908 Journal of Clinical Psychology, November. 1986, Vol. 42, No. 6

GAA, J. P., LIBERMAN, D., & EDWARDS, T. A. (1979). A comparative factor analysis of the Bern Sex Role Inventory and the Personality Attributes Questionnaire. Journal of Clinical Psychology, 35, 3 .

L m E m , D., & GAA, J. P. (1980). Response tendency on the Bern Sex Role Inventory. Journaiof Psychology, 106, 259-263.

NUNNALLY, J . P. (1978). Psychometric theory. New York: McGraw Hill.

CONCURRENT VALIDITY OF THE CLINICAL SYMPTOM SYNDROME SCALES OF THE MILLON CLINICAL MULTIAXIAL INVENTORY ROBERT C. MCMAHON ROBERT S . DAVIDSON

University of Miami Corai Gables, Florida

Veterans Administration Medical Center Miami, Florida

This study examined relationships between the clinical symptom syndrome scales of the Millon Clinical Multiaxial Inventory (MCMI) and the various clinically meaningful mood or symptom states measured by the six Profile of Mood States (POMS) scales (N = 243). The MCMI symptom scale- POMS symptom/mood scale relationships found in this study were com- pared with MCMI symptom scale-MMPI and SCL-90 symptom/mood scale relationships reported in the MCMI manual (Millon, 1983). Results of the present investigation, when combined with results of the previous analyses reported in the MCMI manual, reveal a number of consistent associations of moderate strength between MCMI symptom scales and selected mood or symptom scales from the MMPI and Symptom Checklist-90 as well as from the POMS. Although most relationships between the MCMI symptom scales and the symptom/mood scales of the POMS, MMPI and SCL-90 were consistent with expectation, the Anxiety, Dsythymia and Psychotic Depression scales of the MCMI show limited ability to discriminate appropriately between anxiety and depression in several of the concurrent validity analyses considered herein.

The Millon Clinical Multiaxial Inventory (MCMI) was developed to assess and differentiate among basic maladaptive and pathological personality styles and various clinical symptom syndromes. The 11 MCMI personality scales are compatible with each of DSM 111’s Axis 11 personality disorder categories and are assumed to reveal enduring and pervasive traits. The 9 MCMI symptom scales are designed to be compatible with common DSM I11 Axis I clinical symptom syndrome categories. Over the past few years, an accumulating body of research has been published that deals with the stability (McMahon, Flynn, & Davidson, 1985), factorial validity (Flynn & McMahon, 1984; Millon, 1983), and clinical utility (Craig, Verinis, & Wexler, 1985; McMahon & David- son, 1986; Robert, Ryan, McEntyre, McFarland, Lips, & Rosenberg, 1985) of the MCMI. In addition, McMahon and Davidson (1985) recently completed a study that examined Millon’s hypothesis that particular MCMI personality types are disposed not only to certain patterns of cognition, affect, and behavior, but also to particular clinical symptomology - particularly under stress. Indeed, the illumination of “interplay between longstanding characterological patterns and the distinctive clinical symptomology a pa- tient manifests under psychic stress” is a major mission of the MCMI (Millon, 1977, p. 2). Although considerable interest has been devoted to the MCMI personality scales, comparatively little has been done to establish the external validity of the MCMI symp- tom scales beyond Millon’s initial efforts.

This study examined relationships between the clinical symptom syndrome scales of the MCMI and the various clinically meaningful mood or symptom states as measured

MCMI 909

by the six Profile of Mood States (POMS) scales (McNair, Lorr, & Droppleman, 1971). The MCMI symptom scale - POMS symptom/mood scale relationships found in this study will be compared with MCMI symptom scale- MMPI and SCL-90 (Derogatis, Lipman, & Covi, 1973) symptom/mood scale relationships reported in the MCMI manual (Millon, 1983).

METHOD Subjects

The data used in this investigation were collected from inpatient alcoholics who were treated at the Veteran’s Administration Medical Center in Miami, Florida. Data were collected from all consecutive admissions to the Alcohol and Drug Dependence Unit between July 1977 and spring of 1981. The subjects were 243 predominantly Caucasian males with a mean age of 41.4 (SD = 12) years and an average educational attainment of 12.7 (SD = 2.5) years. All subjects participated in the inpatient alcohol treatment unit at the Miami VA Medical Center, which incorporates both individual and group therapy and is designed after a behaviorally oriented therapeutic community model. Procedure

All of the patients in the study completed the MCMI and POMS after detox and during the first week after admission to the inpatient unit. Subjects were instructed in- dividually in the proper completion of the forms by research personnel in the Psychology Research Department of the Veterans Administration Medical Center.

