Leake G.J., King A.S.
Alcohol Health & Research World: 1977, 1, p. 16–22.
Could negativity about the recovery chances of disadvantaged drinkers become self-fulfilling prophecies? This hidden gem study from the late ’70s suggests they can, and that bolstering counsellors’ expectations promotes commitment to treatment and recovery among the least promising of clients.
Summary Published in 1977 by the US government’s National Institute on Alcohol Abuse and Alcoholism, the featured article was written as much in the style of a magazine as an academic journal, lacking some of the information and methodological safeguards expected of modern-day research papers. On a few points where we felt confident of the meaning, we have made this clearer than in the original. The article is not easy to obtain and copyright restrictions prevent us making it routinely available. However, we have generated an email-able PDF from the poor-quality paper copy, and can send this to individuals for their own private study: use this template email to request a copy.
Sociological and psychological literature as well as medical, job-training, and educational studies suggest that when people interact, one person’s expectations of how another will behave can become self-fulfilling. In particular, stereotypes and attitudes regarding the ‘low motivation and apathy’ of underprivileged people can act as self-fulfilling prophecies of poor status and performance. One possible instance is investigated in this study: that counsellors’ expectations may significantly influence the adjustment and recovery of disadvantaged ‘skid-row alcoholics’. The hypothesis is that rehabilitation, detoxification, and affiliated organisations can in effect ‘cue’ or ‘condition’ these patients to ‘fail’ by anticipating poor progress. If so, the converse should also to be true: that disadvantaged drinkers can be motivated to recover appreciably better when their counsellors hold more favourable expectations. For ethical reasons, this was the proposition tested in this study.
The study posited that low therapist expectations of the recovery chances of disadvantaged clients may become self-fulfilling prophecies.
To test this, at the beginning of patients’ detoxification treatments counsellors at three US alcohol rehabilitation programmes were falsely led to believe that certain clients could be expected to show “remarkable recovery”.
These clients actually did consistently show signs of making greater progress during their first year of counselling.
Practice recommendations seek to balance engendering optimism among clients with not unrealistically expecting too much.
To test this proposition, at the beginning of detoxification treatments counsellors at three government-sponsored US alcohol rehabilitation unts were falsely led to believe that certain clients could be expected to show “remarkable recovery” during the course of counselling. Specifically, they were told that the study was seeking further validation of a “specially designed psychological personality test for the hard-core alcoholic” designed to identify those with “high alcohol recovery potential”, and that this test would be administered at the beginning of detoxification episodes. Though it was stressed that all their clients had “satisfactory personality profiles” and should progress in the programme, counsellors were told that those with high recovery potential were highly motivated to accept counselling and could as a result be expected to attain exceptional recoveries. In fact, the test was merely a standard personality/temperament test, and clients designated as having high recovery potential had been chosen at random.
At the three units, respectively 12, 17, and 22 clients (51 in total) with similar drinking histories and socioeconomic backgrounds joined the study as they started detoxification programmes followed by long-term counselling. From among these a randomly selected three, four and five respectively were designated as “high alcohol recovery persons”, while the remainder formed the control groups against whom the progress of these supposedly most promising clients would be benchmarked. Between these two sets of clients there were no significant differences in age, drinking habits, socioeconomic backgrounds, or on the personality test. All had lost their jobs because of drinking problems and were unemployed, had police records, and been members of Alcoholics Anonymous. Any differences between ‘high recovery potential’ clients and the remainder in their progress towards recovery could reasonably be attributed to expectations existing only in the minds of their counsellors.
At all three units there was evidence that the study’s attempts to induce expectancies in the counsellors generated real advantages for the supposedly high recovery potential clients.
On each of eight dimensions assessing response to therapy, counsellors rated the 38 out of the 51 clients who completed the counselling schedule for a year (all 13 missing were among the 39 controls) from “has improved” through “no change” to “gone back”. In each unit and on every dimension there were more favourable ratings for the supposedly most promising clients than for the controls. They were seen as being more motivated to accept counselling, more punctual in attending appointments, presenting a neater or more attractive appearance, more able to exert self-control, more ambitious [presumably for their recovery], more cooperative, trying harder to stay sober, and generally showing the best recovery. Amalgamated across these dimensions, in each unit the difference in counsellor ratings was statistically significant.
