Social Research Group Working Paper F-145, April, 1977.  This report was part of work performed under Task 4 of a contract from the National Institute on Alcohol Abuse and Alcoholism, ADM-281-76-0027, "Selected Studies on the Societal-Epidemiological Aspects of Alcohol Use and Alcohol-Related Problems."


Barriers to Alcoholism Treatment

(including a "Postscript, 1999" section at the paper's conclusion)

Ron Roizen
Social Research Group
School of Public Health
University of California, Berkeley

By most accounts the number of alcoholics who have never had alcoholism treatment far exceeds the number who have:  conservative estimates might place the untreated-to-treated ratio at from 3:1 to 5:1 (Manis and Hunt, 1957; Harris, 1971; Edwards, 1973), though some data suggest estimates as high as 9:1 (Edwards, 1973) or even 13:1 (Gibbins, 1954; Baekeland and Lundwall, 1977).  Of course, opinions vary on the significance and the policy implications of the prevalence of untreated alcoholics in the United States. Some observers, especially those who regard alcoholism treatment as something the alcoholic must himself want and seek-out, tend to regard the phenomenon as an unfortunate but unavoidable reality.  Within that perspective, pressing into treatment, more or less unwilling alcoholics (or, as this vernacular describes them, alcoholics who have not "hit bottom") is probably both difficult to do and, anyhow, unlikely to accomplish much.  Other observers, for example representatives of the "societal reaction" or "labeling" school of thought, might well regard the untreated alcoholic as better off than the treated one.  And, still other observers have regarded the preponderance of untreated alcoholics as seriously problematic and as a call to action: Baekeland and Lundwall (1977, p. 322) call it " of the most dismal failures to detect and treat a serious and widespread medical-psychiatric condition known in the field of public health."

But the differing opinions over the value of alcoholism treatment are quite distinct from the most obvious questions presented by the untreated-to-treated ratio among American alcoholics:  Why have so many alcoholics remained untreated?  And, what accounts for the fact that one group of alcoholics has been treated and the other not?  This paper takes up these two questions.

The Structure of Commonsense Knowledge about Alcoholism Treatment

As it happens, U.S. culture provides us with a rich stock of both commonsense and expert knowledge about these two questions.  Significantly, most  discussions of alcoholism-treatment-reluctance in the literature cite a body of well-known causes (for example, the stigma on alcoholism, the alcoholic's tendency toward denial and the hiding of his condition, and so on) without specifically referencing to the experiments, to the observations or even to the authorities who support them.  It seems that commonsense knowledge is particularly easy to swallow in this domain, and that it is backed up by a considerable accumulation of informal lay and professional experience.

It will be helpful to think of this knowledge as reflecting two distinct but nevertheless related explanatory motifs:  the first explanatory style is universalist in that it concerns forces and factors that might apply to all alcoholics; the second style is differential in that it focuses on differences among alcoholics that might account for one being in treatment and the other untreated.  Clearly, these two styles correspond easily to the crucial difference between the first and the second question raised above: the question "Why have so many alcoholics remain untreated?" readily gives way to (universalist) discussions of the overall unpleasantness or undesirability of alcoholism treatment as it is viewed by alcoholics (or by the nonalcoholic population as well); on the other hand, the question, "What accounts for the fact that one group of alcoholics has been treated and the other not?" equally readily requires that we discuss (differentially) the qualities of some alcoholics that make them good treatment candidates and the qualities of others that are likely to keep them out of treatment.  Thus, our two master questions correspond more or less directly to two styles of explanation: one that concerns negativist attributes of treatment as they might be perceived by all alcoholics, and another that relies on a variety of thought-to-be salient differences among alcoholics with respect to treatment susceptibility.

Universalist Explanations

Perhaps the most durable theme in the discussion of alcoholism treatment reluctance is stigma.  It is thought that alcoholics resist treatment because the acceptance of the designation "alcoholic" brings upon himself great discredit.  As other stigmatized statuses, the alcoholic is thought to be  reduced in our minds from a whole and usual person to a tainted, discounted one" (Goffman, 1963, p. 3).

In part because of alcoholism's social stigma, since the repeal of National Prohibition most advocates of reform in the social management of alcoholism have centered their efforts on the diffusion of the notion that "alcoholism is an illness." If alcoholism were thought of as an illness, the reasoning might go, then, that neither the alcoholic himself nor those around him'would blame him for this condition. And if he were not blamed for his alcoholismic behaviors, then the moral stigma on alcoholism might be expected to dissipate.  And, in time, the barrier to treatment presented by the stigma on alcoholism  might be reduced or even disappear.

Early alcohologists and alcoholism reformers saw that there were two quite distinct problems facing any effort to reduce alcoholism's stigma.  On the one hand, alcoholics themselves had to be persuaded that they suffered from an illness, thus making for more willingness to seek out treatment. But, on the other hand, the general public also had to be educated to the belief that alcoholism was an illness:  the disease conception of alcoholism hardly would provide a defense against public stigma if only the alcoholics themselves in the community happened to hold the disease veiw. Stigma, after all, is first a property of public beliefs and attitudes. Riley (1949, p. 303), upon reporting that only 20% of the American public in 1946 thought the alcoholic sick, clearly stated the significance of public opinion:  "If the alcoholic could look upon himself as sick rather than as a spineless failure, he would obviously be more likely to seek help.  But with 80% of the public unwilling to concede this point, it is difficult for him to do other than accept this definition of himself."  And for such reasons early opinion often placed a great deal of importance on remolding both the opinions of alcoholics and of the public in general.  Anderson (1942), for example, wrote, "The chief obstacle to progress in the scientific solution of problems concerning alcohol lies in public opinion."

Thus, in part, the problem of getting alcoholics into treatment was defined as the joint domain of the health educator and the public opinion researcher, and a modestly-sized research tradition grew up around both the measurement of public acceptance of the disease concept and estimations of the significance and impact of that acceptance.  By some contemporary standards, at least, the diffusion of the disease concept can be regarded as an accomplished fact.  Room, in 1972 (pp. 1049-1050), for example, writes, "As a 'health education' campaign in the United States, the disease conception must be judged an astonishing success.  A majority of the American public will now agree with a general formulation of the disease concept; alcoholism is an acceptable medical diagnosis for entitlement to many social services for the disabled; not only the medical profession but also the U.S. Attorney General have proclaimed that 'alcoholism is an illness'; the Federal budget for alcoholism as a health problem has risen in the space of a few years from essentially nothing to $85,000,000."

But in spite of the seeming success of the alcoholism movement, most students of public opinion on alcoholism have tended to be quite circumspect about the degree to which public opinion actually has changed or the extent to which such changes influenced people's actions.  It will be appropriate, then, to review for a moment, some of the findings and difficulties uncovered in this research tradition. Given that we are here centrally interested in the relationship between reluctance-to-be-treated among alcoholics and public opinion on alcoholism, it will be useful to organize this review around two questions:  (1) To what extent has the disease concept of alcoholism been adopted by U.S. public opinion? (2) To what extent is the adoption of the disease concept associated with the adoption of corollary views in the public health perspective on alcoholism, especially the view that alcoholics should be treated for their illness?

The Adoption of the Disease Concept in the U.S.: 1946-1975

Alcohologists in the 1960s were happy to find that an apparent major shift in public attitudes toward alcoholism took place in the mid-fifties (Haberman and Sheinberg, 1969, p. 1211; Jellinek, 1960, p. 183).  Though this shift to  majority acceptance of the "disease concept" was thought to be "not deeply  rooted as yet" (Jellinek, 1960) and often "little more than lip service" (Haberman and Sheinberg, 1969), it nevertheless was regarded as progress, and even astonishing progress, in an area in which opinions might have been expected to be slow to change.  Thus, some researchers in the sixties regarded  the slogan "alcoholism is an illness" as well-diffused in American public opinion and saw the remaining tasks as deepening it roots and drawing out its actionable implications.

But, if we stop to look over the evidence for a major shift in (even) the superficial acceptance of the disease concept, it seems to be arguable how much public opinion actually had,changed since the forties.  Jellinek (1960) based his conclusion that a major mid-fifties shift had taken place on two surveys, one done in 1946 by Riley (1949) and the other by Roper, et al. in 1958.  These two studies suggested to Jellinek that public acceptance of the disease concept had jumped from 20% in the mid-forties to 58% in the mid-fifties  (Jellinek, 1960), p. 183).  Haberman and Sheinberg (1969) also drew on the survey data reported by Riley and Roper, et al. and, in addition, on reports by Gallup (1955), Mulford and Miller (1961), Maxwell (1952), Star (1955), and Lemkau and Crocetti (1962).

But these several surveys studies, in fact, could only hint at directions and magnitudes of change in public opinion. They harbored several important drawbacks that made solid conclusions impossible to draw.  Consider the arbitrary year, 1954, as the watershed point in U.S. public opinion: this gives us three "pre-change" studies (Riley's, done in 1946; Maxwell's, 1951; and Star's, 1950) and four "post-change" studies (Gallup, done in 1955; Roper, et al., 1958; Mulford and Miller, 1958; Lemkau and Crocetti, 1960).  None of the three "pre-change" studies used a "disease-concept" item that readily corresponded to the "post-change" studies.  Riley's item is not spelled out in his paper but it appears to have asked respondents simply whether alcoholics could stop drinking by themselves.  Riley (1949) reports that it was found that ". . . half of the people in the United States believe that if he wanted to stop drinking, he could" (p. 302).  The remaining half of the sample seems to have been asked why the alcoholic could not stop drinking, and among this group a total of 20% (of the total sample) replied, "because he is sick" or something to that effect.  This series of questions is probably similar to the series asked by Maxwell (1952) in 1951 of his sample of adults in Washington State; Maxwell, at any rate, reported quite similar findings (41% of his sample thought the alcoholic "could stop"; 54.6% thought he could not stop; and a total of 20.1% gave the reason "because he is sick").  Maxwell, like Riley, concluded that "only one in five states that the alcoholic cannot stop drinking because he is sick."  Of course, this question format is a poor one for assessing the total fraction of the population that might agree to the assertion "alcoholism is an illness."  Only the respondents who thought alcoholics unable to stop appear to have been given the opportunity to express a reason for that inability, and thus about half of both the Riley and Maxwell samples did not have the opportunity to say the alcoholic was sick.  Both researchers found that saying "the alcoholic is unable to stop" was not tantamount to saying "alcoholics are sick"--both reported that roughly half of those who thought alcoholics unable to stop thought they were weak-willed.  But the corresponding fraction of respondents who thought the alcoholic could stop but nevertheless thought alcoholism a disease was not queried into.  In any case, certainly it was a question format in which the apparent level of approval of the disease conception of alcoholism would be kept to a minimum.

The third "pre-change" study, Star (1955), asked of a national sample whether alcoholism was a "mental illness," a question whose positive response hardly can be regarded as a sign that the classical disease conception of alcholism has been adopted by the population.

No "post-change" survey employed an item like the Riley-Maxwell item.  The Star mental-illness item was repeated in 1960, but in a sample of Baltimore rather than in a comparable national sample.  The Roper, et al. (1958) national study used an item that did not mention alcoholism at all, but instead asked:  "If you knew someone who habitually drank so much that it affected his job and his relations with people, would you say that he is morally weak or would you say that he is sick?"  In this item, of course, all respondents are asked if the drinker is "sick," and the "sick" response is provided in an "either-or" framework that both (1) does not require the respondent to volunteer a sickness view on his own, as in Riley and Maxwell's studies, and (2) asks the respondent to choose between only two alternative conceptions of alcoholism.  Thus, the Roper et al. item, seems to stack the deck in favor of pro-illness responses about as much as the earlier items stacked it against them.

But even these relatively substantial differences in question formats would not have disabled efforts to follow the historical course of public opinion if the idea that "alcoholism is an illness" had revealed itself to be a clearly formulated and easily tapped concept in American thought.   In other words, if different studies had shown that the illness conception was accurately and similarly measured in similar populations in spite of substantial differences in question wordings and formats, then the criticisms of the differences between the Riley and Maxwell and the later studies might be set aside as not too serious.  As it happened though, what evidence existed in the mid-sixties began to suggest that question wording was, in fact, quite strongly related to the response distributions.

