MMPI scales for DSM-III personality disorders: their derivation and correlates

Scales tapping the dimensions of personality disorder as represented in the DSM-III were derived using a combined rational/empirical strategy. The final version of the scales demonstrated both content validity as well as internal consistency. Correlations between the derived scales and between these scales and the original MMPI clinical scales were generally found to be in the expected direction providing preliminary evidence of criterion related validity. It is suggested that the derived scales may represent an advance toward the reliable assessment of DSM-III personality constructs.     

Estimation of MCMI DSM-III axis II constructs from MMPI scales and subscales

Wiggins, Harris and Lingoes, and Serkownek Minnesota Multiphasic Personality Inventory (MMPI) scores were used to predict Millon Clinical Multiaxial Inventory (MCMI) scores in a 100-patient sample. Equations from the first sample were cross-validated on a sample of 212 inmate subjects. We conclude that scores on 19 of the 20 MCMI scales can be successfully predicted by the Wiggins, Harris and Lingoes, and Serkownek subscales of the MMPI. In further cross-validation, the equations were used to predict the Morey, Waugh, and Blashfield MMPI composites for the prison sample, again with strongly positive results. The results appear quite promising for the estimation of personality disorder constructs from MMPI scales and subscales.     

Factor structure of the MCMI basic personality scales and common-item artifact

The eight basic personality scales of the Millon Clinical Multiaxial Inventory (MCMI) were derived from Millon's theory of personality, but the adequacy of the MCMI for measuring Millon's personality constructs has never been assessed. One major problem with using factor analysis to illuminate the structure of the MCMI personality scales is that artifactual structure may result from item overlap among the scales. To analyze this, item-overlap coefficients were factored and compared to the factor structures of five subject samples. For the eight basic personality scales, three factors emerged for the overlap matrix and each of the five sample matrices: Aloof-Social, Aggressive-Submissive, and Lability-Restraint. It was concluded that these three factors are inconsistent with Millon's theory and that they will be found artifactually across a wide variety of populations due to overlapping items.     

The relationship between the MCMI personality scales and DSM-III, axis II

Compared were the personality scales of the Millon Clinical Multiaxial Inventory (MCMI) to the diagnosis of personality disorder, according to the Diagnostic and Statistical Manual of Mental Disorders (3rd ed. [DSM-III]; American Psychiatric Association, 1980), obtained by means of the Structured Interview for the DSM-III Personality Disorders (SIDP). The results from 272 psychiatric outpatients show a good correspondence for the Avoidant and the Dependent scales, a fairly good correspondence for the Schizotypal, the Histrionic, the Borderline, the Narcissistic, and the Paranoid scales, and no correspondence for the Schizoid, the Passive-Aggressive, and the Compulsive scales. The Passive-Aggressive scale seems to be positively correlated to personality disorders in general, whereas the Compulsive scale seems to be negatively correlated to a number of personality disorders.     

Refining personality assessments by combining MCMI high point profiles and MMPI codes. Part II. MMPI code 27/72

The relationship between the MMPI 27/72 code type and MCMI high point codes is examined for 228 psychiatric patients. MCMI high point codes that correspond highly with MMPI 27/72 code enable the clinician to identify three discrete personality subgroups entitled: Fearfully Dependent, Conforming-Dependent, and Ambivalently Dependent. The central features of each of these groups are described, and their clinical characteristics used as a basis for differentiating ambiguities and contradictory interpretations commonly given the MMPI 27/72 code type.     

Circular structure of the MCMI-III personality disorder scales.

Millon's (1987) circular model of personality disorders was examined in a large sample of psychiatric patients (N = 2,366) who completed the Millon Clinical Multiaxial Inventory-III (MCMI-III; Millon, 1997) as part of routine assessment after presentation for treatment. Principal components analyses were conducted to identify the first two dimensions in MCMI-III base rate scores, weighted and unweighted raw scores, and nonoverlapping scale scores. Similar analyses were made on these scores when acquiescence was partialled out. Circular plots of the scales were examined against Millon's hypothesized arrangement and the model was tested using confirmatory factor analysis. Results replicated those of Strack, Lorr, and Campbell (1990) with the MCMI-II. Millon's horizontal Impassive-Expressive dimension was recovered in both regular and residual scores but the vertical axis appeared to represent an Impulsivity-Compulsivity dimension rather than the Autonomous-Enmeshed continuum envisioned by Millon. Although scale order followed Millon's predictions for the most part, a number of departures from theoretical expectations were noted and none of the score sets yielded a good fit to the hypothetical structure. Millon's model appears to have promise as a circumplex that can encompass all of the personality disorders but changes are needed to rectify discrepancies between the theory and empirical findings.     

The use of the MMPI special scales in the assessment of delinquent personality.

The MMPI was administered to 401 adjudicated male juvenile offenders, ages 13 to 17, who had been referred to a juvenile court mental health unit for recommendations in court disposition. Mean T scores generated by 86 MMPI clinical and supplementary scales were tabulated and analyzed as to their relationships with completed court offense histories. Twenty-eight of these scales that best described and discriminated between high and low levels of recidivism, crime severity, assault, and property offenses were identified, profiled, and discussed. Data were analyzed by a suggested T-score cutoff of 1.5 standard deviations, univariate, and multivariate (discriminant) analysis. Scales that best discriminated between high and low levels of recidivism and property offenses were Pd (Psychopathic Deviance), Ma1 (Amorality), Pd2 (Authority Problems), Re (Social Responsibility), and MAC (MacAndrew Alcoholism). Low scores on Re (Social Responsibility), Es (Ego Strength), Cn (Control), and St (Social Status) best discriminated between high and low levels of crime severity. High and low levels of assault were best differentiated by Ma3 (Imperturbability), Scale 5 (Masculinity-Femininity), and Pa2 (Poignancy). Although the linear combination of Scales F, 4, 8, and 9 has some utility in the areas of recidivism and property offenses, its discriminative ability is extremely poor for levels of assault and crime severity.     

