A Simplistic Understanding of the Five-Factor Model

To the Editor: Jonathan Shedler, Ph.D., and Drew Westen, Ph.D. (1), argued that the five-factor model “omits key clinical constructs and may not capture the complexity of personality syndromes seen in clinical practice” (p. 1743). To demonstrate this, they constructed an abbreviated measure of the five-factor model using a small subset (30%) of the 200 items contained within the Shedler-Weston Assessment Procedure (SWAP-200). It was a foregone conclusion that the results of a factor analysis of 30% of the SWAP-200 items would not correspond to a factor analysis of the entire SWAP-200. In addition, one should ask whether their SWAP-200 items provided an adequate assessment of the five-factor model. They never attempted to validate their five-factor model measure, and a visual inspection of the items indicates inaccurate representation of the five-factor model. Finally, it is highly unlikely that their small set of SWAP-200 items would provide anything close to a reasonably comprehensive assessment of the five-factor model. In sum, there was little reason to expect that an incomplete and inadequate assessment of the five-factor model with a subset of the SWAP-200 would account for the variance within the entire SWAP-200.

Drs. Shedler and Westen suggested that the constructs assessed by the 12 SWAP-200 scales are outside of the realm of the five-factor model. They ignored many, many studies that indicated otherwise. For example, two other compelling dimensional models of personality disorder have been developed by Dr. Livesley (the Dimensional Assessment of Personality Pathology) and Dr. Clark (the Schedule for Nonadaptive and Adaptive Personality). A complete description of and references for these measures, as well as others, is provided by Widiger and Simonsen (2). The clinical constructs assessed by these scales include narcissism, identity problems, eccentric perceptions, affective lability, aggression, detachment, self-harm behaviors, and compulsivity that resemble closely the scales of the SWAP-200, and many studies have documented well that the constructs assessed by the Dimensional Assessment of Personality Pathology and the Schedule for Nonadaptive and Adaptive Personality are well within the realm of the five-factor model.

Drs. Shedler and Westen (1) derogatorily characterized the five-factor model as providing a simplistic lay description of personality. What was simplistic was their characterization of the five-factor model. The five-factor model is a rich dimensional model of general personality structure that has been used successfully in many areas of science and practice, including (but not limited to) the study of temperament, gender differences, heritability, behavioral medicine, and aging. An integration of the DSM-IV personality disorder nomenclature with the five-factor model would go far in integrating DSM-IV with basic science research on personality structure (3). We regret that Drs. Shedler and Westen (1) argued instead for a distinct separation of our clinical understanding of personality disorders and basic science research on personality structure.