Mulford and Miller (1964), for example, discovered that Iowa respondents did not necessarily adopt illness conceptions of alcoholism to replace their earlier moralistic views but more often simply added the idea that alcoholics are sick to the idea that alcoholics are weak-willed.  Hence, a question that asked respondents whether alcoholics are sick could find 65% of the sample agreeing, but only about a quarter of the total sample defined the alcoholic as sick without also defining him as morally weak.  Somewhere between 24% and 65% of Iowans, therefore, could be regarded as agreeing with the disease concept, depending on what the questioner meant by agreement.  Serious questions also were raised concerning precisely what the respondent meant when he said that the "alcoholic is ill."  Wexberg in 1951 had shown that the slogan had at least eight different possible meanings, and Marcus (1961) counted 16 different meanings attached to the assertion in his survey of an Ottawa Valley town.  He wrote: "It is a sobering thought to realize that a presumably simple statement such as 'alcoholism is an illness' can be interpreted in so many ways . . . It is apparent that the statement means many things to many people; it is not so apparent whether the belief has any functional significance."  Upon reviewing the divergent results of four studies employing different disease concept item-wordings, Marcus (1962, p. 5) was led to conclude, ". . . people's expressed opinions about this issue are rather sensitive to variations in the interviewing procedure and in the phrasing of the question--which leads one to suspect that such opinions are neither very stable nor very strong."

The problem of non-comparable survey items was avoided in a survey of the adult population of Vancouver, Washington conducted by Linsky (1970) in the mid-sixties. Linsky makes the provisional assumption that ". . . differences in attitudes and beliefs of succeeding age groups are assumed to reflect the differences in popular orientations prevalent during the formative years of each age group.  Since both the formal educational system and the mass communications media are considered major agencies of cultural change, it was assumed that attitudes of persons having most contact with these institutions would be indicative of the direction of cultural change."  Of course, age differences are not necessarily the result of historical effects, but may be due to cohort effects and to aging per se; in addition, age, exposure to media, and education are very likely to be positively associated with one another in Linsky's sample, further obscuring what role each variable plays and how accurately the historical tendency of public opinion is reflected.  In any case, Linsky found modest but monotonic relationships between each of the three independent variables and an attitudinal scale concerning views of alcoholism and alcoholics.

In light of the several difficulties of interpretation surrounding the public acceptance of "alcoholism is an illness" it is not surprising that several more recent investigators have broadened their data collection to include some of the attitudinal correlates and corollaries of the slogan's acceptance.  By itself, it seems that knowing the public acceptance of the slogan alone reveals little of its meanings, implications, and depth of internalization.  Thus, for example, the most recent national survey on alcohol and alcoholism (NORC, 1975) employed a four-point scale that provided the respondent with a rough gradient of relative contributions for illness and personal control in alcoholism. The question read as follows:  Which of these statements most nearly describes alcoholism? (1) An illness that strikes without warning; people do not have personal control over whether they get  it or not; (2) An illness people have little personal control over whether they get it or not; (3) An illness that people have a great deal of personal control over whether they get it or not; (4) It's not a real illness.  People bring it on themselves. In the sample as a whole, 7% chose 1, 14% chose 2, 39% chose 3, and 32% chose 4 (the remaining 8% said "don't know").   Again, the acceptance of the wholly involuntarist picture of alcoholism is uncommon in U.S. public opinion (only 7% selected response 1), though two-thirds of the sample give some role to involuntary forces  To cloud matters further, in this same survey 65% of the sample thought it either "basically" or "partially" true that alcoholism is a sign of moral weakness.

But the NORC national survey brought a noteworthy innovation in the manner in which it assessed the contributions of personal control and involuntary forces to alcoholism. Survey respondents were asked the same question and given the same answer categories in items concerning pneumonia and heart attack. And though 38 and 51% of the sample (respectively) regarded these illnesses as entirely involuntary (6 or 7 times the rate for "alcoholism!'), the response# showed that even illness phenomena about which a great majority would agree that medical care is the most appropriate response nevertheless involve substantial contributions from "personal control":  concerning pneumonia, 53%, and concerning heart attack, 41% of the sample gave some role to personal control in the description of the illness.

The Disease Concept, Corollary Views, and Attitudes Toward Treatment

Of course, alcohol health educators did not want only to gain public acceptance of the idea that "alcoholism is an illness." It was hoped that a whole new gestalt about alcoholism would be adopted as well.  Underlying these educational efforts was the tacit ideal of the "enlightened community," a place where all alcoholics sought treatment because all of the disinclining attitudes and beliefs in the community had been educated away.  While these enlightened attitudes and beliefs probably should not be looked upon as separate and discrete precepts, but rather as an interrelated system, nevertheless certain central themes or important propositions may be regarded as forming a skelton for the ideology: (1) Alcoholism is an illness, and thus, (2) should not be sitgmatizing.  (3) The alcoholic needs treatment, (4) in part, because the condition is serious and progressive.  (5) Early treatment is better than late treatment, all else being equal.  (6) Alcoholism treatment is effective, and (7) one should know the danger signs of alcoholism in order to identify it in oneself or in others.  (8) Alcoholism can strike someone irrespective of sex, age, social class, and so on--it is wrong to regard the alcoholic in the old Skid Row stereotype of the past.  (9) Because it is a disease, alcoholism is an appropriate problem for physicians and psychiatrists.

For our purposes, of course, our central concern surrounds the question of public acceptance of the idea that alcoholics need treatment and the pattern of associations between this idea and the other elements of enlightened ideology.

The notion that "alcoholics need treatment" is not, of course, a necessary implication of the notion that "alcoholism is an illness": alcoholism might be an untreatable illness, for example, or one might support the treatment of alcoholism even if one did not believe it a disease. Relatedly, more than one sort of meaning (again) might be attached to the "alcoholics-need-treatment" idea: it might mean, "alcoholics cannot cure themselves alone or without some sort of outside help"; or it might mean, "alcoholics should have medical or psychiatric treatment as opposed to other kinds"; or, in addition, it might mean, alcoholics should be treated rather than punished',' or"alcoholics should be treated rather than protected and allowed to hide their condition."

A variety of studies have delved into the several elements of the enlightened community's alcohol ideology and the relations between these beliefs and opinions and receptivity toward treatment.  It will be worthwhile briefly to review this literature.

McCarthy's and Fain's (1959) study of three Connectucut communities examined the public acceptance and relationships among four elements of the community's "knowledge about alcoholism": (1) alcoholism is an illness, (2) alcoholism is treatable, (3) physicians or psychiatrists are among the best qualified-to help the alcoholic, and (4) the respondent is aware of two or more community resources for the treatment of alcoholism. They found that the latter pattern of responses to these items approximated a Guttman scale, or in other words, public knowledge might be regarded as formed into steps or stages such that, say, agreement to the third proposition (i.e., physicians are qualified to treat alcoholism) appears after agreement, first, to the notion that alcoholism is an illness, and secondly, to the notion that alcoholism is treatable).'  This finding suggests that the four articles of knowledge they investigated act as necessary but not sufficient conditions for the adoption of the next article.  Also suggested in this picture of these articles is the implication that the adoption of the disease concept, by itself, is only the first step, and is a long way from the final step, on this sort of attitudinal or knowledge scale.  In their sample (this discussion is based on Table 2, p. 650), McCarthy and Fain found no disagreement with the notion that alcoholism is an illness, and only about 6% of the sample thought alcoholism untreatable (meanings?).  The consensus dropped sharply, however, when the third question was reached: only a little more than 40% of the sample thought a physician to be among the best two sources of treatment, and less than 15% actually knew of at least two places in the community where an alcoholic could go for help.  Thus, acceptance of the disease concept was only slightly more prevalent than belief in the treatability of alcoholism, but agreement to both of these opinions did not imply that the respondent thought medical treatment most appropriate or knew where to find (two) treatment sources.

Mulford and Miller (1961) explored the public acceptance of the notions that alcoholism is an illness and alcoholism requires expert attention by orienting their analysis as follows: "It seems reasonable to suppose that prevailing public definitions of the alcoholic influence the individual's perception of his own drinking as a problem or not, his recognition of the need for help, the nature of the help which he seeks and even the nature and effectiveness of the assistance which may be available to him."  Thereupon, Mulford and Miller analyzed the responses of their 1958 statewide Iowa sample to five questions: the first item concerned acceptance of the disease concept; the second, the respondent's readiness to reveal a drinking problem in his family to friends (perhaps an intermediate step between complete concealment and the seeking of outside medical help); the third, the respondent's readiness to seek help outside the family; fourth, the first sort of outside help that would be sought; and, fifth and finally, the acceptability of various sources of continuing help, e.g., AA, private sanitaria, and so on.

In this study, Mulford and Miller found that 51% of the sample thought an alcoholic best described as a "sick person."  Roughly the same proportion of the sample (49%) would discuss a family drinking problem with friends, as one might do "in the case of other illnesses." But, unlike McCarthy and Fain's findings, Mulford and Miller did not see the adoption of the disease concept as a necessary condition for "willingness to discuss" the problem with friends.  Instead, the two-items were only weakly associated with each other: among respondents who thought alcoholics sick, 53% "would discuss"; and among respondents who thought alcoholics weak-willed, 43% would.

Based on the data reported, it appeared that the respondent's willingness to seek outside help, as well, was not strongly associated with acceptance of the disease concept.  Mulford and Miller's item read:  "If you or one of your family happened to become an alcoholic, or developed a serious drinking problem, what do you think you would do about it? Try to solve it within the family, 25%; Seek outside help or advice, 12%; First try to solve it in the family, then if necessary seek outside help, 61%; Other, 2%."  It is noteworthy that almost three-fourths of the sample would seek outside help at some point, a fraction of the sample half again as large as the group approving of the disease concept. Cross-tabulated by belief in the disease concept, Mulford and Miller reported that 16% of the disease-concept-approvers, "Would seek outside help or advice" straight off; among those with a moral definition of alcoholism,,this was reduced to 10%.  Again,adoption of the disease concept was not strongly associated with turning to outside help.  Finally, the data showed that "attitude toward outside help" was not necessarily the reflection of an adoption of the view that medical help was most appropriate: among those who would seek outside help straight away, 58% preferred it from the family physician; but even among those who would keep the problem inside the family, 44% nevertheless would seek help first from the family physician if they sought help at all.

Thus from Mulford and Miller (1961) we get the impression of a good deal of slack in the relations among (1) acceptance of the disease concept, (2) belief in the need for outside help, and (3) belief in the appropriateness of medical help. They suggested that at least some of this slack might be coming from the previously-mentioned tendency for many approvers of the disease concept also to hold a moralist view of alcoholism as well.  In a second analysis, Mulford and Miller (1964) sorted respondents into three categories:  those who viewed the alcoholic as sick but not weak-willed, 24%;  . . . viewed the alcoholic as both sick and weak-willed, 41%; and . . . weak-willed but not sick, 34%.  Again the associations between the disease concept and various plans for action were analyzed, this time though with believers in a purely disease view and a purely moralistic view separated from each other and from believers in the hybrid view. Surprisingly, perhaps, cleaning up the disease concept variable did not much strengthen the relationships between it and the treatment (or action) plans.  Fully 91% of those holding the medical view of alcoholism agreed that alcoholics could not give up drinking by themselves but needed help, but two-thirds of those with a "weak-willed" view of alcoholism thought so too.  In all, three-quarters of the sample thought the alcoholic needed help, many more than the fraction adopting the pure disease view.  Forty-one percent of the sample would turn first to the family physician for outside help, and the physician was the most often mentioned source of help "regardless of whether the alcoholic was defined as sick, weak, or both sick and weak."

Implied in relatively low associations among these attitudes and "plans" was the finding that only a few respondents would show agreement with all of the elements of the health educator gestalt. Mulford and Miller (1964) had examined five of those elements: they wrote, "The present study  indicates that only 14% of the entire "package" -- that is, only 14% (1) agreed that the alcoholic is a sick person (this includes those who define him as both sick and weak), (2) agreed that he needs help, (3) would discuss a drinking problem with friends, and (4) would seek help from either a medical source or AA.