RESULTS Table 1 reveals correlations between the nine symptom scales of the MCMI and

the six POMS scales administered at admission to a VA alcohol dependence treatment program (Nie, Hull, Jenkins, Steinbrenner, & Bent, 1975). The comparison data reported below consist of correlations between MCMI symptom scales and both MMPI (Basic and Wiggins Content) scales and SCL-90 scales that measure depression, anxiety/phobias, hostility and mania, which were administered to large, mixed psychiatric samples that included men, women, inpatients, and outpatients (Millon, 1983).

Table 1 Correlations Between MCMI Symptom Scales and POMS Scales

MCMI scale

~~ ~~~~~

POMS POMS POMS POMS POMS POMS Tension- Depression- Anger- Vigor- Fatigue- Confusion- Anxiety Dejection Hostility Activity Inertia Bewilderment

Anxiety .45 .44 .21 - .28 .46 .49

Somatoform .35 .26 .13 - .14 .39 .36 Hypomanic .14 .10 .25 .26 .I2 .12

Dsythymic .46 .47 .22 - .33 .46 .49

Alcohol Abuse .27 .29 .19 - .02 .21 .26 Drug Abuse . l l .13 .27 .20 .07 .13

Psychotic Thinking .36 .47 .38 - .22 .34 .49

Psychotic Depression .47 .54 .36 - .21 .46 .56

Psychotic Delusions .01 .10 .13 .04 .08 .05

Note.-Coefficients 2 .10 are significant at .05 level. Coefficients 2 .20 are significant at .01 level.

910 Journal of Clinical Psychology, November 1986, Vol. 42, No. 6

The MCMI Anxiety scale is designed to measure restlessness and apprehension related to physiological overarousal, which may be connected with either specific phobic stimuli or experienced in a more generalized manner. The MCMI Anxiety scale showed moderate (2 .35) correlations with POMS Tension-Anxiety, Depression-Dejection, Fatigue-Inertia, and Confusion-Bewilderment. Millon (1983) reports moderate associa- tions between MCMI Anxiety and SCL-90 Anxiety (.67), SCL-90 Phobic Anxiety (.47) and MMPI Wiggins Phobias (.47). Moderate associations also are reported between the MCMI Anxiety scale and MMPI Wiggins Depression (.67), MMPI Basic Depression (.57), and SCL-90 Depression (.41). Thus, the MCMI Anxiety scale appears to be associated positively with various aversive symptom/mood states as measured by three independent diagnostic instruments. In addition, this MCMI scale was associated con- sistently and moderately with independent measures of both anxiety/phobia and depres- sion and does not appear to distinguish between these conditions in the clinical groups considered in this analysis.

The MCMI Somatoform scale was designed to tap a pattern that included com- plaints of recurrent fatigue, weakness, and a variety of dramatic, but nonspecific discom- forts in unrelated parts of the body. This MCMI scale, which reportedly measures the tendency to express psychological difficulties through somatic channels, showed moderate correlations with POMS Tension-Anxiety, Fatigue-Inertia, and Confusion-Bewilderment. Millon (1983) reported moderate associations between the MCMI Somatoform scale and MMPI Wiggins Depression (.54), MMPI Basic Depression (.42), SCL-90 Depression (.38), MMPI Wiggins Phobias (.42), SCL-90 Anxiety (.36) and SCL-90 Phobic Anxiety (.34). Thus, the MCMI Somatoform Scale was found to be associated moderately not only with a variety of the aversive mood/symptom states measured on the POMS, but also with all independent measures of anxiety and all but one measure of depression (POMS Depression-Dejection) considered in this analysis.

The MCMI Hypomania scale was designed as a gauge of labile emotionality, restlessness, overactivity, distractability, and impulsivity. This MCMI symptom scale showed no associations of moderate or stronger strength with the mood/symptom states measured by POMS scales. Millon (1 983), however, reports moderate associations be- tween the MCMI Hypomania scale and MMPI Wiggins Hypomania (.67) and MMPI Basic Manic (.66).

