controlled drinking. First, while attempts at moderation may last for a while, the alcoholic is bound to relapse eventually. Second, some problem drinkers may be able to control their drinking, but severe alcoholics, who are physically addicted, will find it virtually impossible to do so. Third, controlled drinking, even if possible in a few cases, is always risky and unnecessary.

The results of the second Rand report contradicted all three assertions. It found a substantial minority who were drinking moderately four years after therapy—including a significant number of those who were heavily addicted when they had entered treatment. The report, like nearly every other reputable study, found that most alcoholics undergo periods of heavy drinking followed by stretches of abstinence or moderation. The Rand subjects often relapsed. But it was not only those attempting to drink moderately who relapsed—and, indeed, for certain types of drinkers, trying to abstain worked worse than drinking moderately. For single men under forty, the goal of abstinence was actually counterproductive.

Ultimately, suspicion of the Rand reports is not founded on issues of procedure and accuracy; taken together, they were careful and soundly conceived studies. Rather, the ill will toward them stems from feelings that the Rand researchers were unaware of the horrors of the alcoholic's personal situation, his troubled homelife and work life. To anyone who has ever worked closely with an AA group, say critics, an emphasis on abstinence seems only common sense. To publicize any contradictory finding is to make treatment that much more difficult.

George Vaillant, formerly of Harvard University, expresses such reservations about the Rand reports in his recent book, The Natural History of Alcoholism. The book has been greeted as a landmark, reporting results of the first large-scale, systematic study since the Rand reports themselves. Sixteen years ago, Vaillant took over two large research studies that had been established in 1940 to follow the lives of some 660 men: 204 of them Harvard graduates, 456 of them members of the urban working class. Vaillant published Adaptation to Life in 1977, analyzing the reasons for personal success and failure in the Harvard group. In The Natural History of Alcoholism, he not only discusses the Harvard and working-class men but also follows up on patients of a medical clinic, the Cambridge Hospital alcohol program, of which he was a director.

Vaillant (now at Dartmouth Medical Center) is strongly sympathetic toward Alcoholics Anonymous and endorses the disease view of alcoholism and a medical model for treating it; if nothing else, he says, the term disease "is a useful device both to persuade the alcoholic to admit his alcoholism and to provide a ticket for admission into the health care system." Yet this position places Vaillant curiously at odds with his own data, tables of which fill a large fraction of the book.