Blizard (1969) found, in a sample of urban New Zealanders, that though two-thirds of the sample regarded an alcoholic vignette as a person with a mental illness, the alcoholic was subject to much social rejection nevertheless.  Moreover, the level of social rejection was unrelated to the treatment of choice or to whether treatment or no treatment was called for. Thus, Blizard's study suggests that the model "adoption of disease concept leads to reduced rejection leads to greater propensity to medical treatment" may be wrong in each of its particulars: high acceptance of alcoholism, in his sample, went along with high social rejection; and social rejection, in turn, was unrelated to treatment preferences.  Haberman and Sheinberg (1969) found that though 64% of their New York City sample thought alcoholism an illness, such acceptance did not necessarily imply that alcoholism could be recognized or identified in others or in oneself.  Thus, again, even acceptance of the disease concept and the belief in the notion that alcoholism requires treatment might generate few new treatment cases because alcoholism symptoms often go unrecognized as such.  Linsky (1970) found that positive attitudes toward alcoholism treatment had enjoyed greater popular acceptance than had been enjoyed by the disease concept, and in a later study (Linsky, 1972), that images of the causes of alcoholism were not strongly associated with opinions about the most effective sort of treatment.  Thus, Linsky (1972) concluded that ". . . beliefs about alcoholism are not tightly integrated into a consistent ideology among the public."

In summary, this literature seems only to suggest caution. The acceptance of the disease concept appears to be high but not easily interpreted:  it may coexist in the same mind with a moralist view of alcoholism and may take on a variety of different meanings from person to person. In turn, it is usually not tightly associated with other elements of the enlightened ideology.  Public opinion appears rather selective about that ideology, fully capable of adopting a precept here and rejecting an associated precept there.  With regard to the desirability of treatment for alcoholism the picture is equally unclear:  treatment may mean different things to different hearers, and it appears that the proportion of the population favorable toward treatment may often exceed the total proportion adopting the disease concept; similarly, attitudes toward treatment may not be strongly dependent on other attitudes in the enlightened ideology.

New Data on Public Opinion, Alcoholism and Treatment

Something of the sportiness and sensitivity to slight gradations of meaning in public opinion items is easily visible in a recent study of adults in San Francisco conducted by Cahalan and Treiman (1976). This study, conducted in 1975, asked a variety of questions concerning respondents' views on alcoholism
 and various elements of the gestalt that contemporarily defines most professional opinion on alcoholism. A selection of items from this survey is shown in Table 1.

Seventy-seven percent of the sample agreed that "most alcoholics are physically addicted to alcohol," (la) seemingly a substantial consensus on behalf of a relatively strong formulation of the disease concept of alcoholism.  The sample also showed a relatively strong concensus behind the notion that drinking problems are often progressive (3a): fully 90% agreed that "if you don't get help when you have a drinking problem, it's likely to get worse and worse."  With regard to the question of recognizing alcoholism, 78% indicated a knowledge of alcoholism's "danger signs" (4a), and only 8% of the sample agreed to the statement, "if I had alcoholism, I'd rather not know-about it," (4b) -- suggesting that at least consciously most respondents with a knowledge of alcoholism's danger signs would not choose to ignore them in oneself should they arise. These items are accompanied by a relatively strong consensus in favor of the need for treatment for alcoholism: 80% of the sample agreed that "if you're having trouble with drinking too much, it's probably better to get help than to try to work it out yourself" (6a), and 78% of the sample indicated that they would "definitely" or "probably" search out treatment if they began to think that they had a drinking problem (6b). Thus, on these several attitudinal dimensions there is a substantial consensus in agreement with key elements in the public health perspective on alcohol problems.

 At least as many as three-quarters and in a couple of instances as many as 9 out of 10 respondents adopted the view that alcoholism is an addiction, that untreated drinking problems grow worse and worse, that drinking problems should be treated, and that they themselves would "search out" treatment if the need arose; and, finally, most respondents suggested that they are aware of alcoholism's danger signs and would be attentive to them if they appeared.

But the relatively high and "right-direction" consensus surrounding these indicators of the public health image of alcoholism are accompanied by other indicators where public attitudes either show less consensus or are contrary to those of our,"enlightened community": though three-quarters of the sample thought the alcoholic was "addicted to alcohol" (la), about half of the sample agreed to the voluntarist statement that "most alcoholics drink because they want to" (lb).  The data showed only a very slight association between these two opinions: thus among respondents who said the "alcoholic drank because he wanted to," 74% nevertheless regarded alcoholism as an addiction; among those who disagreed with the volitionalist statement, 80% regarded alcoholism as an addiction.

It is interesting to compare these data to Mulford and Miller's (1964) collected over a decade earlier among Iowans: in the 1975 San Francisco sample, among respondents who answered both questions (N=1105), 40% agreed to both the involuntarist (addiction) and voluntarist items (Mulford and Miller found 41%), 14% agreed to the voluntarist but not the involuntarist items (less than half the fraction that in the Iowa survey regarded the alcoholic as weakwilled with no mention of sickness); and finally, 37% agreed to the involuntarist statement but disagreed to the voluntarist one (half again as many as thought the alcoholic sick with no mention of weak-willed in the Iowa survey). of course the items and the populations differ substantially but, perhaps significantly, in both samples it is the combined voluntarist-involuntarist group that makes up the modal response in the survey. In the San Francisco-1975 survey, consistently involuntarist conceptions were next most common (37%), while in Iowa the second commonest opinion was the consistently moralist one (34%).

It is noteworthy that although three-quarters of the San Francisco sample agreed that alcoholics are physically addicted to alcohol, barely more than half believed that "medical tests for alcohol addiction are now widely available."  Thus, though alcoholism is widely regarded as an illness, it seems not to be widely regarded as one that is detectable by medical tests.  Again, the implication seems to be that broad agreement to an addiction picture of alcoholism does not imply that many of the particulars of the disease conception have been adopted as well.

With regard to the stigma on alcoholism, the San Francisco sample deviated sharply from the Utopian responses: A substantial minority (41%) agreed to the statement that "To be known as an alcoholic would destroy a man's reputation." Of course, it is arguable whether some respondents interpreted the item to mean "to be known as an active alcoholic. . ." while others thought in terms of an arrested or inactive alcoholic. Perhaps a better, if more subtle, indicator of the prevalence of alcoholism stigma is that 45% of the sample agreed that alcoholism treatment should be conducted away from one's neighborhood in order to protect the patient's privacy: conditions involving no social discredit (for example, for a limb fracture or an ear infection) presumably would not suggest this premium on privacy. That half of the sample thought "news would get around" about their alcoholism treatment (1c) and that nearly 40% believed that one would not be treated with "much respect" in county and city alcohol programs also suggest that in terms of moral distaste alcoholism persistently remains rather a different matter than condtions regarded as wholly physical illnesses.

About half the sample indicated that drinking problems ofte  are transitory and of no particular significance (3b), a view that might leave to many respondents much discretion concerning which drinking problems will be regarded as the first steps in an ever-worsening progression (3a) and which regarded as benign and not to be bothered about. Only slightly more than half of the sample agreed that "most people who get treatment for alcohol problems" overcame them.  And, finally, though 8 in 10 showed a regard for the desirability of alcoholism treatment, almost half of the sample (47%) indicated that they would only seek help for a drinking problem if all other alternatives had been exhausted.

We see in these data, a mixed picture of public acceptance.  It reveals considerable variations across the several elements of the "enlightened view" and a good deal of variety within a given element depending upon the wording of the question and the fine-grained implications that wording might carry.  On balance, these data do suggest that some elements of the public health perspective are better entrenched and matters of more consensus than others.  It appears widely accepted that alcoholism is an addiction, that it is often progressive, and that alcoholism requires outside treatment. On the other hand, these data suggest that public opinion is much more dubious about the contentions that alcoholism is not sitgmatizing, that alcoholism treatment is effective, or that treatment for alcoholism should be sought as readily as treatment for a toothache.

Of course, rather than suggest that public views of alcoholism form an intricate and sensitive system that is highly responsive to small shifts in question meanings and particularist in its response patterns, it might equally well be argued that for most of us "the alcoholism problem" is not something we spend much time thinking about and thus not something on which public opinion will show the logical structure that professional opinion might hope for.  In the past, this alternative explanation for the peaks and valleys and the intercorrelative independencies of attitudinal items have led various researchers to check their findings against subsamples in which more awareness of alcoholism and consistency of opinion might be expected.  Maxwell (1952), McCarthy and Fain (1959) and Mulford and Miller (1961, 1964), for example, compared the views of respondents who were acquainted with an alcoholic with those who were not acquainted.  Thus, Mulford and Miller (1964), for example, wrote, "It is reasonable to suppose that respondents who have a friend or relative who is a problem drinker would pay more attention to alcoholism education materials and might therefore be more knowledgeable on the subject" (p. 320).  If in turn, a subsample thought to be better informed and more likely to evidence a consistent or at least a more evenly "enlightened" viewpoint did indeed show much less sportiness and variation in response across the elements of the public health perspective, then this finding would lend support to the hypothesis that much of the volatility in public opinion was due to inattention or indifference on the part of the general population,

We have tested this hypothesis against two subsamples in the San Francisco study:  first, as in previous studies, we examined the response patterns of persons with a personal acquaintance to an alcoholic (N=278 or 24% of the sample) as measured by agreement to the following item:  deeply affected by someone else who is an alcoholic."   "My life has been deeply affected by someone else who is an alcoholic."  Second, in order to test for a possible difference between "exposure to alcoholism" as it might be measured by the above item, and knowledge or concern about alcoholism (the presence of which might make the attitude profile more intelligible and consistent), the sample was culled of respondents who indicated a knowledge of the "danger signs of alcoholism" a knowledge indicator) and who answered that they were "very" or "somewhat concerned" about the alcoholism problem in their community (an indicator of concern or interest).  Together, these two items split the sample almost exactly in half and provide us with a group which presumably contains most of the respondents whose knowledge and interest in alcoholism exceeds the population median.  But, as is shown in the second and third columns of Table 1, the pattern of responses in either the "acquaintance" or the "high knowledge and interest" subgroups did not differ markedly from the sample as a whole.

Thus, at least insofar as this analysis tests the hypothesis, variation in the public acceptance of various elements of the public health perspective on alcoholism are not due to superficiality, inattention, and unconcern; subsamples with relatively higher exposure, knowledge, and concern show equivalent variations of response.

Illness and Sin

That the modal group of respondents characterizes alcoholism in both volitional and involitional (or both medical and moral) terms has been traced to several possibilities. Linsky (1970, citing Mulford and Miller, 1964) suggested that public attitudes are frequently characterized "by inconsistency and ambiguity."  Jellinek (1960), Philp (1960, cited in Mulford and Miller, 1964), Mulford and Miller (1964), and Knupfer (1964) all have argued that the content of educational efforts, themselves, are often "...vague, ambiguous, logically inconsistent and confusing..." (Mulford and Miller, p. 321).  Knupfer (1964), giving the point a concrete example, noted in the press two medical sources quoted as stating that "with recent years alcoholism has been looked upon as a disease rather than a personal weakness.  It is not a matter of willpower," but continuing that "the most important factor in rehabilitation is the alcoholic's willingness to help solve his own problem."  The ideological overlap has been attributed to "ambivalence" (Orcutt, 1976, and Roman and Trice, 1968), and also regarded as a possible way-station in a transitional phase toward a more consistently medical ideology (Orcutt, 1976).

But it is possible, of course, that the seemingly philosophical inconsistency in popular opinion should be traced more directly to the substance itself of the classical disease concept.  As it has been disseminated through countless high school classrooms, mass media spots, and the propagandist efforts of AA members and their families, the classical AA Jellinekian conception of alcoholism involves both choice and compulsion.  In this view alcoholics might have a good measure of control over whether or not to begin drinking on a particular occasion, but a little control over stopping drinking once started.  Thus, this picture of alcoholism might lead one consistently to answer both -- that alcoholism is an addiction and that an alcoholic drinks because he wants to.  As it happened, almost half (46%) of the San Francisco sample agreed with the classical formulation that "alcoholics can decide whether or not they want to drink, they just can't decide when to stop once they have started drinking."  If, in fact, this classical image of the places of will and addiction in alcoholism accounts for the readiness of many respondents to attribute alcoholism to both will and addiction, then we ought to find that respondents who agreed to the classical view provide more of the both-addiction-and-will cases than are provided by respondents who did not express the classical view.

