Bond UniversityePublications@bond

Humanities & Social Sciences papers Faculty of Humanities and Social Sciences

5-11-2000

Discriminant validity of the Illness BehaviorQuestionnaire and Million Clinical MultiaxialInventory-III in a heteregenous sample ofpsychiatric outpatientsGregory J. BoyleBond University, Gregory_Boyle@bond.edu.au

Loick Le DéanBond University

Follow this and additional works at: http://epublications.bond.edu.au/hss_pubs

Part of the Psychology Commons

This Journal Article is brought to you by the Faculty of Humanities and Social Sciences at ePublications@bond. It has been accepted for inclusion inHumanities & Social Sciences papers by an authorized administrator of ePublications@bond. For more information, please contact Bond University'sRepository Coordinator.

Recommended CitationGregory J. Boyle and Loick Le Déan. (2000) "Discriminant validity of the Illness BehaviorQuestionnaire and Million Clinical Multiaxial Inventory-III in a heteregenous sample of psychiatricoutpatients" Journal of Clinical Psychology, 56 (6), 779-791: ISSN 0021-9762.

http://epublications.bond.edu.au/hss_pubs/588

Discriminant Validity of the Illness Behavior Questionnaire and Millon Clinical Multiaxial

Inventory-Ill in a Heterogeneous Sample of Psychiatric Outpatients

Gregory J. Boyle

Department of Psychology, Bond University

and

Department of Psychiatry, University of Queensland

Australia

and

Loick Le Dean

University of Aston, United Kingdom

___________________________________________________________________________

_ Correspondence concerning this article should be addressed to Gregory J. Boyle, Ph.D.,

Department of Psychology, Bond University, Gold Coast, Queensland, 4229, Australia.

Abstract

The discriminant validity of measures of abnormal illness behaviours and

psychopathology was examined in three samples differing in illness prone- ness: a sample

of young healthy university students. (N = 38), a general community sample (n = 36),

and a sample of clinical psychiatric outpatients (N = 36). Adjustment to illness was

measured using the Illness Behaviour Questionnaire (lBO: Pilowsky & Spence. 1994),

while the Millon Clinical Multiaxial Inventory-Ill (MCMI-III, Millon. 1994) was used to

measure clinical syndromes and personality. MANCOVAs were performed across the

three groups on the lBO and the MCMI-III categories separately. As expected clinical

outpatients obtained significantly higher scores than did nonclinical groups on most of the

lBQ scales, suggesting discernible discriminant validity. However, the lack of discrimination

between groups on several of the MCMI-III scales raises questions about the test validity

of this multidimensional instrument.

Introduction

The test validity of measures of clinical symptoms and personality remains

problematic for both clinicians and personality researchers alike (see Carson, 1991;

Clark, Watson, & Reynolds, 1995; Nelson Gray, 1991; Sher & Trull, 1996).

Perception of changes in moods associated with an illness may play an important

role in implementing adaptive behaviours (Turk & Salovey, 1995). Also,

individuals judge the severity of symptoms against prior experiences and stored

schemata of an illness (Leventhal, 1983; Turk, Holz- man, & Kerns, 1986). Illness

behaviours refer to the ways in which individuals perceive and report symptoms

associated with their conditions. Such a concept has often been used in clinical

practice to distinguish between individuals suffering from a physical ailment and

those whose somatic complaints are mainly attributable to psychological factors

(Pilowsky & Spence, 1994; Turk & Salovey, 1995).

Illness Behaviour Questionnaire

Pilowsky and Spence (1994) designed the illness Behaviour Questionnaire

(IBQ), a self-report inventory that assesses an individual's ideas, affects, and

attributions of clinical symptoms. The IBQ is a 62-item questionnaire with a

dichotomous yes/no response scale that yields scores on seven subscales

measuring an individual's attitudes toward illness. The IBQ is primarily used to

detect abnormal illness behaviours and to identify physical complaints that are

manifestations of psychosomatic disorders. Examples of IBQ items are "Do you

ever think that you have an illness which is punishment for something you have

done wrong in the past?" and "Do you get the feeling that people are not taking

your illness seriously enough when you are sick?" Clinically, the IBQ to some

extent distinguishes between patients with a physical syndrome and those who

have a largely psychosomatic component to their illness behaviours. Several of the

items pertain to psychological attributes, rather than to physical illness behaviours

per se (e.g., "Do you worry or fuss over small details that seem unimportant to

others?" and "Do you often find that you get depressed?"