Such ratings might not reflect the reality of the clients’ progress, but merely the biased perceptions induced by the study’s manipulation of the counsellors’ expectations. However, not so easily explained away are the remaining differences in evidence of the clients’ progress. Near the end of the study the same number of clients rated by counsellors (paragraph above) were rated also by their fellow clients, who encountered them in group discussions. They were asked to identify who they would most like to talk with, be with, and who had the best overall recovery. Despite not having been told these clients were more promising, in each unit and on each dimension the clients also rated the supposedly most promising of their number most highly.
Clients also estimated for each of their peers the number of days they had been sober [presumably over a year]. Averaged across all three units, what otherwise was an estimated 33 days sober more than doubled to 74 days among clients counsellors had been told had the makings of a “remarkable recovery”. At each of the three units the differences were statistically significant. These estimates were available for all the supposedly most promising clients and 29 of the 39 controls.
Particularly significant given the units’ emphasis on completing their programmes were records of absences from scheduled sessions and ‘drop-outs’, records available for all 51 clients. In each of the three units, clients spotlighted to counsellors as particularly promising were to a statistically significant degree less likely to miss sessions or leave treatment early. Averaged across all three units, their absences were 57% fewer and none dropped out compared to an average of a third of those not spotlighted.
Another way the clients came to differ was the greater likelihood that those spotlighted as particularly promising would find and (indicated by fewer jobs transitioned through) keep jobs, differences which exceeded or were near statistical significance. Additionally, though they had not been told they were particularly promising, these clients somehow came to see themselves as such, self-assessing as having fewer ‘slips’ and greater endurance [in their recovery] than other clients. Towards the end of the study all available clients completed a written “Test Yourself – Rate the Program” exercise which asked them to be honest and frank. At all three units clients spotlighted as particularly promising averaged significantly higher on recovery orientation and saw their programmes as more beneficial than the other clients, differences which were statistically significant in five of the six cases.
An amalgamation of scores on several dimensions confirmed the overall advantage given to spotlighted clients. Ratings given them by other clients, attendance records, employment, estimated periods of sobriety, and responses to written exercises were ranked and then averaged to give each client a single rank. In each unit those from whom counsellors were led to expect exceptional recovery recorded an average rank significantly higher than other clients.
The featured study also provided some clues to how the false information provided to counsellors might have been communicated to and/or affected clients, perhaps in simple [or obvious] ways, but possibly too via cues and signs emanating from the counsellors so subtle that clients were unaware of them. Post-study interviews with clients included a test which revealed a possible example. It was based on studies which showed that pupil size communicates attitudes and expectations, ranging from highly dilated pupils communicating favourable attitudes towards another person, to extreme contraction (‘pinpoints of hate’) communicating unfavourable feelings.
After the trial had been completed clients were shown a pair of photographs of their counsellors which were identical, except that one had been modified to substantially enlarge the pupils. Then they were asked whether they “saw any differences in these pictures of your counselor” and, regardless of their answers, to “select the photo that shows how you usually see the counselor looking at you”.
Some clients did notice a difference between the photos, but even they could not say what it was. Nine of the 12 clients who had been designated as particularly promising (and who had as a group shown the greatest progress) chose the photo with enlarged pupils as representing how their counsellor usually looks at them, while most (l6) of the clients not spotlighted as having an exceptional prognosis chose the other photo. Eye contact in face-to-face relations is likely to serve as an unintentional but nevertheless striking indicator of the attitude, interest, and expectations counsellors hold for clients. Although clients were not aware of the subtleties involved, these and other complex and unnoticed cues operating in interpersonal clinical relations may have come to shape their attitudes, motivations, and progress.
More clearly identifiable processes may also have been at work. For example, counsellors were rapidly informed of local job opportunities and, not too surprisingly, the supposedly high-potential clients were first to be referred to these opportunities. With more opportunity for meaningful employment [and actually more and more stable employment], these clients had every reason to feel a greater involvement in and responsibility for their recovery.
[Suggesting that their relationships with clients had been affected by the study’s manipulations,] post-trial interviews revealed that counsellors could better recall the names of those designated a year earlier as particularly promising than the names of other clients.
Incidentally, the study examined whether the personality/temperament test intended as a bogus indicator of prognosis actually did relate to signs of how well the clients were progressing. There was no reliable evidence of such relationships.
The results strongly suggest that among these disadvantaged clients, those from whom counsellors were led to expect exceptional progress actually did progress better than other similar clients. In the form of premature termination of treatment and other variables, the remaining clients experienced real disadvantages by not having been similarly spotlighted. It appears that more favourable expectations are needed if disadvantaged alcoholics are to become quickly integrated and adjusted into mainstream society. Disadvantaged persons should not be assessed against social or health standards they are not expected to fulfil.