A. Total sample (N=1088) B. Subsample agreeing with AA-Jellinekian conception (N=516) C. Subsample rejecting AA-Jellinekian conception (N=556)
medical & moralist:
medical & moralist:
consistently medical:
consistently medical:
consistently medical:
medical & moralist:
consistently moralist:
consistently moralist:
consistently moralisty:
neither medical nor moralist: 9% neither medical nor moralist: 6% neither medical nor moralist: 13%
total: 100% total: 100% total: 100%
Table 2 shows this prediction is not disappointed.  Table 2 shows the frequencies of the "consistently medical, medical-and-moralist, consistently moralist, and neither medical nor moralist" views of alcoholism in the sample as a whole (A) and in the two subsamples defined by agreement (B) or disagreement (C) with the AA Jellinekian conception of alcoholism.  The table suggests a number of interesting implications.  First of all, the data suggest that classical AA-Jellinekian (AAJ's) are indeed more likely than non-AAJ's to combine the medical and moralist viewpoints:  47% of AAJ's and only 32% of non-AAJ's gave the hybrid view.  But it is still more interesting that rejectors of the classical model (non-AAJ's) do not congregate in the "consistently moralist" but rather in the "consistently medical" category.  Among non-AAJ's the modal ideological category, counting for 44% of the sub-sample, thought alcoholism an addiction and did not think alcoholics drank "because they want to." And that, in turn, suggests some interesting speculations:   it suggests that classical ideology, rather than secularizing alcoholismic behavior and enhancing the role of involuntary factors in it, in fact may perform quite contrarily.  Against the background of increasingly entrenched secularism in the U.S., three-quarters through the 20th century (Glock, 1964; Linsky, 1970; Orcutt, 1976), the disease concept that once championed scientific humanism may have instead become a sort of ideologically reactionary portal through which human will and spirit reappear in the explanation of alcoholismic drinking. In both the AAJ and non-AAJ subsamples, the presence of involuntarism seems well established: 78% of AAJ's and 76% of non-AAJ's agreed to the addiction formulation; but with regard to the presence of moralistic or voluntarist assumptions, the two subsamples differ sharply: among AAJ'S, 63% gave will a place in alcoholismic behavior, while among rejectors of the classical disease concept the proportion giving will the same explanatory role dropped to, 44%, less than half of that subsample. By implication, moralism is more strongly a part of the classical disease concept than it is a part of the background, non-AAJ climate of opinion. Of course, we must await replications of this table in other samples and settings, but in the interim these findings may serve as a whetstone for putting a new edge on contemporary views of alcoholism and the disease concept.

Viewed in its capacity as a vehicle for keeping open a lively role for personal choice, the classical disease concept of alcoholism seems poorly suited to reducing stigma on active alcoholism, not necessarily on enchancing the receptivity toward treatment. Instead, the picture is more intricate and more subtle.  In the presence of the classical disease concept, the alcoholic was equipped with an intact "will" and consciousness held captive by an overpowering biological (and later biopsychological) force.  The force, though, could itself be contained by never drinking. Thus, the alcoholic was Jekyll and Hyde, responsible for drinking the potion but helpless to counter its effects, and the moral status of the classically-defined alcoholic was arranged accordingly:

1. Jekyll sneaking out to Hyde's adventures is a pretty unsavory fellow, either as Jekyll or Hyde. Certainly a Jekyll-and-Hyde was on the moral scale far down below Jekyll before Hyde was discovered.
2. But Jekyll after forswearing Hyde was another matter:  suddenly, if the forswearation proves genuine, and if Hyde were terminated before some really unforgivable act spoiled even the reform, Jekyll seems not only restored to full moral status but even perhaps elevated slightly above pre-Hyde Jekyll: before, J's good behavior was innocent but, after Hyde's rejection, it took on the color of a triumph. Even the improprieties of the now dead Hyde might be excused or chalked up to J's ignorance of the peculiar effects of the potion on him.  J was offered, then, a status free of stigma in the future and considerable forgiveness for the past as well in exchange for Hyde.
3. But what if the informed J refused the contract or seemed to keep on breaching it every now and then?  Then, it seems plausible that J's moral status would sink below that J-and-H before J was educated about potion.  Now J was not only going into his Hyde act now and then, but also flauntingly rejecting the claims of others that he should stop taking potions.

We might now more fully elaborate the moral structure of the situation by varying certain critical features of the situation: for example, how are moral matters affected if:
a1 Jekyll agrees that Hyde's behavior is reprehensible, or
a2 He thinks Hyde's behavior, in fact, isn't too bad.
b1 Jekyll regards the potion as the cause of his transformation to Hyde, or
b2 Jekyll sincerely believes the potion is only coincidental and incidental to the transformation.
c1 Jekyll swears he doesn't care for Hyde any more than anybody else does, but continues taking potion nonetheless.
c2 Jekyll is discovered to be fond of Hyde's adventures in the right company, in spite of the assurances that Hyde is disliked.

The elaborations and permutations of these moralist elements are countless, and fascinating in their structure. But we will confine our attentions, for the present to four relatively simple statuses, involving two moral coordinates:  First, J's state of knowledge about potion -- is he aware and has he been informed that potion is making him Hyde and Hyde is bad, or is he innocent of this knowledge; second, is Jekyll making himself Hyde or has he stopped drinking potion for good. These two coordinates imply four Jekyll moral statuses.

                 J "drinking" potion:
 J knows potion's effects:
No Yes


In Cell A, we have Jekyll before Hyde has ever come on the scene -- his moral standing is independent of potion, Hyde and knowledge.  In Cell B, Jekyll is unconscious of potion's effect on him and Hyde's effects on others.  Strictly speaking, no blame should attach to Jekyll in this cell, but others who are aware of his potion practices might stigmatize and reject him for their own safety and wellbeing. But, especially because he appears uninformed about the potion's powers over him, the desirability of treatment will appear very high indeed among J's neighbors, while, of course, since he is innocent of the "realities," J himself will see no point for treatment.  In Cell C is Jekyll reformed, morally elevated above Hyde by an informed triumph over potion and Hyde, not stigmatized, unrejected, and, when his recovered status is established, able to use or not use supportive treatment as he wishes.  Finally, in Cell D is the unrecovered Jekyll, most blameworthy of all the cells -- and perhaps even more stigmatized than a straight Hyde character who never passes into a Jekyll.   Such a character would have no option but to be Hyde, while Jekyll has the knowledge and therefore the choice to not be Hyde.

The point of all this is simple enough -- with no knowledge of the science of the transformations of Jekyll and Hyde, the basic moral calculus consists of a Jekyll, a man to be accorded full moral status, and a Hyde, a man with something less than full rights and moral status. Let us define the moral distance between Jekyll and Hyde by the line,


But, now let us inform the world that Dr. Jekyll and Mr. Hyde are in fact the same person, that a special potion transforms one to the other, and so on.  Now the moral distance between these poles must be extended:  the distance from C to D (see Table) then becomes longer:


And, thus, the impact of this knowledge is not so much to reduce rejection (D seems more rejectable than a straight Hyde character), but rather to extend the sweep of moral definition and thus the range of social sanction (positive and negative) available for the social control of Jekyll.  The metaphor is directly applicable to the classical disease concept. Before disease alcoholism, the world consisted of drunks and respectables divided by a moral distance; but after the disease concept's widespread adoption, the world consisted of "recovered alcoholic," morally superior to themselves before alcoholism took hold, and vastly superior to active alcoholics who reject the theory that freely proffers them the moral status of recovered alcoholic.

 The disease concept, then, can be thought of as a conceptualization that upped the moral stakes surrounding deviant drinking. For recovered alcoholics, the promise was no stigma; but for alcoholics who kept on drinking, the rejection actually might have had reason to increase.  Thus the disease concept, instead of secularizing alcoholism -- taking morality out of the issue -- provided the social control of deviant drinking with a bigger carrot and a bigger stick.  And thus the most efficacious consequence was very likely, that the exchange of the moral status "active alcoholic" for the status "recovered alcoholic" came to be more profitable; or, put the other way around, sticking with the status "active alcoholic," as others around you grew to define you that way, became more and more morally expensive, and, ideally, the option to assume the recovered status might become more appealing.

Beliefs About Alcoholism and Willingness to be Treated

So far, we have confined our attention to the apparent acceptance of various elements of the public health view of alcoholism and we have not, as yet, analysed the patterns of relations among or the dependencies that exist between various elements of that perspective. We have seen that public agreement is both selective and quite responsive to differences in wording or slightly different cuts at the various elements of the public health view.  With respect to the reception of the notion that alcoholics should receive treatment, we have also seen that about 8 in 10 respondents give nominal support to the desirability of treatment, although the commitment to treatment suggested in these responses is less than complete: about half of the sample would seek treatment only when all other alternatives had been exhausted.  We turn you now to the general question of in what ways receptivity to treatment is associated with the adoption of the various elements of the public health view. We begin this analysis by looking in a little more detail at the measures of receptivity to treatment and the internal relations among them.

As can be seen back in Table 1, four different survey items tapped one or another aspect of the respondent's receptivity to alcoholism treatment.  The first (Table 1, item 6a) is a rather vague and weakly-worded item, one that no doubt touches upon the presense or absence of a general, fuzzy sense of favorability toward treatment.I  The second item2 differs from the first in two respects: first, it phrases the question in terms of the respondent himself rather than in terms of the generalized "you" used in the first item; second, the respondent is given a selection of four answer-categories (rather than merely "agree" or "disagree") with which to express his willingness to find alcoholism treatment if he thought he were an alcoholic.  The third item deals with the question of when the respondent might seek treatment, earlier or later in the course of the condition. And, finally, the fourth item broaches the question of whether the "public" support of such alcoholism clinics may be a source of'reluctance to be treated.

In Table 3 these three indicators of willingness to be treated or favorability toward alcoholism treatment are cross-tabulated against the various elements of the public health belief system on alcoholism. In addition, the fourth column of Table 3 reports the frequency of "high" (see footnote 1, Table 3) scores on a combined index of treatment willingness constructed out of items 6(a), 6(b), and 6(c).    In general, the "beliefs" variables show from modest to no associations with the four measures of willingness to be treated. The only apparent exception is belief 3a -- an item concerning alcoholism's progressiveness -- which seemed strongly to affect favorability toward treatment: among believers in alcoholism's progressiveness, 80% "probably" or "definitely" would seek out treatment for themselves if the need arose and 58% scored high on the treatment willingness index (TWI); but among rejectors of alcoholism's progressiveness, these percentages dropped, respectively, to 58% and 25%. But before we jump to the conclusion that belief in alcoholism's progressiveness provides the strongest ideological handle on receptivity toward alcoholism treatment, we should recall that the belief in progressiveness is nearly universal in this sample. Thus, rejectors of that belief may represent the deviant attitudinal sediment left over after 90% of the sample has been poured off; more importantly, the strongly skewed progressiveness marginal implies that most treatment-likers and treatment-dislikers are being drawn from the ranks of people who believe in alcoholism's progressiveness.  The findings imply that if all of the disbelievers in progressiveness could be converted to believers, the overall fraction of the sample approving of alcoholism treatment would increase on each of the measures by only about 3 percentage points.

Aside from the belief in progressiveness, Table 3 shows only modestly strong associations at best, varying between epsilons (i.e., simple percentage differences between positive and negative "believers" in treatment receptivity) of zero and 12 points. Looking at the TWI, the stronger of these associations appeared for (1) belief that alcoholics are addicted, (2) that one would not want to remain ignorant of one's own alcoholism, and (3) that "most alcoholics do overcome" their troubles if they get treatment. But even these relatively stronger variables show some instability of associational strengths across the different measures of treatment receptivity in Table 3: thus, for example, the availability of medical tests shows a relatively stronger association with treatment willingness as it is operationalized by variable 6(b) but a rather weaker one when run with the TWI.

In Table 3a, willingness to be treated as it is measured by item 6(b) [" likely would you be to search out an alcoholism treatment program?"] has been regressed on the full collection of beliefs about alcoholism that we examined in Table 3.  Taken together, this collection achieved a multiple correlation coefficient of .24 with the treatment-willingness variable, thus, only 6% of the willingness variance was explained by these belief variables. While the multiple correlation is low, it does suggest some additivity among the beliefs. The zero-order correlation coefficients ranged from 0 to small; most were tiny.

