Harold R. Miller

Performance of psychiatric inpatients and intellectually deficient individuals on a task that assesses the validity of memory complaints.

A brief psychological screening test, devised by Rey (1964), was developed to assess the validity of memory complaints. Although Rey hypothesized that malingerers would be mislead to perform poorly while even individuals with severe concentration problems could succeed, a review of the literature did not reveal any empirical reports that examined the actual performance of nonmalingering though disturbed patients. Therefore, Reys test was administered to samples of acutely disturbed psychiatric patients (N = 50) and intellectually deficient individuals (N = 16). The results confirmed the criterion proposed by Lezak (1983) and suggested that malingering should be considered among individuals who deny remembering at least 9 of the 15 times of the Rey test.

What's in a name? The MMPI-2 PTSD scales

The MMPI-2 post-traumatic stress disorder scales (PK and PS) were examined with a sample of 96 outpatient anxiety disorder and 97 outpatient traffic accident subjects. PK and PS correlated .96 and showed highly similar correlations with MMPI-2 validity, clinical, content and supplementary scales for both groups, which indicated that these scales are indices of MMPI first-factor variance. This was supported by separate principal components analyses for each group, in which PK and PS contributed the most variance to the first component. Although the groups showed different MMPI-2 mean profiles, neither PK nor PS contributed to discriminant function classification of group members. It is suggested that PK and PS are indices of general emotional distress and maladjustment; implications for their use in diagnosing post-traumatic stress disorder were discussed.

Using the Millon clinical multiaxial inventory's scale B and the MacAndrew alcoholism scale to identify alcoholics with concurrent psychiatric diagnoses

The ability of the MacAndrew Alcoholism Scale (MAC) and Scale B of the Millon Clinical Multiaxial Inventory (MCMI) to discriminate independently defined alcoholics with psychiatric diagnoses from other psychiatric patients was examined for males and all alcoholics, using three different criteria of alcoholism. The MAC identified from 80% to 87% of male alcoholics and 76% to 82% of all alcoholics, and from 25% to 52% of male nonalcoholics and 55% to 75% of all nonalcoholics, resulting in a large number of false positives. Scale B identified from 31% to 60% of male alcoholics and 33% to 43% of all alcoholics, and from 73% to 85% of male nonalcoholics and 85% to 94% of all nonalcoholics. The operating characteristics of Scale B showed that it was not as efficient in identifying alcoholics as previous work had indicated. We suggest that Scale B should not be used and the MAC should be used cautiously to discriminate alcoholics with psychiatric disorders from patients without alcoholism.

Development of a three-scale MMPI: The MMPI-TRI

A 60-item short form of the MMPI with very high content validity and items that appear on both the MMPI and MMPI-2 was developed and named the MMPI-TRI. It contains three 20-item scales--the Subjective Distress, Acting-Out, and Psychosis scales. These three scales have excellent internal consistency and sufficient independence from each other. An anxiety and depression group of patients, prison inmate group, and a schizophrenic and other psychotic group had the highest mean scores on Subjective Distress, Acting-Out, and Psychosis, respectively. Correlations with the 13 regular scales of the MMPI and MMPI-2, their content and supplementary scales, and four other psychometric instruments provided very strong evidence for validity. Norms are provided.

Maximum Likelihood Estimates of the Accuracy of Four Diagnostic Techniques

This study examined the sensitivities and specificities of four diagnostic techniques used in psychiatry; the clinical interview, the Diagnostic Interview Schedule (DIS), the Psychiatric Diagnostic Interview (PDI), and the Millon Clinical Multiaxial Inventory (MCMI). Subjects consisted of 207 patients drawn from in- and out-patient units, and 32 people from a correctional institution. A maximum likelihood estimation approach was used to estimate the sensitivity and specificity of each instrument. It was found that no one technique was clearly superior to the others. The psychometric properties were highly dependent upon the diagnosis. It was concluded that structured interview schedules are preferred for making diagnoses, but that further development of them is necessary. Self-report scales may be most useful in the description and prediction of individuals behaviors in different situations.

The Harris-Lingoes subscales: Fact or fiction?

The subjectivity of the Harris-Lingoes MMPI content subscales was examined by asking expert judges (N = 13) to group items from appropriate clinical scales that represented similar content, attitudes or traits. The mean subgroups were compared to the Harris-Lingoes subscales, and item groupings were consensually validated and replicated. Judges developed more content categories per scale than Harris and Lingoes, but showed relatively little agreement on item groupings. Nine consensually validated and replicated subscales were highly similar to nine Harris-Lingoes subscales, while nine other replicated subscales were moderately similar to seven Harris-Lingoes subscales. Twelve Harris-Lingoes subscales were unrelated to the replicated subscales.

Short, Shorter, Shortest The Efficacy of WAIS-R Short Forms With Mixed Psychiatric Patients

The accuracy of different subtest combinations in predicting Verbal, Performance, and Full Scale IQ, scores on the revised edition of the Wechsler Adult Intelligence Scale (WAIS-R) was examined. A best subset multiple regression technique with sample sizes ranging from 1,062 to 1,084 individuals was used. All possible subtest combinations were analyzed to determine the ones that showed the three best and the single worst multiple Rs with IQ scores from the complete WAIS-R. All short forms provided accurate estimates of Full Scale scores, and predictive accuracy increased slightly as more subtests were used. The data suggest that statistically searching for a “best” short form is largely futile. Short forms should be selected on the basis of their efficiency to provide the information required.