In a projective test, the examiner presents unstructured, vague, or ambiguous stimuli (such as the inkblots of the Rorschach test) with the belief that responses to the test represent revelations about the unconscious mental processes of the respondent. As of the mid-1990s, ﬁve of the ﬁfteen most frequently used psychological tests were projective techniques. This is somewhat surprising, given that the psychoanalytic approach in which the tests are based has been out of favor in the mainstream of American psychology for more than ﬁfty years, yet the tests demand just such an approach in scoring and interpretation. The popularity of the tests is even more surprising given the lack of solid proof that they are even capable of providing any useful diagnostic information, along with a substantial body of evidence indicating that the tests lack reliability and validity—this is sometimes referred to as the “projective paradox.” Validity is the extent to which a test is actually measuring what it claims to measure, as well as its ability to predict behavior. Reliability simply refers to the extent to which a person taking the same test more than once will obtain the same results each time, as well as the extent to which the test will yield similar results regardless of who scores it.
Certainly the best-known projective test is the Rorschach test, introduced in the 1920s by the Swiss psychiatrist Hermann Rorschach, in which people are asked to describe what they see in a series of ten inkblots. It is far and away the most popular projective technique, even now given to many hundreds of thousands of people annually. It came under harsh attack as long ago as the 1950s, due to its lack of standardized procedures and norms—averaged results from a representative sample of the population, used as a reference point. Without them it is impossible to determine whether an individual’s results are “normal” or not. Standardization is important because apparently minor differences in how a test is given can strongly inﬂuence a person’s responses.
Since the 1970s Rorschach users have felt protected against such criticism by John Exner’s Comprehensive System, which provides detailed procedures for standardized administration of the test as well as norms for both children and adults. Unfortunately, the test continues to have major problems with reliability and validity, largely because of the continuing subjective nature of many of the scoring criteria. The person scoring the test rates the subject’s responses on more than 100 characteristics, including such things as whether the person described the whole blot or just parts, whether the response was typical or
The Rorschach test consists of symmetrical inkblots similar to the one shown here, which was produced by the author (the actual Rorschach stimuli may not be reproduced). Photo by author. unusual, whether the response was based on shape or color or both, whether the person focused on the dark portions or the white spaces, and many other details. As a result, two well-trained examiners may come up with strikingly different interpretations of a single person’s responses.
The result of this lack of reliability is a remarkable lack of validity. The Rorschach is quite poor as a diagnostic tool for most psychiatric conditions, with the possible exception of schizophrenia and other thought disturbances, and even then the evidence is mixed. Quite clearly false, however, are the claims by some Rorschach proponents that the method can reliably detect depression, anxiety disorders, sexual abuse in children, antisocial personality disorder, tendencies towards violence, impulsivity, and criminal behavior. Furthermore, the norms that exist for the test are unrepresentative of the U.S. population, and their use results in substantial overestimation of maladjustment. In one California study of blood donors, for example, one in six appeared to have schizophrenia, according to their Rorschach scores.
The test also is remarkably susceptible to faking, an important consideration for a test so frequently introduced as evidence in court. A 1980 study is typical: Rorschach responses of twenty-four people were submitted to a panel of experts for diagnosis. The proﬁles actually came from the following four groups: six actual mental patients with a diagnosis of paranoid schizophrenia, six “uninformed” fakers instructed to try to fake the responses of a paranoid schizophrenic, six “informed” fakers who listened to a detailed tape about schizophrenia ﬁrst, and six normal control subjects who simply took the test under standard conditions. Each test taker was rated by six to nine judges. The informed fakers were diagnosed as psychotic 72 percent of the time, versus only 48 percent for the actual psychotics. The uninformed fakers were also diagnosed as psychotic almost half of the time, and even the normal controls were diagnosed 24 percent of the time.
The reliability and validity of other projective tests also raise serious doubts. The Thematic Apperception Test (TAT), almost as widely used as the Rorschach, has neither standardized administration procedures nor an established scoring procedure. In the TAT, respondents are shown a series of ambiguous scenes drawn on large cards. For each picture, the respondent must make up a story. One card takes the projective approach to an extreme: it is totally blank. Individual clinicians choose the number of cards to show, up to thirty-one, as well as which particular cards are used. Although many standardized scoring systems have been created for the TAT, a survey of North American psychologists practicing in juvenile and family courts found that only 3 percent used any of them. Research suggests that using them would not help in any case. The systems show poor reliability and are unable to differentiate normal individuals from people who are either psychotic or depressed. Furthermore, these scoring systems provide no norms.
A third projective approach in wide use, again mostly by the courts, asks the person to draw a picture. The most widely used drawing test is the rather self-explanatory Draw-a-Person test. Interpretation proceeds in what has been called a “clinical-intuitive” manner, based on “signs” (features of the body or clothing, for example), usually guided by rather tentative psychodynamically based hypotheses. Large eyes might indicate paranoia and long ties might suggest sexual aggression, for example. A house with no windows might indicate feeling trapped. A person whose genitalia, or hands, or knees, or other features, depending on the interpretive guide used, are prominently visible might indicate a history of sexual abuse or latent homosexuality. There is no evidence, however, supporting the validity of this approach. Clinicians, in other words, have no grounds for believing any particular signs indicate any particular problem, other than their own prejudices and those of whoever trained them. Furthermore, studies suggest that clinicians will often attribute mental illness to many normal individuals who simply don’t draw very well.
At this point, it is clear that projective tests fail to meet even the most basic standards of reliability and validity. In fact, a recent review of projective tests commissioned by the American Psychological Society (APS) concluded “that, as usually administered, the Rorschach, TAT and human ﬁgure drawings are useful only in very limited circumstances” (author emphasis). Given this, how shall we interpret the projective paradox? Why are they still among the most popular tests? Of the various possible explanations, two seem especially important here. As human beings, clinicians are as susceptible as anyone to conﬁrmation bias, or the tendency to take into account evidence that supports one’s own beliefs and expectations while failing to consider evidence which fails to do so. A clinician who believes large eyes indicate paranoia, for example, will place great importance on the single client who drew large eyes and actually was paranoid, while remaining unmoved by (and possibly not even having noticed) the many large-eyed pictures drawn by clients who were not paranoid.
A second, more positive possibility to consider is the fact that many clinicians who use projective methods do not use them as tests or diagnostic tools at all, but rather as auxiliary tools in clinical interviews. They help the clinician to form initial, tentative hypotheses about the client, to be tested by closer examination with better tools. This use of the tests seems more appropriate, given the apparent uselessness of projectives where diagnosis is concerned. Unfortunately, surveys suggest that many clinicians, despite the clear evidence to the contrary, continue to believe in the diagnostic efﬁcacy of projective tests.
- Gregory, R. J. Psychological Testing: History, Principles, and Applications. 3rd ed. Boston: Allyn & Bacon, 2000;
- Lilienfeld, S. O., Wood, J. M., and Garb, H. N. “The Scientiﬁc Status of Projective Techniques.” Psychological Science in the Public Interest, 1(2) (2000).