Finally, in Table 3b is presented the varimax factor pattern for the "belief" and "willingness to be treated" variables. This pattern shows all three of the measures-of-treatment-willingness with respectable loadings on the first factor, a willingness-to-be-treated factor. But, except for variable 3(a) concerning alcoholism's progressiveness (and about whose association with willingness we have already discussed and suggested caution), the treatment-willingness items are alone on that factor. On the second factor, various indicators of stigma seem to have congregated together, and the third factor suggests dimension reflecting "belief in medicine" per se: Joined together here were the beliefs that medical tests for alcoholism are widely available, that they treat one respectfully in alcoholism treatment programs, and that people who get treatment do overcome their problems.  Regarded with a little imagination, the fourth factor suggests a sort of "anti-treatment/pro- rationalization" dimension: strung together, the variables picked out in factor four suggest a convenient vocabulary for expressing fear and distaste of treatment: "news would get out" if one went to treatment; alcohol problems are often "nothing serious" anyway; I know the danger signs of alcoholism; and, to be known as an alcoholic would destroy a man's reputation.  Factor five seems to reflect the ideological axis stretched between volitional and involitional conceptions of alcoholism; and factor six selects out two variables joining the opinion that alcoholism treatment depends completely on the alcoholic's motivation, on the one hand, with the opinion that one would only seek alcoholism treatment when all other alternatives had been exhausted, on the other: a seeming "hit-bottom" factor.

The principal component solution for this matrix of belief and treatment-willingness variables (not shown here) congregated around the three measures of treatment-willingness.  Belief variables with loadings higher than .3 on the principal factor were: [no] shame attached to going to a public clinic (Table 1, item 6(d), actually a willingness item too); belief in alcoholism's progressiveness; [disagreement with the notion that] alcoholism would destroy a man's reputation (item 2a); and [disagreement with the statement that] it would be better not to know about one's alcoholism if one "had" it.

Of course, the form of these data deviates substantially from the assumptions of correlation-based multivariate techniques -- for instance, the attitudinal variables are dichotomous rather than continuous, and the distributions are often quite skewed in addition to being, non-normally shaped and thus they should be interpreted with a grain of salt.

Denial of Alcoholism

One of the best-known and most often cited explanations of reluctance to be treated among alcoholics is contained in the concept "denial." Hartocollis (1968), for example, writes:

The fact that alcoholism has been officially recognized as an illness ...may indicate a certain change in the attitude of the public but not necessarily in the mind of the patient.   Indeed, one of the most frequent and most baffling experiences for those who attempt to treat patients with drinking problems is to hear patients deny that they have any problems for which they need help.

Similarly, Moore and Murphy (1961) wrote that "The characteristic difficulty of the alcoholic in accepting the reality of his illness is familiar to all who have attempted to treat him." Here we will examine the hypothesis that alcoholics often avoid treatment because they do not regard themselves as alcoholic.

Ironically, the fact that denial if often regarded as a major obstacle of improvement among alcoholics in treatment implies that "denial" may not provide a very good explanation of why alcoholics avoid treatment: if alcoholics inside and outside of treatment proved to be equally insistent deniers of alcoholism, then logically denial could account for why some alcoholics were in and others not in treatment. Thus, explanations built around the concept seem to be better regarded as universalist rather than differential accounts.

As a theory of treatment reluctance, the denial concept suggests that alcoholics often are unable or unwilling to recognize or accept a diagnosis that to others appears obvious. In the literature, "denial" is often handled in two rather different ways: one focusing on it as a psychological defense mechanism, and the other as a consciously constructed program of deceit and evasion. Often, of course, elements of both unconscious and conscious are present in the same putative alcoholic (Moore and Murphy, 1961) though, of  course, the clinical response would be affected by which of the two was thought to define its essence. The question of whether denial should be regarded in psychoanalytic or in legalistic terms is also important in sorting out what sort of a theory of treatment reluctance the concept provides. In its more psychoanalytic form, a denial theory of reluctance fits well with a broader psychoanalytic interpretation of alcoholism itself indeed, it could be argued that the patient's inability to recognize his own deviant drinking amounted to a forceful proof of the drinking's status as symptom of underlying intrapsychic conflict. If alcoholism were a "simple addiction" which did not involve an ultimately psychological seat, then it might be expected that the addiction would not so often occasion denial.  Thus, in its psychoanalytic form the theory of denial blends easily into a more psychoanalytic sort of picture of treatment reluctance involving unconscious rather than conscious sources of resistance.

On the other hand, a more legalistic conception of denial tends to give way to discussions of the "reasons" why alcoholics persist in denying their illness.  These, of course, may include fear of loss of social standing because of alcoholism's stigma, fear of the abstinence requirement, or even arguable differences between the patient and the diagnoser with respect to alcoholisms' essential meaning and diagnostic criteria. Bigus (1973) reproted that patients in alcoholism treatment often reported performing drinking experiments on themselves in order to test (and usually to disconfirm) the existence of "loss of control" over their own drinking. In its more legalistic form, a denial theory of reluctance takes on a bargaining or game-theory appearance in which, fundamentally, different parties might be thought of as regarding the advantages and disadvantages of treatment differently.

Cahalan and Treiman's (1976) data again provide an unusual opportunity to examine the role of denial in the resistance to alcoholism treatment.  Their survey included both a series of items concerning alcohol-related problems and symptoms and, as well, another series concerning the respondent's view of his own drinking and its implications. Thus, by comparing these two sets of data it may be possible to assess the prevalence and significance of denial among untreated alcoholics in the general population.

Of course, in order to look for denial among alcoholics it will be necessary to discriminate alcoholics from nonalcoholics in the sample. For the purposes of this analysis respondents were sorted into a five-category typology of drinking statuses: (1) Treated Alcoholics: all respondents who were currently in treatment for alcoholism or who reported having been treated in the past (N=29); (2) Current Abstainers: all respondents who reported drinking no alcohol during the past year (N:=127); (3) Probable Alcoholics: respondents who scored three or more points on an alcoholism diagnostic scale based on Woodruff et al., (1973) (N=36); (4) Possible Alcoholics: respondents scoring 2-3 points on this same scale (N=58); and finally, (5) Current Drinkers:  respondents who qualified for none of the other categories (N=889).  In the sample as a whole, then, we found the following frequency distribution:  Treated Alcoholics, 2.5%; Abstainers, 11.2%; Current Drinkers, 78.1%; Possible Alcoholics, 5.1%; and,-Probable Alcoholics, 3.2%.

A number of questions in this survey examined the question of the respondents' perception of their own drinking and the degree to which respondents may have regarded their drinking as problematic. For example, respondents were asked, "Which one of the following terms would best describe your drinking at the  present time?"  The response categories and distribution were as follows: 20% of the sample described themselves as "A teetotaler or nondrinker"; 29% as "Almost a teetotaler or a very light drinker"; 42% as "A social drinker"; 2% as "A problem drinker or alcoholic"; 1% as "An ex-problem drinker or ex-alcoholic"; and 6%, other or no answer. Thus, in this self-characterization item only about 3% of the sample called themselves past or present alcoholics or problem drinkers, while in the diagnostic typology described above a little over 10% of the sample either had had treatment for alcoholism or qualified as a "probable" or "possible" alcoholic based on their responses to a variety of drinking-problem items.  The cross-tabulation of the self-characterization and the diagnostic typology is of interest:


                         Drinker Self-Characterizations:
Other Tot. (N)
Treated Alcoholics 14% 14 21 28 14 10 101% (29)
Abstainers 88% 5 2 0 0 6 101% (127)
Current drinkers 13% 36 46 * * 5 100% (889)
Possible alcoholics 0% 7 69 5 2 17 100% (58)
Probable alcoholics 3% 3 39 33 11 11 100% (36)
Total Sample 20% 29 42 2 1 6 100% (1139)
Several interesting findings are suggested in Table 4: First, if we regard as deniers respondents who qualified for one of the three alcoholic labels but who did not indicate alcoholism or problem drinking in their response to the self-characterization item, then, clearly, most of these alcoholics seem to deny their alcoholism. But, on the other hand, denial appears to be far from universal among the untreated alcoholics (indeed, fully one-third of the "probable alcoholics" identified themselves among untreated alcohOI3"es as among treated cases. Though the small numbers must be kept in mind, the much greater admission of alcoholism among probable rather than possible alcoholics suggests the possibility that judgments of drinking practices and problems become rather sensitive up at this end of the drinking-problem distribution.  It is also noteworthy, perhaps, for its significance as a subtler form of denial, that all three of the alcoholic groups and particularly the "possible" alcoholics relatively more frequently employed the self-styled option of writing in an "other" response or leaving the question blank.

Of course, the self-characterization question on which Table 4 was based concerned only the respondent's best characterization of his drinking, and thus it does not pick up gradations of self-definitions or some of the middle ground between defining oneself as an alcoholic or a nonalcoholic.  Fortunately, two other survey items registered something of this gradient:  The first asked whether the respondent was sure he was not an alcoholic; a yes answer to this item can be regarded as the strongest declaration of nonalcoholism that the respondent could make in this survey, a no answer, on the other hand, may be thought of as an indication even of quite weak suspicisons of alcoholism -- even on the basis of a passing doubt about it, the respondent might indicate that he was not completely sure of his nonalcoholism.  The second item served to extend the time frame of the respondent's self-characterizational response:  it asked, "Were you ever afraid that you might be an alcoholic or that you might become one?" The response categories were: Yes, during the last 12 months; Yes, but it was 1-3 years ago; Yes, but it was more than three years ago; and No, never. Thus, these two items better may reveal self-characterizational points between the two poles, or the gradient of denial that may be found in the alcoholic part of the sample.

 Table 5 reports the frequency distributions on these items. Though in Table 4 from five to thrity-three percent of the alcoholic subgroups regarded "alcoholic or ptoblem drinker" as their best drinking self-characterization, a considerably larger fraction has  considered the possibility and not wholly rejected it.  Among probable alcoholics fully 63% did not indicate that they were sure that they were not alcoholic, and roughly three-quarters agreed that they were afraid they might be alcoholic or might become one.  Among "possible" alcoholics the proportions showing doubts about nonalcoholism run about half as high as among "probable" alcoholics, but they are still substantial:  31% indicated they might be alcoholic, 38% that they had feared alcoholism or the development of alcoholism in the last year, and 45% that they had ever xperienced such fears.  (One caveat, however, should be entered here about the figures in the second and third columns: this item is independent of the diagnostic typology by which respondents were sorted into possible and probable alcoholic categories -- it was among some 46 items that figured in that diagnostic scale. The first-column item was not part of the diagnostic inventory, but, of course, it is highly correlated with the included item.)

In Table 6, below, the items employed in Tables 4 and 5, above, are combined into a single Self-Characterization Scale which sorts respondents into those who pretty definitely regard themselves as alcoholics (those who gave "alcoholic" responses to all of the items), those who equally definitely regard themselves as nonalcoholies (all "nonalcoholic" responses, including agreement to the statement, "I am sure that I am not an alcoholic"), and those with mixed patterns of response suggesting some unsureness (all others).  This table, then, suggests that denial is neither ubiquitious among untreated alcoholics nor always complete among those who do some denying.  More than a quarter of the untreated "probable" alcoholics and at least 5% of the "possibles" flatly regard themselves alcoholic; 41% of the ,possibles" and 17% of the "probable" alcoholics indicated a flat denial of alcoholism; and over 50% of these two groups fell in between.

