Most test items are so specific that they usually have little relevance to the overall assessment. For example, the item, "I like to go to parties," may or may not be indicative of a histrionic personality; not everyone who likes parties is a histrionic. Some items, however, are so dramatic that they are interpreted on their own terms. For example, if a patient responds "true" to the item, "I have been thinking strongly about killing myself," the assessing clinician has the responsibility to establish the likelihood of suicidal intent by interviewing the subject. Such critical items are literally critical to the clinical situation. These are usually built into the inventory by intent, though some may be identified through research after the test has been constructed. After the patient has completed the inventory, answers to all the items, especially the critical items, can be quickly scanned by the clinician, suggesting issues that should be explored further during the clinical interview.
Because individual test items usually refer to highly specific behaviors, they do not make broad predictions about behavior. For this reason, items are usually grouped into scales. Taken together, items such as, "I like to go to parties," "I am a dramatic and emotional person," and "I like to be the center of attention," begin to point to a histrionic pattern. The scale thus makes a broader prediction about behavior but loses some specificity in the process. Not all histrionics will answer affirmatively to, "My thoughts are scattered and hard to focus."
Ideally, every scale item should tap some aspects of the construct the scale is intended to assess. When all the important aspects of a construct have been anchored to different items, the scale is said to possess content validity. The narcissistic personality, for example, consists largely of the traits of grandiosity, exploitiveness, and lack of empathy. As such, any scale lacking items that assess grandiosity cannot be a valid measure of the narcissistic construct, as content essential to the construct is missing. Careful consideration of the different facets of every construct is, therefore, essential to scale development. Scales that perform in accordance with the expectations of psychological theory are said to possess the additional property of construct validity (Cronbach & Meehl, 1955). If a new antisocial personality scale fails to correlate highly with an established measure of substance abuse, for example, this calls the validity of the antisocial scale into question.
Tests are given to a large number of subjects, called the normative sample, to determine what is expectable and what is statistically deviant. Although any scale can be interpreted on its own, whole inventories consisting of many scales can be constructed simultaneously using the same sample. When any one person completes the inventory, his or her scores can be graphed as a profile configuration. The two or three highest scales in the profile are usually called a codetype. The profile stands in place of the person just as a set of items stands in place of its scale. For interpretive purposes, the profile is the person. Accordingly, the scales of an inventory should exhaust all of personality, just as the items that assess a construct tap every aspect of its content. The scales must have content validity for the person. Inventories developed according to some theoretical or methodological rationale provide some assurance that the individual has been assessed along the essential dimensions of personality and thus ultimately support the content validity of the clinical report that will eventually be written on the subject.
A variety of self-report instruments are available that assess the personality disorders.
Minnesota Multiphasic Personality Inventory-2nd Edition (MMPI-II)
With more than 550 items, the MMPI-2 (Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989) is not so much a standardized test as a standardized item pool that belongs to psychology itself. Literally hundreds of personality scales have been derived from the MMPI throughout its long career. In fact, there are now more auxiliary scales than there are items on the MMPI (Graham, 1990). Morey, Waugh, and Blashfield (1985) constructed a set of MMPI-I scales to represent the 11 DSM-III personality disorders, based on the strategy used by Wiggins (1966) in the construction of the Wiggins content scales. Item selection proceeded through two stages. In the initial phase, scales were rationally derived by four experienced clinicians who culled the item pool for items representative of DSM-III personality disorder criteria. Those items selected by two or more clinicians formed the preliminary scales; items could be assigned to more than one scale, mirroring the diagnostic overlap of DSM-III. These were then subjected to empirical refinement. Nonoverlapping scales were constructed by assigning each overlapping item to the scale with which it exhibited the highest correlation. The final scales consist of from 14 to 38 items for the overlapping scales and from 13 to 20 items for the nonoverlapping scales. As should be expected, the internal consistencies of the longer, overlapping scales are appreciably higher, ranging between 0.675 (compulsive scale) and 0.859 (avoidant scale). Those of the nonoverlapping scales range from 0.619 (histrionic scale) to 0.791 (schizotypy scale). These internal consistencies are superior to those of the clinical scales and comparable to those of the Wiggins content scales. Specific item assignments are available in Morey et al. Norms for the updated MMPI-2 have been supplied by Colligan, Morey, and Offord (1994).