The IBQ was derived from factor-analytic studies of chronic pain patients (Pilowsky

& Spence, 1994). Recent factor analyses have provided some support for the construct

validity of the seven subscales in the instrument (Pilowsky & Spence, 1994; Zonderman,

Heft, & Costa, 1985). Although some studies have failed to provide unequivocal support

(Main & Waddell, 1987; Stretton, Salovey, & Mayer, 1992), other studies have established

the concurrent validity of the IBQ with various measures of anxiety and depression (Grassi,

Rosti, Albierti, & Marangolo, 1989; Harkins, Price, & Braith, 1989; Pilowsky & Spence,

1994). Pilowsky and Spence reported moderate to high stability coefficients over a 1 to 12

week retest period ranging from .87 to .67, with a mean coefficient of .80. The test manual

also provides correlations between patients' scores and others' perceptions of patients'

responses ranging from .50 and .78 (mean=.63).

Patients suffering from physical complaints without demonstrable organic

pathology (viz., psychiatric outpatients) usually exhibit significantly higher scores on the

IBQ scales labelled Disease Conviction, Affective Disturbance (involving both anxiety

and depression), and Irritability than do patients with somatic complaints where organic

pathology is established (Pilowsky, Smith, & Katsikitis, 1987). More recent studies

(Pilowsky & Katsikitis, 1994; Pilowsky & Spence, 1994) have shown that psychiatric

outpatients also tend to exhibit higher levels on General Hypochondriasis and Affective

Inhibition, than do patients with somatic disorders.

Millon Clinical Multiaxial Inventory-III

The Millon Clinical Multiaxial Inventory-III (MCMI-III; Millon, 1994) is a

175-true/false-item inventory that yields scores on 11 personality patterns, three

severe personality patterns, seven clinical syndromes, and three severe clinical

syndromes. Clinical symptoms are assumed to be transient maladaptive states of the

individual's basic personality pattern that emerge under perceived stressful situations.

Several factor analyses have provided limited support for the purported MCMI

structure (Choca, Shanley, & van Denburg, 1993; Craig & Weinberg, 1993; Lorr,

Retzlaff, & Tarr, 1989; Lorr, 1993; Lorr, Strack, Campbell, & Lamnin, 1990; Retzlaff,

Lorr, Hyer, & Ofman, 1991). However, the claimed factor structure of the instrument

may in part be an artifact produced by item overlap. While Millon asserted that item

overlap is consistent with the clinical model of psychopathology, overlapping items

cause scales to be linearly dependent (Gibertini & Retzlaff, 1988; Lumsden, 1988;

Retzlaff & Gibertini, 1987). Indeed, in clinical settings, Cantrell and Dana (1987), and

Piersma (1986) found only limited support for the use of the instrument as a tool in

the screening of psychopathology

The MCMI-III manual reports overall test-retest reliability coefficients ranging

from .95 to .82, with an average of .90 over a 5 to 14 day retest interval. However, no

long term stability coefficients are provided. The manual reports Cronbach alphas ranging

from .95 (Debasement scale) to .66 (Compulsive scale), with a mean of .83: No fewer

than 20 of the scales have alphas exceeding .80, suggesting the possibility of some

item redundancy in the MCMI-ill subsca1es (see Boyle, 1991; Clark & Watson, 1995;

Cortina, 1993).

Studies have generally supported the concurrent validity of the MCMI scales

(e.g., Craig & Olson, 1992; Dyer, 1994; Gabrys et al., 1988; McCann, 1991; Montag

& Comrey, 1987; Morey & Levine, 1988; Schuler, Snibbe, & Buckwalter, 1994;

Terpylak & Schuerger, 1994). Also, there is some evidence of discriminant validity

(e.g., McMahon, Flynn, & Davidson, 1985; Morey, Blashfield, Webb, & Jewell,

1988).

Using the MCMI, Vollrath, Almes, and Torgersen (1994a) found that

individuals with personality disorders tend to rely excessively on dysfunctional coping

strategies, to discard social support, to use disengagement, and to vent negative

emotions. Similarly, individuals with elevated clinical syndrome scores tend to discard

adaptive emotion- focused coping (e.g., seeking social support), and to use

maladaptive emotion-focused coping strategies such as disengagement (Vollrath et al.,

1994b).

Leaf, Alington, Ellis, DiGiuseppe, and Mass (1992) found that individuals

diagnosed with Schizoid, Avoidant, Dependent, Passive-Aggressive, Schizotypal, and

Borderline personality disorders on the MCMI were characterized by acute clinical

syndromes. Such individuals reported more distress, more negative life events, and

more social problems than individuals without disorders. Leaf et al. corroborated

previous studies in which normal subjects have consistently scored more highly on

the MCMI Narcissistic scale than did psychiatric patients (Dubro, Wetzler, & Kahn,

1988; Reich & Troughton, 1988b; Wetzler, Kahn, Cahn, van Praag, & Asnis, 1990).

As the psychiatric condition improved, the score on the scale increased (Joffe &

Regan, 1988; McMahon, Flynn, & Davidson, 1985; Reich & Troughton, 1988a).