Observations on employment opportunities show that expectations aroused in counsellors affected clients through clear and obvious means as well as via subtle psychological cues such as pupil size. At all three units, the possibility that counsellors gave more attention and preferential treatment to those earmarked for exceptional recovery could not be ruled out.
More favourable attention from counsellors may have led to other indirect and subtle recovery incentives for the spotlighted clients, for example, via the reactions of their fellow clients. Clients’ evaluations of each other were initially uncontaminated by the attitudes and beliefs of counsellors, and clients were unaware of the test results which supposedly led to predictions of exceptional recovery potential, yet nevertheless they came to hold more favourable opinions of those designated as high potential. Conceivably, counsellors’ expectations of these clients’ progress were communicated to clients as a whole during weekly group discussion sessions, cuing others to hold corresponding expectations for their recovery. Such collective expectations could serve to augment the ‘high potential’ clients’ expectations of recovery and reinforce their motivations to ‘work’ their programmes.
The results suggest that to motivate disadvantaged clients to seek greater recovery benefits from counselling, counsellors must expect them to be capable of more than hitherto appreciated. The disadvantaged can be convinced not only by seeing fellow ‘alcoholics’ like themselves gaining sobriety, but also by feeling that this is to be expected of them. Trying to locate explanations of low motivation to engage with treatment within disadvantaged alcoholics themselves is of limited value. Counsellors might have been seeking to find in their clients what [ie, low expectations and negativity] they should have been looking for in themselves. To counter the tendency to under-expect from these sorts of clients, counsellors should be made aware of the effect of their expectancies and, if possible, trained in developing and communicating high expectations for clients’ potential for recovery.
Hopefully this description of counsellors’ expectations as determinants of the behaviour of disadvantaged people with alcohol problems will serve as a reminder that human progress and the potential for recovery from these problems are the combined outcomes of external social, as well as internal psychological, factors. Failure to recognise the social basis of differences between human beings in their motivation and progress is to limit understanding to half of reality.
commentary The featured study’s importance was apparent in a review published in 2014 in the Addiction journal, authored by William Miller – originator of motivational interviewing – and colleague Theresa Moyers. Their aim was to assess the relative importance of features specific to a therapy versus factors shared by (or ‘common’ to) bona fide approaches. The family of common factors they highlighted were “Therapist characteristics,” among which was their “Expectancy” in relation to the client. Under this subheading, the featured study provided the only evidence. A related factor also explored was the therapist’s “Allegiance” to an approach (their belief that it will work), which itself may work partly by instilling an expectation in the therapist that the client will improve.
How what the therapist expects could appreciably determine how well the client actually does became clear when the reviewers turned to client factors related to good outcomes. These included “initial optimism about treatment effectiveness, motivation, self-efficacy and hope”. They explained that “In addition to the importance that clients attach to change, their confidence in doing so is a good predictor of outcome. In health behavior models, importance and confidence flow together to predict change. Clients who believe that they can do what is needed to make and sustain a change have a distinct advantage as they begin treatment.” It is easy to envision how these client factors would be affected by the expectations of their therapist – occupying a position of authority on treatment – about how well they will do.
As interpreted by the authors and others, the implications of this unique study are profound both for practitioner training/supervision and for treatment practice. However, this small trial with just a dozen clients falsely spotlighted to therapists as particularly promising is a slim basis for asserting that therapist expectations actually do affect addiction treatment outcomes. In its favour is the plausibility of the processes it assumed would affect outcomes and the consistency of the findings suggesting they actually did. However, substance use and substance-related problems were not assessed and neither were longer term post-therapy outcomes, leaving question marks over whether the observed processes produced the desired outcomes. In particular, ratings by fellow clients similar to those on which considerable store was placed by the featured study have been found no better than a chance-level predictor of post-treatment outcome. Among methodological gaps are that modern studies are expected to adjust results for clients missing from follow-up data. Presumably because no attempt was made to contact clients who left treatment, these formed a substantial proportion of the comparison clients – for example, over a third when it came to ratings of clients by their therapists. However, if their data could have been included, it is likely that these ‘drop-outs’ would have further weighted the results in favour of supposedly exceptional-prognosis clients, all of whom were retained in treatment and in the study.