Millon Clinical Multiaxial Inventory (MCMI)
Now in its third edition (MCMI-III; Millon, Davis, & Millon, 1996), the MCMI is by far the most widely used personality disorders inventory. A principal goal in constructing the MCMI-III was to keep the total number of items constituting the inventory small enough to encourage use in all types of diagnostic and treatment settings, yet large enough to permit the assessment of a wide range of clinically relevant behaviors. At 175 items, the final form is much shorter than are comparable instruments, with terminology geared to an eighth-grade reading level. As a result, most subjects complete the MCMI-III in 20 to 30 minutes. The inventory is intended exclusively for subjects believed to possess a personality disorder and should not be used with normals. The MCMI is frequently used in research. More than 650 publications to date have included or focused primarily on the MCMI, with approximately 65 new references currently published annually.
The inventory itself consists of 24 clinical scales (presented as a profile in Figure 4.1) and three modifier scales—Disclosure, Desirability, and Debasement—which identify tendencies to overdisclose or underdisclose pathology, favor only socially desirable responses, or endorse only those highly suggestive of pathology, respectively. The next two sections constitute the basic personality disorder scales. The first section contains moderately severe personality pathologies, ranging from schizoid to masochistic, and the second section represents the severe personality pathologies—the schizotypal, borderline, and paranoid. The masochistic and sadistic personalities, included in the third revised edition of the DSM, but not in the fourth edition, have been retained in the
The Assessment of Personality
CONFIDENTIAL INFORMATION FOR PROFESSIONAL USE ONLY
SCORE RAW BR O
PROFILE OF BR SCORES 60 75 85
2A 2B 3
CLINICAL 4 PERSONALITY S PATTERNS 6A
ll ll l2
SEVERE S PERSONALITY C PATHOLOGY P
BIPOLAR: MANIC DISORDER
SS CC PP
FIGURE 4.1 Millón Clinical Multiaxial Inventory-III.
MCMI-III. The next two sections cover the Axis I disorders, ranging from the moderate clinical syndromes, such as anxiety and dysthymia, to those of greater severity, such as thought disorder and delusional disorder (Millon, 1997).
The MCMI can be used on a routine basis in outpatient clinics, community agencies, mental health centers, college counseling programs, general and mental hospitals, the courts, and private practice offices. The division between personality and clinical disorders scales parallels the multiaxial model and has important interpretive implications. The resulting profile is helpful in illuminating the interplay between long-standing char-acterological patterns and current clinical symptoms. Scores on the personality and clinical syndromes scales run from 0 to 115, with those above 85 suggesting pathology in the disordered range. More comprehensive and dynamic interpretations of relationships among symptomatology, coping behavior, interpersonal style, and personality structure may be derived from an examination of the configural pattern of the clinical scales. To further increase its coordination with the DSM, the items that assess the personality disorders have been written to reflect the diagnostic criteria of their respective constructs. For example, the first criterion for the DSM-IV dependent personality disorder reads: "Has difficulty making everyday decisions without an excessive amount of advice and reassurance from others," and its parallel MCMI-III item reads: "People can easily change my ideas, even if I thought my mind was made up."
Computer-generated clinical reports are available at two levels of detail. The Profile Report of scale scores is useful as a screening device to identify patients who may require more intensive evaluation or professional attention. Individual scale cutting lines on the MCMI-III can be used to make decisions concerning primary behavior disorders or syndrome diagnoses. Similarly, elevation levels among subsets of scales can furnish grounds for judgments about impairment, severity, and chronicity of pathology. More comprehensive and dynamic interpretations of relationships among symptomatology, coping behavior, interpersonal style, and personality structure may be derived from an examination of the configural pattern of all 24 clinical scales. Alternatively, the Interpretive Report provides both a profile of the scale scores and a detailed analysis of personality and symptom dynamics as well as suggestions for therapeutic management.
