Reconceptualizing Personality Disorders within the Context of the Five-Factor Model
There remains debate as to exactly how personality disorders should be classified (see the Appendix). This issue is more than just a matter of curiosity, since our entire conception of personality disorders is an essential factor in how we approach their treatment. As perhaps the most widely accepted and scientifically validated trait perspective on normal personality, it stands to reason that the Five-Factor Model (FFM) ought to also provide a basis for classifying abnormal personality and, in particular, the personality disorders. Thus, Costa and Widiger (1994) brought together a group of experts, including Theodore Millon, to address personality disorders within the context of the FFM.
The DSM-III and DSM-IV use diagnostic categories for the personality disorders, whereas the FFM suggests a dimensional approach. The categorical approach has several advantages. It is relatively easier to conceptualize disorders as either having them or not, clinicians are familiar with the current categories, and when clinical decisions are categorical they tend to be consistent. However, the dimensional approach offers the advantages of not being arbitrary in defining specific symptom cut-off points for a diagnosis, they allow for retaining information on those patients who just miss the cut-off point for a diagnosis (and could, therefore, simply be classified as not having the disorder), and the dimensions are more flexible than a categorical diagnosis. More importantly than just speculating on advantages and disadvantages, however, is that the majority of empirical data seems to support the dimensional approach (Widiger & Frances, 1994). For example, borderline personality disorder patients do not show a specific profile on the MMPI, but rather a nonspecific elevation across most scales. Diagnosis is typically made following a clinical evaluation including an interview. A similar challenge faces clinicians using the NEO-PI (the assessment tool specifically designed for the FFM), but useful and relevant data are available from looking at the specific trait scores within factors, particularly within the factor neuroticism (Trull & McCrae, 1994). In addition, factor analysis on the dimensions of personality disorder yielded results that fit very well with the FFM, with several aspects of personality disorder (but not all) again being linked to neuroticism (Schroeder, Wormworth, & Livesley, 1994). It is important to remember, however, that the very idea of using a dimensional approach is based, in part, on an assumption:
…If one assumes that disordered personality is qualitatively different from normal personality, then the inclusion of a dimensional model of personality may be insufficient or inappropriate for investigation. If one assumes that disordered personality reflects quantitative differences in the manifestation or severity of normal personality traits (i.e., a dimensional approach), then the adoption of a personality taxonomy for use as a structural referent becomes a necessary or even fundamental conceptual task. (pp. 73-74; Wiggins & Pincus, 1994)
What, then, might personality disorders look like in terms of a dimensional description? Clark, Vorhies, and McEwen (1994) take an integrated approach based on two basic propositions pertaining to traits: first, that they are continuously distributed and exhibit wide individual variation; and second, that they are not fixed, but rather they are adaptations to the environment that are consistent within one’s individual range. These two points lead to the notions that a single trait structure can represent both normal and abnormal personality, that within the normal range there is great individual difference in each person’s characteristic and adaptive styles of thinking, feeling, and behaving, and that personality disorders are characterized by extreme and inflexible expressions of the normal personality structure. When examining data from individuals diagnosed with personality disorder, they have identified symptom clusters that form dimensions, or factors, which once again fit well with the FFM (Clark, et al., 1994). Widiger and several colleagues have actually offered five-factor translations of the standard categories of DSM-III and DSM-IV personality disorders (see Widiger, et al., 1994). The purpose of these translations is to take the personality disorder categories that psychologists are familiar with and put them in terms of the FFM. Consider two examples:
Paranoid Personality Disorder: Paranoid personality disorder (PAR) involves interpreting the actions of others as threatening or deliberately demeaning. These individuals tend to be suspicious, mistrustful, hypervigilant, and argumentative. According to the FFM, PAR is characterized primarily by excessively low agreeableness, particularly on the suspiciousness facet (a facet is one of the traits that makes up a factor). They are also characterized by the low agreeableness facets of very low straightforwardness and compliance, which represent the PAR tendencies to be secretive and oppositional. PAR is also characterized by the angry hostility facet of neuroticism, low extraversion, and low openness.