The proposed causal chain linking false information to counsellors at one end to client outcomes at the other is plausible, but there was no direct evidence that counsellor expectations of their clients had actually been affected by the information, leaving a major gap right at the start of the chain. Another link would run from the client’s expectations of how well they will do in treatment to how well they actually do. Across psychotherapy this link is statistically significant and, relative to other possible influences, appreciable, though in respect of substance use clients the evidence is weak. By implication the reviewers who found this result agreed with the featured study’s recommendations that therapists be trained in how to hold and communicate high expectations for clients. Specifically they suggest therapists deploy “persuasion tactics” with clients regarding the likely efficacy of psychotherapy, especially when explaining the basis for the treatment, for example, by mentioning that the approach is prestigious and supported by research, and offering vignettes of successful cases.
Countering ‘therapeutic nihilism’ regarding the prospects of dependent drinkers and the associated feeling that one is unlikely to be able to help, have long been seen as key components of training, though training of any kind has less chance of an impact if work environment and priorities mitigate against working with problem substance use (1 2). There is evidence that a US course implicit in which was “increasing participants’ optimism about intervention” did actually augment optimism among social work trainees, generating greater willingness to seek and undertake work related to substance use.
The study itself and later documents citing it suggest it means therapists should develop high expectations of all clients, but there is a fundamental logical weakness to drawing this conclusion. It could be that the act of distinguishing the dozen ‘high potential’ clients from the majority – communicated somehow by counsellors to clients – was an essential ingredient in promoting their progress. If therapists are trained to have high expectations of all clients, that ingredient would no longer be present. The featured study suggests just such a possibility by saying that at all three units, “the possibility that counselors gave closer attention and preferential treatment to those earmarked for better recovery could not be ruled out,” and stating that this did indeed happen in respect of employment. It is not possible to give “closer attention and preferential treatment” on the basis of expectations if those expectations are uniformly high for all clients. For example, in that scenario all would be referred equally to employment opportunities, eliminating the competitive advantage given to the dozen ‘high potential’ clients.
Another knotty practice dilemma arises from warnings about ‘over-egging’ a client’s chances of success, placing a question mark against the degree to which therapists should boost client expectations. With relapse the norm in substance use treatment, these warnings are especially relevant. It seems highly likely that across a caseload, instilling optimism is usually on balance positive, but perhaps not if it leads to greater disillusion and distrust if treatment fails. The review of expectation in psychotherapy cited above cautioned that therapists should temper their hope-inspiring statements so they neither too quickly threaten a patient’s beliefs or sense of self [ie, their own estimations of how well placed they are to make progress], nor promise unrealistic change – a caution reminiscent of the featured study’s warning that “Disadvantaged persons cannot be assessed against social health standards that they are not expected to fulfill.” This warning broadens out the influences on treatment outcomes beyond client and therapist to society’s possibly unrealistic expectations that clients labouring under multiple disadvantage will through treatment become sober and conventionally productive members of society. It seems to counter the risk of expecting too little with the risk of setting too high a bar for success, or at least too high in unrealistic directions. However, these mandates are not necessarily contradictory. Setting the recovery bar higher than is realistic is likely to engender expectations of client failure among staff, negativity which may be communicated to clients. Setting the bar lower may be expecting less, but could also engender optimism that within these parameters even multiply disadvantaged patients can succeed.
These considerations remind us that the study and the understandings on which it was based specifically concerned disadvantaged ‘skid-row alcoholics’, of whom therapists of the time often seem to have expected little in terms of engagement with and profiting from treatment. Another early US study published in 1970 addressed the same strata of the drinking population who turned up at a Massachusetts emergency department, but ‘failed’ to follow through on referral to treatment for the drink problems possibly underlying their ill-health. Dismissed by staff meant to help them as no-hopers uninterested in helping themselves, their engagement with treatment was elevated to normal levels by replacing staff’s dismissive attitudes with more positive attitudes and more assertive, practical and wide-ranging assistance. “In our view, these people do seek to be helped, and to perceive them as untreatable is a disservice as much to ourselves as to them,” said the treatment-unit manager who orchestrated the interventions and the study.
We owe a debt to William Miller of the University of New Mexico for consistently bringing this study to light.
Last revised 07 October 2021. First uploaded 29 September 2021
REVIEW 2018 Cohesion in group therapy: a meta-analysis
REVIEW 2018 Congruence/genuineness: a meta-analysis
DOCUMENT 2013 Sometimes best to break the rules