A number of other self-report instruments are available. Notable are two variants of the MMPI and MCMI designed for adolescents; the first, the MMPI-A (Butcher et al., 1992) is a bit briefer than the MMPI, as is the Millon Adolescent Clinical Inventory (MACI) (Millon, 1993). Another recent variant of the MCMI is the M-PACI, the Millon Preadolescent Inventory (Millon, Tringone, Millon, & Grossman, in press) for use with youngsters in the 9- to 12-year age range. The Personality Diagnostic Questionnaire (e.g., Hyler & Rieder, 1987) is now in its fourth revision. F. L. Coolidge and Merwin (1992) reported on the reliability and validity of the Coolidge Axis II Inventory. The Personality Assessment Inventory (Morey, 1992) consists of 344 items on 4 validity scales, 11 clinical scales, 5 treatment scales, and 2 interpersonal scales. Only three scales, however—Paranoia, Borderline Features, and Antisocial Features— directly assess personality pathology. The Dimensional Assessment of Personality Pathology-Basic Questionnaire was constructed by Livesley (1987) and his associates (Livesley & Schroeder, 1990; Livesley et al., 1992; Schroeder, Wormworth, & Lives-ley, 1992) based on an extensive literature review and a comprehensive scale development effort. The Schedule of Nonadaptive and Adaptive Personality (Clark, McEwen,
Collard, & Hickok, 1993) is a 375-item true-false instrument primarily oriented to trait dimensions derived from factor analytic research. However, diagnostic scales for the DSM personality disorders are also included. The Tridimensional Personality Questionnaire (Cloninger, Przybeck, & Svrakic, 1991) is based on novelty seeking, harm avoidance, and reward dependence, temperament dimensions proposed by Cloninger (1987b). Finally, the Wisconsin Personality Disorders Inventory (Klein et al., 1993) is a 360-item inventory developed to operationalize the interpersonal theory of Benjamin (1996). Items were developed to represent the internal experience of each personality disorder as conceived from the perspective of the SASB. The NEO-PI-R (Costa & McCrae, 1992), originally designed to assess normal personality factors, has also been employed recently to evaluate clinical personality traits.
A number of clinical interviews are available for the personality disorders. Two of the more widely used are reviewed.
Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II)
The SCID-II (First, Gibbon, Spitzer, Williams, & Benjamin, 1997) is a semistructured diagnostic interview assessing the 12 personality disorders included in DSM-IV; the sadistic and masochistic personality constructs from the third revised edition of the DSM are not included. According to First and colleagues, the interview has often been used in research settings to describe the personality profiles found in particular samples or to select patient groups for further study. In clinical settings, the SCID-II may be used routinely as part of a standard intake. Alternatively, a subset of the interview may be used to confirm the presence of a suspected personality disorder. Phrased to coordinate with the language employed in the DSM diagnostic criteria, the interview questions are grouped by personality disorder (e.g., all the avoidant personality questions are asked together) and scored absent, subthreshold, true, or inadequate information to code. All available sources of information can be used for scoring, not just the subject's own report. The SCID-II can even be administered to an informant. Specific probes are included to assist in exploring the presence of each interview item. The SCID-II also includes a 119-item self-report screening questionnaire that can greatly reduce administration time. Each self-report question corresponds to an interview item but is asked in such a way that it elicits a much greater frequency of true responses. The questionnaire thus serves as a screening inventory, for the interviewer need only scan the completed form and inquire into positive admissions.
Structured Interview for DSM-IV Personality Disorders (SIDP-IV)
The SIDP-IV (Pfohl, Blum, & Zimmerman, 1997) is a semistructured clinical interview that assesses all the personality disorders of the DSM-IV, plus the self-defeating personality from the revised third edition of the DSM (the sadistic personality is not included). Whereas the questions of the SCID-II are grouped by disorder and closely rephrase the diagnostic criteria, those of the SIDP-IV are phrased more conversationally and grouped into 10 topic areas, such as interests and activities and emotions. Using this more natural format, information elicited by previous questions is more readily applied to others with the same theme. Interviewing and scoring typically take 80 to 120 minutes. An informant may consume another 20 to 30 minutes. The authors suggest that administration time can be reduced by dropping questions from the optional personality disorders (the self-defeating, depressive, and negativistic personalities). An alternative form of the SIDP-IV is available with the questions grouped by disorder. Subjects are encouraged to respond according to "what you are like when you are your usual self." Because personality is enduring over time, interviewers are required to use the five-year rule, meaning that "behaviors, cognitions, and feelings that have predominated for most of the last five years are considered to be representative of the individual's long-term personality functioning." Items are scored not present, subthreshold, present, and strongly present. Tentative ratings may be made during the interview, but a final rating, based on all available data, is deferred until the end. The manual states that the interview has been used in more than 60 studies and translated into several languages.
Other interviews have been developed specifically to research particular personality disorders, including the depressive personality (Gunderson, Phillips, Triebwasser, & Hirschfeld, 1994), the narcissistic personality (Gunderson, Ronningstam, & Bodkin, 1990), and the borderline personality (Zanarini, Gunderson, Frankenburg, & Chauncey, 1989). Each includes questions geared to traits associated with its respective construct, as manifest in various domains of functioning. The Diagnostic Interview for Narcissism, for example, assesses narcissism in terms of grandiosity, interpersonal relations, reactiveness, affects and mood states, and social and moral adaptation. Because these interviews focus closely on one personality alone, they require only about an hour to administer.
Was this article helpful?