Thus, these data should reopen the question of how much the resistance to treatment among untreated alcoholics should be attributed to denial.  About one in eight untreated alcoholics in our sample accepts the label but nevertheless has not sought out treatment; a much larger fraction of the sample, perhaps more than half of the group that qualified as treatment candidates, is on the fence about their drinking status but apparently unexposed to treatment and not particularly anxious to seek it out. And thus, these data suggest the desirability of keeping separate the level of denial among alcoholics and their receptivity toward treatment. The utility of this separation is emphasized in Table 7, below, in which are cross-tabulated the "level of denial" and the "willingness to be treated if one were an alcoholic" among the possible and probable alcoholic subgroup in the sample. This table, then, provides two explanatory axes for the untreated alcoholic's untreatedness: as we move from right to left across the columns, the amount of denial increases, and, equivalently, the potential for an explanation in terms of denial increases as well; as we move from top row to bottom on the table, however, the amount of openly-expressed disinclination toward treatment (regardless of denial) increases, and equivalently, the potential for an explanation of untreatedness in terms of disinclining factors is increased. The corner cells of the table correspond to the potential relations between these two sorts of explanation. The upper-left corner cell contains cases who are the best candidates for an account of their untreatedness that relies heavily on denial -- they have reported in their questionnaires that they "definitely would seek out treatment" if they thought they had a serious drinking problems and they have reported that they "definitely" are not alcoholic.  Conversely, the lower-right corner of the table gives us respondents who openly regard themselves as alcoholics but at the same time indicate that alcoholism treatment is not for them -- they, in short, seem to require an explanation that focuses on distaste for treatment rather than mechanisms of denial. In the lower-left corner are cases whose untreatedness is "overdetermined" and might be traced either to denial or distaste-for-treatment or portions of both. And, finally, in the upper-right corner are cases who both regard themselves alcoholic and say that alcoholism treatment is desirable -- these respondents presumably require a situational rather than an attitudinal account of their untreatedness.

(corner percentaged, N=94)
willingness to accept treatment: not an alcoholic (high denial) May be an alcoholic definitely an alcoholic (low denial) total
high 11% 9% 2% 22%
medium 12% 22% 4% 38%
low 10% 23% 7% 40%
total 33% 54% 13% 100%
The distribution of cases in Table 7 suggests a couple of observations:  First, taken individually, "denial" and "distaste for treatment" appear to be about equally prevalent among untreated alcoholics: for example, a third of this subgroup denies alcoholism entirely and 40% indicate no taste for alcoholism treatment "even if I were an alcoholic." If. for the sake of comparison, we consider the lower-left to upper-right diagonal as the axis of cases in which denial and distaste-for-treatment are equally vigorous in the alcoholic's treatment resistance, then the cases above the diagonal represent better candidates for a "denial" explanation (31%) and the cases below the diagonal better candidates for "distaste" explanations (34%).  Clearly, the untreated subpopulation appears to be quite evenly distributed to "diagonals, deniers, and distasters" as they are operationalized in this analysis.

It will prove worthwhile to consider separately "denial" and "distaste" treatment candidates in the analysis to follow, and in a moment I want to turn to an analysis of the sources of denial in-this subsample.  But before turning there is is important for a moment to stand back from these data and consider some of the meanings of denial in the context of a social survey:  all of the data analyzed here came originally from the survey respondents -- they reported on their own drinking practices, drinking-related problems, and as well their definitions of drinking status.  And because of this fundamental fact, we ought to be very careful about how, exactly, denial is defined and regarded.

Denial in the Context of Survey Research

As I previously noted, discussing survey data showing a good deal of independence between objective measures of problem drinking and self-characterizations of drinking status, the "denial" account of this disparity presentsus with some difficulties:

From a clinical perspective it could be argued that we are seeing denial at work. But this explanation is somewhat weakened by the fact that it is the respondent himself who has reported the drinking problems in household surveys.  Thus, denial or distortion does not extend so far as to cause the respondent to gloss over or fail to report drinking problems on the items that register that information.  While denial is no doubt at work in some cases, the simplest interpretation is that most respondents put a milder meaning on their drinking problems than might be put by survey analysts and other observers.  For a substantial portion of the respondents who report even relatively severe drinking problems, these problems do not seem to imply that their drinking is deviant or that they are deserving of a stigmatized label.  Or, stated from the respondents' point of view, normal drinking is often not incompatible with getting into a number of problems with drinking. (Roizen, 1974, p. 19)

The Cahalan and Treiman (1976) study, from which these data were drawn, involved a survey through the mail, followed-up by telephone requests, and, finally, by personal interviews where the preceding methods had failed to generate a response. Of the households sent one, 72% returned a completed questionnaire (see Cahalan and Treimen, 1976, pp. 97-103 for a description of their field methods).  Thus,, some respondents simply refused to contribute information about themselves, and among them the potential respondents who would have qualified as alcoholic (let us call them "Type I Deniers") could be regarded as denying their alcoholism -- though alcoholic, they revealed nothing of their drinking practices and problems in this study's "encounter" with them.

A second group of "denying" respondents -- it is hoped a small one -- sent back completed questionnaires but did not report drinking-related symptoms or heavy consumption practices. In a clinical context, this second sort of denial (Type II deniers) often is regarded as quite common among alcoholic patients and would-be patients. But is is arguable whether this sort is nearly so common in survey studies: in the clinical setting, an interview is surrounded by a "strong field of forces" (Room, forthcoming; Robins, 1966) that involves control over the interviewee's circumstances and fate, and may be regarded by him as essentially accusatory.  Though, when properly handled, it appears that clinical interviews can yield quite valid self-reports of drinking behavior (see Ruprecht, 1970, p. 170 and Sobell and Sobell, 1975), survey studies, of course, are largely unhampered by some of the sources of distortion in clinical settings.  Survey studies have no fate-control and are not, except in very rare cases, regarded accusatorily.  Presumably, assurances of confidentiality or anonymity and the promise to use information in "statistical reports only" further enhance candor.  In a mail survey, moreover, distorted questionnaires may well be more likely never to be mailed back -- it seems plausible that returning a distorted questionnaire in the mail involved not only denial but the desire to sabotage the study, a desire that is presumably rare in the general population.

A third group (Type III Deniers) accurately and consistently reports their drinking behavior, but doesn not regard it as implicative of alcoholism or problem drinking. This sort of denial, too, is common in clinical settings where patients acknowledge arrests for drunkenness, blackouts, or binges but stoutly resist the alcoholic label. Smith (1961, p. describes one such common denial tactic:

It may be difficult to understand how the patient can continue to regard himself as a nonalcoholic in view of his drinking behavior; but this paradox is resolved by his definition of an alcoholic, which always excludes him.  The exclusion is comfortably accomplished by taking some one facet of the problem, which he lacks, as essential to the diagnosis.  For example, the patient who habitually drinks in bars says, 'I never drink alone.'  The alcoholic who drinks alone says, 'I never drink in the morning.'  Perhaps the most widely used rationalizations are, 'I drink like anybody else,' or 'I only drink beer and never touch the "hard stuff" therefore I'm not an alcoholic."'

This sort of "denial," if taken at face value, is in fact "disagreement" over either the essential character of alcoholism or the patient's conformity to that character.  Type III Denial, then, earns its status as denial only if we add to this "disagreement" the tacit belief that the diagnoser (the physician, or in the case of the present study, the survey analyst) is correct and the patient-respondent is wrong.  In other words, without adding in the belief that the patient should adopt the diagnoser's view of the situation (either because of the "facts" of the case of because of the diagnoser's authority) we are left with little more  than a simple disagreement over what the word "alcoholism" means or should mean and whom it should and should not be applied to.

Ordinarily, the forensic position of the diagnoser is strengthened by his rank and by his professional (or paraprofessional) capacity to drawn on a body of expert knowledge that allows him to see the patient's condition more accurately than the patient himself.  But, though this picture of professional wisdom and consensus is an important feature of the diagnostic situation, in fact is often more honestly regarded a performance staged for its therapeutic potential and does not at all convey the squabbling over definitions, etiology, and even questionings of the existence of "alcoholism" in the scholarly and professional literatures.  In short, the state of knowledge about deviant drinking simply does not permit us readily to regard the patient-respondent as incorrect and the physician-survey analyst as correct and to proceed on that basis.  And thus we should preface and introduce the following analysis of "denial" by noting that we use the term, following conventional usage, to denote a person's disagreement with the "alcoholism diagnosis" as it is operationalized by some authority (in this case, my adaptation of Woodruff, et al., 1973) in a diagnostic test for alcoholism -- but, we should not be at all certain whether the "denier" or the analyst is correct in making that diagnosis.  Having said all this, we may turn now to look at some of the putative sources of denial among alcoholics, beginning with just these differences in definition.

Denial and Definitions of Alcoholism

We saw in Table 7 that "possible" and "probable" alcoholics could be sorted into three groups, those high in denial (33%), those medium (54%), and low (13%).  As it happens, Cahalan and Treimen asked respondents about drinking-related behaviors and problems in nine different problem areas.  Table 8 reports the frequencies of these various drinking problems in four sample subgroups:  among nonalcoholic current drinkers, and among high-denying, medium-denying, and low-denying alcoholics as these last three categories have been defined above.  This table, then, ought to show us something of the sorts of problems that behave as criteria of alcoholism among untreated alcoholics and the sorts that do not behave criterially.  If, for example, all treatment candidates who did not deny alcoholism repoted drinking problem A and among treatment candidates who did deny alcoholism drinking problem A were quite rare, then this finding would suggest the possibility that drinking problems A is often regarded as a criterial attribute of alcoholism among untreated alcoholics in the general population.  Of course, strictly speaking the finding is wholly post hoc, ergo propter hoc, and thus should be regarded as exploratory only.  I should also mention another caveat for Table 8:  it will be recalled that "possible" alcoholics were more frequently "high deniers" of alcoholism than were "probable" alcoholics (see Table 6, p. 11).  This implies that there exists not only a difference in the level of denial across the three alcoholic-denier groups, but also a difference in the "average" severity of the alcoholic symptoms.  Controlling for severity of alcoholism in Table 8, though, would make the N's too small for consideration and thus, for now, this is a confounding of variables we will have to live with.

It should also be noted that many of these problems scales (again) were employed in the diagnostic test we have used to cull alcoholics from the rest of the sample.  Thus, it is not the absolute frequency of a particular problem in the alcoholics that we should pay attention to (that frequency will vary from problem to problem, depending, in part, on whether the problem was used in the diagnostic scale), but rather variations in frequencies across high, medium, and low-denying alcoholics.  Nonalcoholic current drinkers are included in the table merely to provide a benchmark for the prevalence of these sorts of problems in the nonalcoholic general population, but again the reader should keep in mind that the relative disparity in frequencies between "alcoholics" and "nonalcoholics" is increased by the fact that many drinking-problem scales were used as criteria of alcoholism as well.

In Table 8, Cahalan and Treiman's drinking problem scales have been arranged into three groups: (A) Role Performance Problems: i.e., frictions or breaks in relationships due to drinking with others at home, at work, or elsewhere, and losses of jobs or trouble with the police; (B) Intake and Health Problems: problems defined by a relatively high level of alcohol consumption, per se, by drinking bouts lasting longer than one day, or by either warnings from a physician that alcohol was hurting the respondent's health or the occurrence of accidents or injuries due to the respondent's drinking; and (C) Existential Problems:  that is to say, problems that may be problematic only to the respondent himself and-not necessarily to others in his social environment -- fear of losing control over drinking, for example, and a variety of problem experiences drawn from Jellinek's (1952) original symptomatology.  Cahalan and Treiman (1976, pp. 18, 20) briefly described the individual problem scales as follows:

Role Performance Problems:

(1) Job Problems: ranging from someone at work showing concern about respondent's drinking, up to losing a job because of drinking.

(2) Spouse Problems: ranges from the spouse's showing concern or getting upset about respondent's drinking, up to separation or divorce because of drinking.

(3) Interpersonal (Nonspouse) Problems: ranging from friend or relative (but neither spouse nor person from work) showing concern about respondent's drinking, up to breaking up a close friendship or other relationship.

(4) Police Problems: drunk driving arrest or other trouble with the law related to respondent's drinking . . .

Intake and Health Problems:

(1) Current Heavy Intake: reported having had at least eight drinks one or more days during the prior month, or four or more drinks upon four or more days during the month.

(2) Bingeing: stayed drunk for more than one day at a time during the last year.

(3) Health or Injury Problems: doctor had advised respondent during the last year that drinking Was having a bad effect on his or her health or drinking caused respondent to suffer an accident or injury in the past year.

Existential Problems:

(1)  Loss of Control: one or more of the following within the year -- getting drunk when not intending to, usually drinking until passing out, unable to stop drinking or to cut down on drinking.

(2)  Symptomatic Behaviors: reported having had one or more of the following within the year -- drinking first thing in the morning, having several drinks before going to a party in order to have enough, sneaking drinks, drinking more when alone, drinking to shake off a hangover, being unable to remember events of the evening upon awakening the next day.