The MCMI Dysthymia scale was constructed to tap a pattern that included behavioral apathy, feelings of discouragement, self-deprecatory cognitions, and guilt. This MCMI scale showed moderate correlations with the Tension-Anxiety, Depression- Dejection, Fatigue-Inertia, and Confusion-Bewilderment scales of the POMS. Millon (1983) reports moderate correlations between the MCMI Dysthymia scale and MMPI Wiggins Depression (.74), MMPI Basic Depression (.70), SCL-90 Depression (.69), SCL-90 Anxiety (.48), and MMPI Wiggins Phobias (.41). Thus, the MCMI Dysthymia scale is associated with a variety of symptom/mood states as measured by three separate instruments. Although the MCMI Dysthymia scale did not discriminate between the measures of anxiety and depression on the POMS, this MCMI scale consistently was associated more highly with the other independent measures of depression than with other measures of anxiety reported in the MCMI manual (Millon, 1983).

The MCMI Alcohol Abuse scale was designed to measure a pattern that included a probable history of alcoholism with distress and disruptions in social and occupational functioning. Although this MCMI scale was not associated moderately with any of the POMS scales in this analysis, it was associated moderately with a variety of MMPI symp- tom/mood scales, including MMPI Basic Manic (.39), MMPI Basic Depression (.35), and with the SCL-90 Anxiety scale (.35).

The MCMI Drug Abuse scale was constructed to tap a recurrent or recent pattern of drug abuse with related difficulties with impulse control and social conformity.

MCMI 91 1

Although this MCMI scale did not show moderate or higher correlations with any POMS scales, the two highest associations found were with POMS Anger-Hostility (.27) and POMS Vigor-Activity (.20). Millon (1983) reported moderate associations between the MCMI Drug Abuse scale and MMPI Basic Manic (.59), MMPI Wiggins Hypomanic (.49), SCL-90 Hostility (.48), and MMPI Wiggins Hostility (.43). Thus, there is evidence of moderate association between the MCMI Drug Abuse scale and each of the inde- pendent scales that measure mania and hostility reported in the MCMI manual (Millon, 1983). Statistically significant associations, which failed to reach the criterion of moderate strength, were found between MCMI Drug Abuse and the POMS Anger-Hostility and Vigor-Activity scales.

The MCMI Psychotic Thinking scale was designed to gauge a pattern that included disorganized, confused, and regressive behavior with associated inappropriate affect, delusions, and hallucinations. This MCMI scale, designed to measure a severe symp- tomology state, reached or closely approached moderate association with POMS Confusion-Bewilderment , Depression-Dejection, Anger-Hostility, Tension-Anxiety, and Fatigue-Inertia. Millon (1983) reports moderate associations between the MCMI Psychotic Thinking scale and MMPI Wiggins Depression (.61), SCL-90 Depression (.56), MMPI Wiggins Phobias (.47), SCL-90 Anxiety (.42), and MMPI Wiggins Hostility (.40). Thus, the MCMI Psychotic Thinking scale was associated positively with a variety of independent measures of symptom/mood states on the POMS, MMPI, and SCL-90.

The MCMI Psychotic Depression Scale was constructed to measure severely de- pressed mood, pervasive pessimism, hopeless resignation, and psychomotor retardation and/or agitation. This MCMI scale showed moderate associations with the Confusion- Bewilderment, Depression-Dejection, Tension-Anxiety, Fatigue-Inertia, and Anger- Hostility scales of the POMS. Millon (1983) reports moderate correlations between the MCMI Psychotic Depression scale and MMPI Wiggins Depression (.83), SCL-90 Depres- sion (.75), MMPI Basic Depression (.63), SCL-90 Anxiety (.63), MMPI Wiggins Phobias (.50), SCL-90 Phobic Anxiety (.37), and MMPI Wiggins Hostility (.39). Thus, the MCMI Psychotic Depression scale consistently was associated moderately to highly with various independent measures of depression, anxiety, and hostility in the clinical samples con- sidered in this analysis.

The last MCMI symptom scale, Psychotic Delusions, was designed to tap a pattern that included interrelated delusions of a persecutory or grandiose nature, periodic belligerence, and irrational behavior. This MCMI scale did not show moderate or higher association with any of the symptom/mood states measured by the POMS. Millon (1983) reports moderate correlations between the MCMI Psychotic Delusions scale and MMPI Wiggins Hostility (.41) and SCL-90 Hostility (.41).