Antisocial Personality Disorder: Antisocial personality disorder (ATS) is characterized by irresponsible and antisocial behavior, and often involves criminal activity and a lack of regard for the rights of others. Within the FFM, they score excessively low on conscientiousness and agreeableness (particularly low on the facets of straightforwardness, altruism, compliance, and tendermindedness). They score high on the neuroticism facets of hostility, anxiety, depression, and impulsivity. However, so-called “successful” psychopaths may be characterized by very low levels of anxiety and self-consciousness.
(see Widiger, et al., 1994)
So, given the possibility of reconceptualizing personality disorders within the FFM, is it something we should do? Millon suggests that we view personality as the psychological equivalent of the body’s biological systems. Personality is, in this conception, a psychic system of structures and functions that lead to characteristic patterns of thought, feeling, and behavior. These characteristics cannot be viewed as simply normal or abnormal, since any specific element of the personality might be adaptive in one situation but maladaptive in another. Thus, the dimensional approach to describing personality provides a comprehensive picture in which little information of potential significance is lost (Millon, 1994). McCrae questions the very validity of Axis II of the DSM system, which appears to have little empirical support. He suggests that clinicians include in their diagnosis of patients a global assessment of the five personality factors. Thus, the diagnostic report would provide the necessary information on personality pertaining to the common symptoms and problems associated with either high or low scores on each factor (McCrae, 1994; Widiger, 1994). For example:
High Neuroticism: chronic negative affects, difficulty in inhibiting impulses, irrational beliefs
Low Neuroticism: lack of appropriate concern for potential problems in health or social adjustment, emotional blandness
High Agreeableness: gullibility, excessive candor and generosity, inability to stand up to others, easily taken advantage of
Low Agreeableness: cynicism and paranoid thinking, inability to trust, quarrelsomeness, too ready to pick fights, exploitative and manipulative, lying, rude and inconsiderate
(see McCrae, 1994)
Perhaps the most valuable aspect of any model used for classifying the personality disorders is its ability to provide guidelines for conceptualizing a treatment strategy. Sanderson and Clarkin (1994) have indeed found the NEO-PI useful in differential treatment planning. For example, the NEO-PI, in conjunction with a clinical interview, helps describe the typical interpersonal patterns of the patient, suggesting areas of difficulty needing treatment regardless of whether the therapy format is individual, family, or a group setting. In addition, the NEO-PI can help to identify which therapy format might be best suited to each particular patient. Although Sanderson and Clarkin (1994) caution that such conceptions still await empirical confirmation, they do offer some examples from their own supportive clinical experience. Likewise, MacKenzie (1994) offers numerous specific examples from cases in which factor scores provided clear target areas for focusing therapy. For example, a women who scored high in agreeableness acknowledged that she repeatedly got into relationships in which she felt used, a teacher who scored very high on openness was overly stimulated in new situations and felt overwhelmed with creative ideas, and a man who scored low on conscientiousness felt stuck in life, having worked only itinerant construction jobs despite having earned a graduate degree in college. In each case, the NEO-PI data matched the clinical presentation quite well, suggesting that the FFM would indeed be an effective conceptualization of treatment strategies for personality disorder issues (as well as, presumably, for other psychological and adjustment disorders).
The diagnosis of personality disorders, whether categorical or dimensional, remains a controversial topic. Of even greater concern, is the resistance of these disorders to treatment. However, the FFM appears to offer an advantageous way of describing personality disorder as an extreme extension of normal personality dimensions, and the NEO-PI scales offer practical direction with regard to treatment strategies. Change, however, may not come easily:
Some observers have said that what is at issue here is the American Psychiatric Association “versus” the American Psychological Association. In other words, the potential conflict between psychiatric/categorical and psychological/dimensional models could stall progress in this field. It has been suggested that the American Psychological Association should issue a rival DSM that uses a dimensional approach. We believe a far better solution would be cooperation between the two approaches, which would lead to more coordinated research and shared clinical experience. (pg. 325; Costa & Widiger, 1994).