In examining the figures reported in Table 8, it is well to keep in mind the very small numbers of "low deniers": I have included them in the table because collapsing them into the "medium" denying group only loses information for the reader, but certainly the low-deniers column should be regarded with caution.  What, then, emerges from Table 8?

It seems that problems of role performance associated with drinking do not figure strongly in whether or not the alcoholic denies: interpersonal (nonspouse) and police problems are equivalent in the three denial groups; and if anything, job and spouse problems are less common among low and medium deniers than among high deniers, but the relationships are neither strong nor monotonic.  But the remaining drinking problem groups (Intake and Health problems and existential problems) suggest an interesting picture of denial:  While going on binges does not differentiate denial levels, heavy intake shows signs of a modest relationship (the more common low intake, the more likely denial) with denial, and "health or injury problems" show signs of a strong one:  though the N's are too small to trust, it appears likely that drinking in spite of health conditions or accidental injuries due to drinking is regarded as strongly criterial of alcoholism in this group.  Similarly, the existential problems suggest that signs of loss of control are closest to the heart of the untreated alcoholic's image of alcoholism (though 43% of "high deniers" nevertheless report loss of control signs):  low deniers universally acknowledge loss of control, three quarters of "medium deniers" acknowledge it, and less than half the "high deniers" do.  On the other hand, "symptomatic behaviors," a scale that taps into some of the other classical symptoms of alcoholism is unrelated, or perhaps even very slightly negatively related to denial.  Thus, perhaps echoing Smith (1961), many symptoms of alcoholism can be reported without implying one is alcoholic, but signs of loss of control stand out among the alcoholism symptoms and may be regarded as virtually pathognomonic to many in this sample.

In short, first "loss of control" and second "health and accident" problems appear to be most strongly criterial of alcoholism in the minds of the untreated alcoholics in this sample, and, equally interesting, many other sorts of drinking-related problems seem by themselves to have little or no criteriality.  Especially job, spouse, interpersonal, and even police problems appear to be unimplicative of alcoholism so long as the respondent maintains a sense of personal control over drinking behavior.  Significantly, only continued drinking in spite of bodily harm may be regarded as an exterior sign of alcoholism -- the other "non-bodily" harms, it appears, are more easily traced to nonalcoholismic sources.

Denial's Commonly Cited Roots

In the preceding analysis we have taken what might be called a "rationalist approach" to denial: "denial" itself has been regarded in the master image of "disagreement," and we have tried to look for the criteria attributes of alcoholism among untreated alcoholics with varying amounts of denial in order to tease out the major features of this "disagreement."  But, of course, rationalist approaches to denial are infrequent, and if we stopped here we    would have ignored most of the hypothetical models of denial suggested by alcohologists. Hence, I would like now to consider, in rapid order, a variety of explanations of denial in the literature and in commonsense knowledge, and I will try, where possible, to test these hypotheses against the Cahalan and Treiman data. First,

(1) Denial and Gradualism (see Ruprecht, 1970, and Clancy, 1964): "When a disease has a march of symptoms that spans years instead of months, a patient's alarm at changes is lulled" (Ruprecht, 1970, p. 168).  The suggestion, of course, is that alcoholics fail to notice their alcoholism in quite the same way that parents do not notice the growth of their children (unless separated from them for a time): it all happens too slowly. In my opinion, the role of time in denial is very likely to be profound but difficult to assess with cross-sectional data (we will have occasion later on to again raise questions about the rapidity of change in donditions in alcoholism without having a satisfactory means for testing the question with these data).  It might be argued that the gradualism hypothesis should show a relationship between age of the alcoholic and level of denial: the reasoning is as follows, if alcoholism is hard to recognize in the slow evolution of its stages, then only when the disease has run or nearly run its full course will the alcoholic see in sharper relief how much distance has been traversed.  The test is flimsy, but the best, at the moment, we have to offer:

Among Currently Drinking Nonalcoholics (N=888) the median age in the Cahalan and Treiman sample fell within the age-group 30-34; 40% were younger than 30.  Among self-acknowledged alcoholics (N=13), the median age fell in the group aged 35-39, and only 23% were younger than 30. Among medium-denying alcoholics, median age was again 35-39, and 23% younger than 30.  But among high-denying alcoholics, median age dropped to 30-34 and fully 47% were younger than 30.  Finally, among all treated alcoholics (N=29) median are rose to 45-49, and only 1 case (3%) was younger than 30.
Thus, age appears to be associated with decreasing denial, but, of course, not only the hypothesis of the importance of "gradualism" in denial would suggest this sort of association between age and denial.

(2) Denial and Avoidance of  the Abstinence Requirement:  "There is considerable support," write Lloyd and Salzburg (1975, citing on the point Brunner-Orne, 1963; Gerard , Saenger, and Wile, 1962; Pattison, 1966 and 1967; and Reinert, 1968), "for the belief that many alcohol abusers continue to deny that they have a problem with drinking until it becomes chronic and excessive, largely because of resistance to being condemned to a life of abstinence" (p. 821).  This hypothesis, of course, has a rationalist ring to it -- it is, in another form, an expression of the ancient fear that the cure is worse than the illness. That alcoholics would prefer not to give up drinking seems so commonsensical as not to require demonstration, but it will prove worthwhile, I think, to consider some of the intricacies of this seemingly simple hypothesis:

First of all, let us state the hypothesis clearly:  "Many alcoholics deny their alcoholism because to accept the label would for them imply that they must give up drinking permanently and entirely."  In propositional form, the hypothesis can be stated:  "If being an A implies B, and B is undesirable to me, then I will not regard myself as an A." In its psychoanalytic form, the alcoholic may be (or, more to the point, should be) unaware that his rationalizations of alcoholism amount to an unconscious defense mechanism deployed in order to protect his symptom. As a conscious or legalistic sort of denial, on the other hand, rejecting alcoholism is regardable as part of a strategy for staying out of the hands of the impreratives implied by acceptance of the alcoholism label.

The hypothesis seems to imply that the denying alcoholic accepts the notion that genuine alcoholics must abstain -- if this premise were not accepted, then the "denial of one's own alcoholism" would in fact slide into the "denial of conventional wisdom on alcoholism   a more openly argumentative and intellectualist stance. In the Cahalan and Treiman sample respondents were asked two relevant questions: first, (agree or disagree) "If a man is really an alcoholic, he will have to quit drinking forever -- he will never be able to drink moderately again," and second, "Many alcoholics learn to 'taper off' and get their drinking under control again." Among untreated alcoholics as a group (N=93), 66% agreed that a genuine alcoholic must give up drinking (among treated alcoholics agreement was 77%; among currently drinking nonalcoholies, 70%) and 70% rejected the idea that alcoholics could "taper off" (among treated alcoholics, 61% rejected; among CDNS, 72% rejected). If we now examine the three denial groups within the untreated alcoholic subsample, the data showed that alcoholic respondents who denied completely that they were alcoholic were less likely than nondeniers to agree that genuine alcoholics must quit (53% agreed among high deniers; among medium deniers; 70% agreed; and among nondeniers, 77% agreed). Complete deniers were also more likely to think that alcoholics often "taper off" to normal drinking: among medium deniers 25% agreed; and among nondeniers 31% agreed).  Thus, it appears that high denial of alcoholism may often involve not only a personal rejection of the diagnosis, but also a more sweeping rejection of perhaps the most important pragmatic implication of that diagnosis, the need for abstinence.

We see in these data, then, the prosepct of a more wide-ranging sort of rejection of conventional vidsom, both in general as it applied to everyone and in particular as it applied to oneself, among high-denying untreate alcoholics (UAs).  Following Paredes (1974), we may expand the frame of denial to include such possibilities as: (1) denial of being ill or requiring some sort of treatment; (2) denial of the right of others to deprive one of drinking; denial of the existence of alcoholism, its disease status, and so on; (3) denial of medical (or other) authority to diagnose or expert knowledge on alcoholism; (4) denial that drinking is a cause of life troubles and presentation of countertheories.

(3) Denial and Stigma:  "The social stigma still attached to alcoholism accounts for considerable rationalization by alcoholics, as well as outright refusal of treatment" ([Canadian] Commission of Inquiry into the Nonmedical Use of Drugs, 1972, p. 42).  The crosstabulation of levels of denial among UAs with four measures of perceived social stigma on alcoholism produces a mixed picture.  As Table 9 shows, high-deniers are more likely than medium- and low-deniers to think the alcoholic label would "destroy a man's reputation," to think the alcoholic's privacy should be protected, and to think "the news would get out" if one went into alcoholism treatment.  Regarding the amount of respect alcoholics will receive in treatment, however, the small group of self-confessed alcoholics indicated agreement more often than the high-deniers.  Moreover, three of the four patterns of association between stigma measures and denial level are nonmonotonic., further suggesting only a cautious interpretation.

(% Agree -- see Table 1 for full wording of items)
Stigma on alcoholism: high-
...destroys reputation 53% 28% 39%
...protect privacy 57% 40% 46% would get out 59% 55% 39%
...treat you without respect 52% 38% 62%
The "Hidden Alcoholic"

Sooner or later, a discussion of barriers to alcoholism treatment will have to take up the subject of the "hidden alcoholic."  As the name suggests, this is a phenomenon that is obscured from our view, We "see" the hiden alcoholic mostly by projecting on-the world outside the door several of the most important assumptions of the alcoholism' conception of deviant drinking.  In some cases this projective character of hidden alcoholism is laid out quite unabashedly:  thus, for example, both O'Hallaren and Wellman (1958) and Rubington (1972) arrived at a figure for the prevalence of "hidden alcoholism" in the U.S. by subtracting from the Jellinek Formula estimate of the total number of alcoholics the number in some sort of alcoholism treatment.  "Where are the remainder?" O'Hollaren and Wellman (1958) ask, "The answer seems to be that they are 'hidden' -- hidden from general recognition."

But the place of the hidden alcoholic in American thought is not only a function of the disparities in size of the populations of alcoholics inside and outside of treatment. Given that disease alcoholism takes the form of a condition that is "no respecter" of social class, sex, education, and so on, it must follow that the condition is more or less evenly distributed throughout the social class (etc.) structure.  Thus, if alcoholism treatment services, on the other hand., cater to only one or two sectors of the needful population, or if, in other words, the population in treatment is not a representative sample, so to speak, of the general adult population, then it follows that a substantial population is "hidden" from these services and from recognition as alcoholic.  Moreover, we can make a pretty good guess at the composition of this population by equating it with the complement group that would fill out the in-treatment population such that it would then become representative of the general distribution of alcoholism in the population.

The case for a substantial hidden alcoholic population based on the assumption of alcoholism's "general distribution" through modern society was stated as long ago as in 1867, by A. Day, a superintendent of the Washington Home for Fallen Men (quoted in Schmidt et al., 1968, p. 5):

It seems to me that the very general prevalence of the disease is not appreciated. Everyone knows of one or more cases, but they are looked upon as infrequent exceptions.  It must be borne in mind that a large proportion of these cases are not visible to the eye of society, but are hidden from the world by the decent pride of friends, or the sensitivity of the patients themselves. No social limits are narrow enough to exclude it, and it knows no distinction of sex.

Even from this early date, the hidden alcoholic was attributed with sufficient wealth and control over himself to keep most of the obligations of conventional life.  Over the years since the repeal of National Prohibition, the relative contributions of "hidden" and "unhidden" alcoholics to the overall alcoholic population seems to have shifted markedly toward greater numbers of hiders:  As recently as 1953, Henderson and Bacon estimated that "less than half, perhaps only a quarter" of the total population of alcoholics fit the traditional stereotypes of the psychotic or deeply neurotic alcoholic and the Skid Row type.  By 1957, Wellman et al. estimated that only 8% of the alcoholic population was on Skid Row and only a small fraction in Alcoholics Anonymous. O'Hollaren and Wellman (1958) estimated 7 percent on Skid Row and 6 percent in outpatient clinics or hospitals or in AA.  And, of course, by 1974, it was not uncommon to read that "3-5%" of the alcoholic population was on Skid Row, Alcohol and Health: New Knowledge, 1974, suggested "far less than 5 percent." (p. 113).