Discriminant validity of the MMPI for borderline personality disorder

This study explores the discriminant validity of the MMPI in relationship to the DSM-III and the Diagnostic Interview for Borderlines (DIB) constructs of Borderline Personality Disorder (BPD). A two-way analysis of variance model assessed differences between inpatients (n = 42) and outpatients (n = 42) and between BPD and nonBPD patients. We compare the best discriminant model for the current samples with previous BPD discriminant functions. The present study assesses a wider range of psychopathology in which the MMPI appears to be relatively insensitive to the BPD construct while retaining the high specificity reported in previous studies. Implications for the use of the MMPI as a diagnostic instrument are discussed.     

Perfectionism and multiphasic personality inventory (MMPI) indices of personality disorder

The present research tested the hypothesis that personal and social aspects of the perfectionism construct are related differentially to indices of personality disorders. A sample of 90 psychiatric patients was examined with respect to their scores on the Multidimensional Perfectionism Scale (MPS) and the personality disorder subscales (PDS) of the Minnesota Multiphasic Personality Inventory. The MPS provides measures of self-oriented, other-oriented, and socially prescribed perfectionism, whereas the PDS assess levels of various personality disorder symptoms. Zero-order and partial correlations indicated that the perfectionism dimensions of the MPS were related to various subscales of the PDS and, perhaps more importantly, that the findings vary as a function of the perfectionism dimension in question. The results are discussed in terms of the importance of perfectionism in personality disorder symptom patterns.     

Maximum likelihood estimates of item parameters using the logistic function

The logistic function is proposed as an alternative to the integrated normal function when estimating parameters of test items. The logistic curve is described; an iterative method for finding maximum likelihood estimates of its parameters is given, and an example of its use is presented.     

On the consistency of the MMPI in borderline personality disorder

14 patients diagnosed as having Borderline Personality Disorder were examined on two separate occasions with the MMPI. Results strongly suggest that, as a group, borderline patients are quite consistent responders on the MMPI under test-retest conditions (1 to 58 mo.). The frequency distribution of scale elevations and code types was also examined which showed the often-noted heterogeneity of profile code types within this population. Thus, a note of caution is advised for those investigators who assume that a prototypical borderline MMPI profile exists. Specifically, considerable heterogeneity is seen among this sample of patients with respect to a two-point coding strategy, and individual codes change over time.     

Empirical evaluation of the MCMI-III personality disorder scales

The MCMI-III personality disorder scales were empirically validated with a sample of 870 clinical patients and inmates. Prevalence rates of personality disorders were in general lower on the MCMI-III than clinical ratings, but trait prevalence was generally higher; thus a base rate of 75 on the MCMI-III could be a guideline in the screening of trait prevalence. However, the sensitivity of some MCMI-III scales was very low. Moreover, the correlations of most personality disorder scales of the MCMI-III were significant and positive with corresponding measures on clinical ratings and MMPI-2 personality disorder scales, but these were, in general, not significantly higher than some other correlations. As a consequence the discriminant validity seems to be questionable. The MCMI-III alone cannot be used as a diagnostic inventory, but the test could be useful as a screening device as a part of a multimethod approach that allows aggregation over measures in making diagnostic decisions.     

The prediction of drug use among college students from MMPI, MCMI, and sensation seeking scales

This study examines the ability of five self-report assessment measures to predict college students' drug use across 12 pharmacological drug categories. Subjects were 125 female and 61 male university undergraduate students. The test battery included the following instruments: the Psychopathic Deviancy (Pd) scale of the Minnesota Multiphasic Personality Inventory (MMPI); the MacAndrew Alcoholism scale (MAC), a special scale of the MMPI; the Sensation Seeking Scale (SSS); the Millon Alcohol Abuse Scale; and the Millon Drug Abuse Scale. Scores from these instruments were utilized in linear combinations to predict individual drug use outcomes as well as polydrug versus single drug use patterns. The drug categories included in this investigation are coffee, tobacco, alcohol, marijuana, tranquilizers, depressants, amphetamines, LSD, other hallucinogens, solvents, narcotics, and cocaine. Results demonstrate significant and meaningful relationships between predictors and drug use levels among college students. Weighted equations derived from linear discriminant function analyses were generally capable of accurately classifying subject's drug use levels across drug categories and in discriminating single drug use from polydrug use patterns. Typically, the Sensation Seeking Scale entered as the most powerful predictor of substance use and abuse.     

Convergent validity of the MCMI-III personality disorder scales and the MMPI-2 scales

The MCMI-III personality disorder scales (Millon, 1994) were empirically validated in a sample of prisoners, psychiatric inpatients, and outpatients (N = 477). The scale intercorrelations were congruent with those obtained by Millon, Davis, and Millon (1997). We conclude that our Flemish/Dutch version shows no significant differences with the original version of the MCMI-III as far as intercorrelations are concerned. Convergent validity of the MCMI-III personality disorder scales was evaluated by the correlational data between the MCMI-III personality disorder scales and the MMPI-2 clinical (Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989) and personality disorder (Somwaru & Ben-Porath, 1995) scales. Improved convergence was obtained compared with previous versions of the MCMI-I. Only the compulsive MCMI-III personality disorder scale remains problematic. The scale even showed negative correlations with some of the related clinical scales and with the corresponding personality disorder scales of the MMPI-2.