That is, the symptom “is worried about how manly he is” occurred equally often with drawings that were muscular and manly and with drawings that weren’t.
After having examined the forty-five drawings and the pair of “symptoms” associated with each, subjects were asked to indicate which characteristics of the drawings went with each of the symptoms. In fact, there were no such relationships. Nonetheless, the subjects’ responses showed that they had perceived in the drawings relationships between drawing characteristics and symptoms that were not really there. Thus, they reported that men who worried about their masculinity overwhelmingly draw figures with broad shoulders. Chapman and Chapman (1969) later showed that trained psychodiagnosticians showed similar illusory correlation effects. Additional research has shown that even explicit warnings about the illusory correlation effect fail to prevent it (Wailer and Keeley 1978; Kurtz and Garfield 1978). As Kurtz and Garfield (1978) note, individuals, even trained clinicians “try to find something in the clinical material presented” (p. 1013). Even if the meaning isn’t really there, it will be found.
Chapman and Chapman (1967; Chapman 1967) have further found that the degree of the illusory correlation effect is due to what is termed the associative strength between variables that are perceived, incorrectly, to be correlated. Thus, the head and intelligence are logically associated. So are broad shoulders and masculinity. Based on these associations, people assume that a broad-shouldered drawing means that the individual worries about his masculinity and that people worried about their intelligence draw large-headed figures. Chapman showed the importance of associative strength in the genesis of illusory correlation in a study that used pairs of words. Subjects saw sets of word pairs, one pair at a time. Some of the pairs contained associated words (e.g., “lion tiger”) and some pairs contained unassociated words (e.g., “lion notebook”). Although each word in the list of pairs that the subjects saw appeared equally often with each other word, subjects reported that words were paired with a strongly associated word much more frequently than was actually the case. Returning to the clinical situation with projective tests, these results show that the associative strength variable produces the illusory correlation. Selective memory is probably responsible for the continued acceptance of the correlations. When one finds an actual instance in which a man worried about his masculinity has drawn a broad-shouldered picture, that instance will stand out in memory much more than one in which a man worried about his masculinity doesn’t draw a broad- shouldered picture. This will be recognized as the same process that maintains belief in psychic predictions, hunches, and prophetic dreams.
Illusory correlations, license to make symbolic interpretations, and selective memory all combine to produce a strong belief that unreliable and invalid projective tests are accurate ways of discovering psychologically relevant information about people.
The failure of projective tests to achieve even minimal standards of adequacy has led Grove et al. (2002) to conclude that the results of these tests should not be admitted in the courtroom as they do not meet the legal standards of scientific evidence. In fact, Grove and Barden (1999) have argued that testimony of mental health experts (clinical psychologists, social workers, and other types of mental health therapists) in general is based so much on nonscientific methods that no such testimony meets the legal standards for admissibility.
Does psychoanalytic therapy work? The question of whether any type of psychotherapy works or not is much more difficult to investigate than might be expected. One of the problems lies in the definition of “work.” If by “work” one means only that the therapy is better than doing nothing at all, then nearly all types of psychotherapies “work.” However, there is a very large placebo effect in psychotherapy. Simply believing that one is being treated can result in beneficial effects. A more rigorous definition of “work,” then, requires that the therapy in question be more effective than a placebo therapy in which patients engage in some activity they believe is therapeutic, rather than receiving the actual type of psychotherapy being evaluated. Another problem in evaluating psychotherapeutic effectiveness is spontaneous remission. Many of the problems about which people consult psychotherapists will simply go away if left untreated. If an individual is seeing a therapist during the period of time when the problem would have disappeared on its own, it is quite natural for both the therapist and the patient to attribute the elimination of the problem to the therapy, not to the passage of time. Such processes result in strong beliefs on the part of many psychotherapists, psychoanalytic as well as nonpsychoanalytic, that their particular brand of therapy works. These same processes also produce scores of testimonials for any type of psychotherapy one wishes to name.
The problems discussed above all operate to make it appear that various types of psychotherapy are more effective than they really are. Actual research on the effectiveness of psychotherapies in general has shown that, as expected, so-called talk therapies are more effective than doing nothing (Smith, Glass, and Miller 1980). However, such psychotherapies show almost no greater effectiveness than placebo therapies (Prioleau, Murdock, and Brody 1983). Although these conclusions may at first seem surprising, they become more understandable after a bit of “demythologizing” of psychotherapy is done. Most people consult psychotherapists not because they suffer from major psychoses—such as schizophrenia or manic-depressive psychosis—but because of smaller neurotic problems or just because they are upset by the “slings and arrows of outrageous fortune” that beset everyone from time to time. Gross (1978), in an excellent and much overlooked book, has cogently argued that we are becoming a “psychological society” in which we are led to define every little disappointment, setback, and depression as something to be treated by some mental health professional. Albert Ellis, founder of rational emotive therapy, has pungently commented on this trend: “I find that increasingly in our society much of what we call emotional disturbance is whining” (quoted in Gross 1978, p. 315, emphasis in original). The type of problems for which many people now consult a psychotherapist, then, are frequently minor in terms of true psychopathology. In past years, such problems would have been talked over with a trusted friend, relative, or clergyman—people who would have had no psychotherapeutic training. It is certainly true that talking about problems with someone else can be very beneficial. The other person may be able to propose solutions, may see the problem from a different perspective, or may simply provide moral support for a difficult and troubling course of action already decided upon. These are real and important benefits of what is generally called “advice.” Psychotherapists, no matter what particular type of therapy they practice, provide advice to their patients, in addition to the special ministrations their own brand of therapy calls for. It is this advice component of every psychotherapy that helps account for the general effectiveness of psychotherapy over doing nothing.
If psychotherapies in general are more effective than doing nothing because the therapist is providing advice to patients, one would expect that professional advice givers (i.e., trained professional psychotherapists) would not be much more effective than individuals lacking formal therapeutic training. This issue has been the focus of considerable research. Two reviews of the literature on this issue (Durlak 1979; Hattie, Sharpley, and Rogers 1984) have actually found that patients treated by trained professionals do worse than those treated by untrained “paraprofessionals.” Berman and Norton (1985) have criticized the Hattie, Sharpley, and Rogers review on statistical grounds and reanalyzed the literature that the original study reviewed. In their reanalysis Berman and Norton find no difference in therapeutic effectiveness between trained professionals and untrained paraprofessionals. Dawes (1994) has penned an especially trenchant critique of the practice of psychotherapy that exposes the many empirical and logical failings of claims that underlie many psychotherapeutic beliefs and techniques.
Returning to the specific issue of the effectiveness of psychoanalytic therapies, what is true for psychotherapies in general is true for psychoanalytic therapies in particular. Repeated reviews of the literature have failed to show any solid evidence that psychoanalytic therapy is superior to placebo therapy (Eysenck 1952; Rachman 1971; Rachman and Wilson 1980; Erwin 1980, 1986).
Does any type of psychotherapy provide a better result than placebo therapy? The answer is yes, and the type of therapy is behavior therapy and its close relative, cognitive behavior therapy (Erwin 1986). Developed as an alternative to the ineffective psychoanalytic treatments in the early 1960s, early behavior therapies concentrated on classical and instrumental conditioning as the explanation of disordered behavior. The idea was that such behavior was learned and could be eliminated using the techniques of reinforcement, punishment, and extinction drawn from work on conditioning animals. As the field of behavioral therapies has matured, it has become much more cognitive, admitting that patients’ cognitions play an important role in disordered behavior and must be addressed by any