Though the origins of these estimates remain unclear, their function has been relatively straightforward: by attempting to distance the concept "alcoholic" from the longstanding Skid Row stereotype, alcoholism reformers tried to break down such denial of alcoholism and failure to recognize it that might flow from the old image. The hidden alcoholic, though often seeming to be little more than a shadowy extension of the disease concept's major rhetorical themes, thus became one of the central figures in the public health approach to alcoholism and one of the most colorful explanatory rubrics for why some alcoholics stayed untreated.

Abditivity Among Untreated Alcoholics

How much hiding is involved in remining an untreated alcoholic? Cahalan and Treiman's study was the first that this author is aware of in which general population respondents were questioned on the subject of the concealment of drinking.  The central item dealing with hiding read: "Some people sometimes use a variety of ways to hide or conceal some of their drinking. This can be done for a variety of reasons -- for example, one may not want to offend someone who has strong beliefs about drinking.  In the past year or so have you done any of the followng?"  Seven hiding maneuvers are described, and to each of which the respondent checked "yes" or "no."

The most frequent hiding tactic reported by untreated alcoholics (possibles and probables combined) and by Currently Drinking Nonalcoholics (CDNs) is the use of a breath "freshener" or mouthwash after drinking: such use was reported by one in three alcoholics and one in eight CDNS (see Table 10).  No other hiding behavior was reported by more than 3% of the CDN group; among untreated alcoholics the frequency of the remaining hiding behaviors ranged from 5% to 22% of the group.  On the whole, then., though the abditivity of untreated alcoholics considerably exceeds that of currently drinking nonalcoholics, most hiding tactics appear to be employed by a small minority of the persons qualifying in this sample as untreated alcoholics.

Used a "breath freshener" or mouthwash after drinking 33% 12%
Fibbed to a spouse or someone else about how much you had been drinking 22% 3%
"I sometimes sneak drinks when no one is looking" (response: "true now") 19% 1%
Stayed away from work or home or some other place on a made-up excuse in order not to show up tipsy or with alcohol on your breath 16% 1%
Specially drank drinks that don't leave alcohol on the breath (vodka, for example) 15% 3%
"Watered" bottles (added water to bottles of liquor or wine) so that the amount which had been drunk wouldn't show 5% *
Tried to keep people from different groups apart, so that the people in one group would not know about the drinking you did with the people in another group 5% *

As shown in Table 11, among untreated alcoholics 44% reported no hiding behaviors at all and 28% only one behavior -- thus 72% of the untreated group can be regarded as reporting only a limited repertoire of such tactics, and among most of these hiding was confined to the use of mouthwash.

Total Number of Hiding Behaviors: 0 1 2 3 4+
Untreated alcoholics (N=94) 44% 28% 14% 06% 09%
Currently drinking nonalcoholics (N=888) 84% 13% 02% 01% 00%
Postscript, 1999:

There, unfortunately, this old paper's narrative abruptly stopped.  Back in April of 1977, when the text was still fresh, I added a brief "note" sketching how my analysis of hiding would be concluded.  The need for such an I.O.U. was occasioned by a contract deadline -- something had to be turned in and so I submitted this nearly complete text without that final element as well as a concluding section.

Now, 22 years later, it is arguably a little too late to finish the job.


Anderson, D. (1942) "Alcohol and public opinion." Quarterly Journal of Studies on Alcohol 3:376-392.

Edwards, G. (1973) A community as case study: alcoholism treatment in antiquity and utopia.  Pp. 116-136 in: Proceedings of the Second Annual Alcoholism Conference of the National Institute on Alcohol Abuse and Alcoholism. Washington, D.C., U.S. Government Printing Office.

Baekeland P. and Lundwall L.K. (1977) Engaging the alcoholic in treatment and keeping him there. Pp. 161-234 in: B. Kissin and H. Begleiter (Eds.) The Biology of Alcoholism Vol. 5: Treatment and Rehabilitation of the Chronic Alcoholic. New York,, Plenum Press.

Bigus, 0. (1973) Becoming 'Alcoholic': A Study of Social Tranformation.  Berkeley, University of California, Ph.D. dissertation.

Blizard., P.J. (1969) "The public image and social rejection of the alcoholic in New Zealand." Quarterly Journal of Studies on Alcohol 30:686-700.

Brunner-Orne, M. (1963) "Comments on normal drinking in recovered alcohol addicts." Quarterly Journal of Studies on Alcohol 24:730-733.

[Canadian] Commission of Inquiry into the Nonmedical Use of Drugs (1972) Treatment. Ottawa, Information Canada.

Cahalan, D. and Treimen, B. (1976) Drinking Behavior, Attitudes and Problems in San Francisco.  University of California, Berkeley, Social Research Group, School of Public Ilealth.

Clancy, J. (1964) "Motivation conflicts of the alcohol addict." Quarterly Journal of Studies on Alcohol 25:511-519.

Gallup, G. (1955) Public Opinion News Service. Princeton, American Institute of Public Opinion.  January (news release).

Gerard D.L., Saenger, G. and Wile, R. (1962) "The abstinent alcoholic." Archives of General Psychiatry 6:83-95.

Gibbins, R.J. (1954) "Alcoholism in Ontario; A survey of an Ontario county." Quarterly Journal of Studies on Alcohol 15:47-62.

Glock, C.Y. (1964) "Images of man and public opinion." Public Opinion Quarterly 28:539-546.

Goffman, E. (1963) Stigma; Notes on the Management of Spoiled Identity. New York, Prentice-Hall.

Haberman, P.W. and Sheinberg, J. (1969) "Public attitudes toward alcoholism as an illness." American Journal of Public Health 59:1209-1216.

Harris, L. and Associates, Inc. (1971) American Attitudes Toxiard Alcohol and Alcoholics; A Survey of Public Opinion. Prepared for NIAAA. December.

Hartocollis,, P. (1968) "Denial of illness in alcoholism.'' Bulletin of the Menninger Clinic 32:47-53.

Henderson, R.M. and Bacon, S.D. (1953) "Problem drinking: The Yale plan for business and industry." Quarterly Journal of Studies on Alcohol 14:247-262.

Jellinek, E.M. (1952) "Phases of alcohol addiction." Quarterly Journal of Studies on Alcohol 13:673-684.

_____ (1960) The Disease Concept of Alcoholism. New Haven, Hillhouse Press.

Knupfer, G. (1967) Discussion of four papers in the alcoholism section of the Society for the Study of Social Problems, San Francisco, Ca., SSSP Annual Meeting, August 26-27.

Lemkau, P.V.  and Crocetti, O.M. (1962) "An urban population's opinion and knowledge about mental illness." American Journal of Psychiatry 118:692-700.

Linsky, A.S. (1970) "The chancing public views of alcoholism." Quarterly Journal of Studies on Alcohol 31:692-704.

_____ (1972) "Theories of behavior and the social control of alcoholism. Social Psychiatry 7:47-52.

Lloyd, R.W., Jr. and Salzburg, H.C. (1975) "Controlled social drinking; an alternative to abstinence as a treatment goal for some alcohol abusers." Psychological Bulletin 82:815-842.

Manis, J.G. and Hunt., C.L. (1957) "The community survey as a measure of the prevalence of alcoholism." Quarterly Journal of Alcohol 18:212-216.

Marcus, A.M. (1961) Popular Beliefs About Alcoholism: An Exploratory Investigation. Studies in Alcohol Education Project 1: First Report. Toronto, Addiction Research Foundation,

_____ (1962) Description and Discussion of Jellinek's Remarks on Public Attitudes Toward the Idea of Alcoholism as an Illness.
Studies in Alcohol Education. Research Memo 4.  Toronto, Addiction Research Foundation.

Maxwell, M.A. (1952) "Drinking behavior in the state of Washington." Quarterly Journal of Studies on Alcohol 13:219-239.

McCarthy, R.G. and Fain, T.G., (1959) "Measuring knowledge of alcoholism in the community." Quarterly Journal of Studies on Alcohol 20:645-654.

Moore, R.A. and Murphy, T.C. (1961) "Denial of alcoholism as an obstacle to recovery." Quarterly Journal of Studies on Alcohol 22:597-609.

Mulford., H.A. and Miller., D.E. (1961) "Public definition of the alcoholic." Quarterly Journal of Studies on Alcohol 22:312-320.

_____ (1964) "Measuring public acceptance of the alcoholic as a sick person." Quarterly Journal of Studies on Alcohol 25:314-323.

O'Hollaren, P. and Wellman, W. (1958) "'Hidden' alcoholics: medical implications of undiscovered addiction." California Medicine 89:129-131.

Orcutt., J.D. (1976) "Ideological variations in the structure of deviant types: A multivariate comparison of alcoholism and heroin addiction." Social Forces 55:419-437.

Paredes, A. (1974) "Denial., deceptive manuevers, and consistency in the behavior of alcoholics." Annals of the New York Academy of Sciences 233:23-33.

Pattison., E.M. (1966) "A critique of alcoholism treatment concepts: With special reference to abstinence." Quarterly Journal of Studies on Alcohol 27:49-71.

_____ (1967) "Abstinence criteria in alcoholism treatment." Addictions 14:1-19.

Philp, J.R. (1960) "Alcoholism is a public health problem." California Alcoholism Review 3:45-50. [Cited in Mulford and Miller, 1964]

Rappeport, M., Labaw, P. and Williams, J. (1975) The Public Evaluates the NIAAA Public Education Campaign, Vol. 1. A Study for U.S. DHEW. Princeton, N.J., Opinion Research Corporation.

Reinert, R.E. (1968) "The concept of alcoholism as a disease." Bulletin of the Menninger Clinic 32:21-25.

Riley, J.W. (1949) "The social implications of problem drinking." Social Forces 27:301-305.

Robins, L. (1966) Deviant Children Grown Up. Baltimore, Williams and Wilkins.

Roizen, R. (1974) "Naming names: a note on drinker self-characterizations." Drinking and Drug Practices Survevor No. 9:18-20.

Roman, P.M. and Trice, R.M. (1968) "The sick role, labelling theory, and.the deviant drinker." International Journal of Social Psychiatry 12:245-251.

Room, R. (1970) Assumptions and implications of disease concepts of alcoholism. Paper presented at the 29th International Congress on Alcoholism and Drug Dependence, Sydney, Australia, February 2-14.

_____ (1972) "Drinking and disease; comments on 'The Alcohologist's Addiction'." Quarterly Journal of Studies on Alcohol 33:1049-1059.

_____ (forthcoming) The accuracy of survey reports of drinking behavior and events. University of California, Berkeley, Social Research Group.

Roper, E. et al.. (1958) Public Sees Alcoholism as Illness. The Public Pulse. Chicago, National Newspaper Syndicate. January (news release)

Rubington, E. (1972) "The hidden alcoholic." Quarterly Journal of  Studies on Alcohol 33:667-683.

Ruprecht, A.L. (1970) "Alcoholism: Denial and the physician." Postgraduate Medicine 47:165-171.

Schmidt, W., Smart, R.G. and Moss, M. (1968) Social Class and the Treatment of Alcoholism; An Investigation of Social Class as a Determinant of Diagnoses, Prognosis and Therapy. Toronto, University of Toronto Press. (Brookside Monograph of the A.R.F., No. 7)

Smith, J.A. (1961) "Problems in the treatment of the alcoholic." Quarterly Journal of Studies on Alcohol Supplement No. 1:129-137, November.

Sobell, L. and Sobell, M. (1975) "Outpatient alcoholics give valid self-reports." Journal of Nervous and Mental Disease 161:32-41.

Star, S.A. (1955) The public's attitude about mental illness. Paper presented at the annual meeting of National Association for Mental Health, Indianapolis, Indiana.

U.S. Department of Health, Education and Welfare. (1974) National Institute on Alcohol Abuse and Alcoholism, Alcohol and Health: New Knowledge, Second Special Report to the U.S. Congress, from the Secretary of Health, Education and Welfare. Washington, D.C., U.S. Government Printing Office.

Wellman, W., Maxwell, M. and O'Hollaren, P. (1957) "Private hospital alcoholic patients and the changing conception of the 'typical' alcoholic." Quarterly Journal of Studies on Alcohol 18:388-404.

Wexberg, T.E. (1951) "Alcoholism as a sickness." Quarterly Journal of Studies on Alcohol 12:217-230.

Woodruff., R.A., Guze, S.B., and Clayton, P.J. (1973) "Alcoholics who see a psychiatrist compared with those who do not." Quarterly Journal of Studies on Alcohol 34:1162-1171.