As mentioned in a prior post, this is a piece of a seven part series examining science and pseudoscience in the field of personality and psychopathology assessment. My goal is to expose the less than stellar underbelly of some of the most commonly-used and well known measures, particularly those that fall under the grouping of “projective” assessments. The first measure I tackled was the Rorschach, and it was definitely found wanting. This post deals with another venerable test, the Thematic Apperception Test.
The Thematic Apperception Test (TAT; Murray, 1943) has a history almost as long as the
Rorschach, and, also like the Rorschach, it has a highly interesting development and history (see
Cramer, 2004 for full details). The two major figures in the development of the TAT were Henry Murray and Christina Morgan. Murray was a surgically trained physician with a PhD in biochemistry before being hired on faculty of the Harvard Psychological Clinic in 1926. Although initially largely unqualified for such a position, Murray underwent extensive training in psychoanalysis, including meeting with Carl Jung, and intensive reading in psychiatric and psychological literature (Robinson, 1992). Morgan was an artist and certified nurse’s aide who was highly influenced by Jung’s theories on personality and psychopathology, having been analyzed by him personally (Douglas, 1993). Although unremarked on in many writings on the TAT (e.g., Groth?Marnat, 2003; Weiner & Greene, 2008), Murray and Morgan also engaged in a long?lasting extramarital affair that ended only with her suicide in 1967 (Douglas, 1993). Murray and Morgan may appear odd choices to develop a major psychological test, but the TAT ranks second only to the Rorschach as the most often used type of projective test by clinical psychologists (Camara et al., 2000). It is less popular, but still quite frequently used by school psychologists (Hojnoski et al., 2006).
Murray appears to have been the theoretical driving force behind the TAT, as it is based on his needs?press concepts of personality. For Murray, an individual’s personality is the result of an interaction between one’s needs (internal motivations) and presses (environmental or situational pressures that impact how one expresses those needs). Morgan, who is absent as an author from the officially published version of the test (Murray, 1943), assisted more in the preparation of the actual testing materials (the pictures on the test cards), some early administration of the measure, and writing the results for publication (Holt, 1949). The instrument itself (in the final version) consists of 31 black and white cards that have pictures of various kinds (14 show a single person, 11 show two people, three have a group of people, two have scenes of nature, and one is blank), although only 20 are used with each individual, since some cards are specific to age or gender. Examiners show the cards to the examinee and ask him or her to tell a story based on the picture. The stories that are told, according to Murray, reveal numerous aspects of personality and can be used to understand how someone thinks and feels in real?world influenced by Jung’s theories, Murray believed that these stimuli would also “expose the underlying tendencies which the subject…is not willing to admit, or cannot admit because he is unconscious of them” (p. 1, 1943).
The TAT manuals (Murray, 1943; 1971) provide very clear and detailed procedures for assessing 28 “needs” and 24 “presses” along a 5?point scale based on the stories told. However, similar to what happened with the Rorschach, numerous other systems and methods of using the TAT soon developed. Methods using a smaller number of cards than the standard 20 (often 8?12) became common (Karon, 2000), and a majority of practitioners do not appear to use any of the available scoring systems, instead relying on “intuitive” interpretations of the stories (Gieser & Stein, 1999; Groth?Marnat, 2003; Rossini & Moretti, 1997). In fact, surveys show that few users even follow Murray’s guidelines to present the cards across two different sessions (Vane, 1981) or even present the same cards in published research (Keiser & Prather, 1990). So, just as with the Rorschach, many users of the TAT, both historically and currently, are not using it as originally intended by the developers or even from the same theoretical viewpoint as them (e.g., Westen et al.’s  psychodynamic, object?relations focused scoring system).
There have been several positive findings regarding scoring on the TAT and relationship to specific areas of psychological functioning, but they have all been found when using a particular scoring system. For example, one meta?analysis found TAT scores under McClelland’s system were superior to self?report scores in predicting long?term career outcomes, such as success in a one’s career and level of income (Spangler, 1992). Westen’s scoring system has been found to differentiate between those with and without personality disorders (Ackerman et al., 1999). Other scoring systems have shown TAT scores to be related to therapy attendance in persons with personality disorders (Ackerman et al., 2000) and general symptom improvement in persons treated in an in?patient unit (Fowler et al., 2004). There have not, however, been any studies that have successfully used the TAT (in any of its scoring variations) to accurately assess the majority of mental disorders.
Unfortunately, as shown above and noted by many (e.g., Groth?Marnat, 2003; Hunsley, Lee, & Wood, 2003), the majority of those using the TAT clinically would not benefit from this information, since most practitioners are not using either standardized administration or scoring procedures. Add in the TAT’s lack of incremental validity (Garb, 1998), the high potential for overpathologizing normal populations based on TAT responses (Lilienfeld et al., 2000), and it can be seen why the TAT “rarely plays a prominent role in clinical diagnostic evaluations” (Weiner & Greene, 2008, p. 469). Indeed, the purpose of the TAT, as originally conceived, was not for it to be used as a diagnostic instrument, but instead as a method of exploring a person’s experience of the world around them and the underlying motives they attributed to others. It was not intended to assess for manifested symptoms seen in a psychological disorder.
So, in summary, the TAT has some limited empirical support in assessing for personality disorders and achievement motives when using particular scoring systems. In none of these systems, however, has evidence shown it to be a useful tool to measure cognitive, emotional, or behavioral symptoms of any mental disorder outside of that limited purview. Further, given that few practitioners use the TAT in the standardized manner that it was intended to be used, those that do use it in a diagnostic fashion are undoubtedly relying on personal experience and judgment, rather than empiricism and sound research, with all the attendant biases and problems relying purely on personal experience entails (Dawes, Faust, & Meehl, 1989). In short, diagnoses based on results from a TAT should be treated with extreme caution.
Verdict – Pseudoscience when used outside of a small domain (personality disorders)
Next time, I will examine projective drawings. This brings us to the halfway point of our series, and from there we will examine sentence completion tasks and the (non-projective but very widely used) Myers-Briggs Type Indicator. I will conclude the series with a look at scientifically reliable and valid measures of personality.
(For a full list of the works I’ve cited above, feel free to email me)