DISCUSSION AND CONCLUSIONS In summary, results of the present investigation reveal a number of significant

associations between MCMI symptom scales and the symptom/mood scales of the POMS. Associations of moderate strength were found between (a) the MCMI Anxiety, Dysthymia, and Psychotic Depression scales and the Tension-Anxiety, Depression- Dejection, Fatigue-Inertia, and Confusion-Bewilderment scales of the POMS; (b) the MCMI Psychotic Thinking scale and the Tension-Anxiety, Depression-Dejection, Anger- Hostility, and Confusion-Bewilderment scales of the POMS; and (c) the MCMI Somatoform scale and the Tension-Anxiety, Fatigue-Inertia and Confusion-Bewilderment scales of the POMS.

Results of the present investigation, when combined with results of the previous analyses reported in the MCMI manual, reveal a number of consistent associations be- tween MCMI symptom scales and selected mood or symptom scales from the MMPI

912 Journal of Clinical Psychology, November 1986, Vol. 42, No. 6

and SCL-90, as well as from the POMS. Consistent associations of moderate strength were found between (a) the MCMI Anxiety, Dysthymia, and Psychotic Depression scales and all available measures of depression and anxiety/phobia; (b) the MCMI Somatoform scale and all available measures of anxiety/phobia; and (c) the MCMI Psychotic Thinking scale and two of three available measures of both depression and anxiety/phobia.

Most relationships between the MCMI symptom scales and the symptom/mood scales of the POMS, MMPI and SCL-90 are consistent with expectation (Millon, 1984). However, a number of exceptions are worthy of note. The MCMI Anxiety scale does not discriminate appropriately between measures of anxiety and depression in any of the samples considered in this analysis. Similarly, the MCMI Dysthymia and Psychotic Depression scales did not discriminate between anxiety and depression as measured on the POMS in this study. However, these two MCMI depression scales were associated more strongly with measures of depression than with measures of anxiety in the con- current validity analyses reported in the MCMI manual (Millon, 1983).

REFERENCES CRAIG, R., VERINIS, J. S., & WEXLER, S. (1985). Personality characteristics of drug addicts and alcoholics

on the Millon Clinical Multiaxial Inventory. Journal of Personality Assessment, 49, 156-160. DEROGATIS, L., LIPMAN, R., & COVI, L. (1973). The SCL-90: An outpatient psychiatric rating scale.

Psychopharmacology Bulletin, 9, 13-28. FLYNN, P., & MCMAHON, R. (1984). An examination of the factor structure of the Millon Clinical Multiax-

ial Inventory. Journal of Personality Assessment. 48, 308-31 1. MCMAHON, R., & DAVIDSDN, R. (1985). An examination of the relationship between personality characteristics

and symptomhood patterns. Journal of Personality Assessment, 49, 552-557. MCMAHON, R., & DAVIDSON, R. (1986) An examination of depressed versus non-depressed alcoholics in

inpatient treatment. Journal of Clinical Psychology, 42, 177-184. MCMAHON, R., FLYNN, P., & DAVIDSON, R. (1985). Stability of the personality and symptom scales of the

Millon Clinical Multiaxial Inventory. Journal of Personality Assessment, 49, 231-234. MCNAIR, D. M., LORR, M., & DROPPLEMAN, L. F. (1971). Profile of Mood States, Manual. San Diego,

CA: Educational and Industrial Testing Service. MILLON, T. (1977). Millon ClinicalMultiaxial Inventory (1st ed.). Minneapolis: National Computer Systems. MILLON, T. (1983). Millon Clinical Multiuxial Znventory (3rd ed.). Minneapolis: National Computer Systems. MILLON, T. (1984). On the renaissance of personality assessment and personality theory. Journal of Per-

sonality Assessment, 48. 450-466. NIE, N. H., HULL, C. H., JENKINS, J. G . , STEMBRENNER, K., & BENT, D. H. (1975). Statisticalpackage

for the social sciences (2nd ed.). New York: McGraw-Hill. ROBERT, J., RYAN, J., MCENTYRE, W., MCFARLAND, R., LIPS, 0.. & ROSENBERG, S. (1985). MCMI

characteristics of DSM-I11 posttraumatic stress disorder in Vietnam veterans. Journal of Personality Asserr- ment. 49, 226-230.