Whyne-Berman and McCann (1995) investigated personality disorders and

defence mechanisms measured by the MCMI. Antisocial personality scores were

significantly related to acting out, Compulsive traits were associated with reaction

formation, and Passive-Aggressive personality scores were linked with displacement.

Paranoid personality tendencies were significantly correlated with projection, while

Dependent personality characteristics suggested reliance on introjection. The MCMI

Dependent scale appears to be important in determining psychological adjustment, as

individuals with elevated Dependent personality were found to have been repeatedly

hospitalized (Overholser, Kabakoff, & Norman, 1989).

Millon claimed that the MCMI-III is not intended for use with nonclinical

populations, and that "base rate scores" are prefaced on the assumption that the

respondent is a member of the "clinical population," broadly defined. However, in

order to assess the utility of the MCMI-III instrument for use with nonclinical

populations, the present study examines the personality and symptom characteristics

of clinical psychiatric outpatients as compared with a nonclinical group of healthy,

young students, and also with a non-clinical sample from the general adult

community.

Hypotheses

The first hypothesis (H1) proposed that the clinical outpatients would

demonstrate significantly more elevated scores than would either of the nonclinical

groups on the IBQ and MCMI-III scales (even though non-patients may tend to score

more highly than patients on some of the MCMI-III personality scales). It was

assumed that a continuum exists ranging from healthy adjustment to psychological

maladjustment. H2 proposed that on the IBQ and MCMI-III scales, a general adult

community group would perform mid- way between samples of clinical outpatients and

healthy university undergraduates, and that scores would differ significantly between

student and community samples. Since the student sample was young and healthy, it

was expected that they would exhibit lower levels of illness, chronic pain, illness

proneness behaviours, and psychological maladjustment than would older individuals,

especially those from lower socioeconomic and educational backgrounds selected

randomly from the adult community at large.

Method

Participants

All data was collected in accord with the Code of Ethics of the World

Medical Association after approval had been obtained from the Bond University

Ethics Review Committee. Informed consent was obtained from the research

participants after the nature of the testing procedures had been explained to them.

The total sample consisted of 110 participants from the Gold Coast area of

Queensland who had responded to the items in the MCMI-III and IBQ

questionnaires. The overall age of respondents (50 men; 60 women) ranged from 19

to 72 years (M = 33.12 years, SD = 12.95 years).

In order to test the utility of the IBQ and MCMI-III instruments, three

approximately equal-sized groups differing on a continuum of illness-proneness

behaviours and psychological maladjustment were administered both instruments.

Individuals comprising the two nonclinical groups and the clinical outpatients'

group all freely chose to fill out the MCMI-III and IBQ instruments.

A college sample consisted of 38 predominantly young, healthy individuals

obtained by randomly surveying undergraduate students enrolled at Bond

University. In general, this student group which comprised 18 males and 20 females

aged ranged from 19 to 45 years (M = 24.53 years, SD = 6.97 years) was expected

to be quite low on physical illness, chronic pain, illness-proneness behaviours and

psychological maladjustment. Although the MCMI-III instrument was designed

primarily to discriminate between different psychiatric syndromes, the present

comparison between outpatients and nonclinical groups was considered a useful test

of the discriminant validity of the MCMI-III and IBQ instruments.

A general community sample consisted of 36 adults who were recruited by

surveying individuals from the general adult community at large, located in the

Gold Coast region of Queensland. Their participation was mainly solicited through

personal contacts of the investigators, and therefore did not represent a truly

random sample. They were not paid for their participation, nor were they screened

explicitly for psychiatric disorders. This sample comprised a much greater diversity

of individuals from differing socioeconomic and educational backgrounds than was

the case with the undergraduate student sample, whose members generally came

from more educated backgrounds. The general community sample, which consisted

of 13 males and 23 females aged between 19 and 72 years (M = 38.86 years, SD =

15.56 years), because of its overall greater age composition was expected to be

somewhat more prone to physical illness, chronic pain, illness behaviours, and

psychological maladjustment than was the university student sample.

The clinical outpatients group (n = 36) comprised individuals referred to

two Gold Coast psychologists for evaluation primarily because of psychosocial

dysfunction that warranted examination by an appropriately trained mental health

practitioner-e.g., employment, relational problems, and/or DSM-IV diagnosable

psychopathology (predominantly Axis I and II disorders). Most of them suffered

from depressive mood, and/or severe personality disorders. Few of the clinical

outpatients exhibited discernible Axis III medical conditions. The outpatients who

chose to participate in the study did so freely after the nature of the testing

requirements had been explained to them. They were informed prior to accepting

therapy that they would be asked to complete a battery of psychological tests, and

that their responses to the IBQ and MCMI-III instruments would be used for

research purposes as well as for psychotherapy. The clinical outpatients sample

consisted of 19 males and 17 females whose ages ranged from 22 to 60 years (M =

36.44 years, SD = 10.16 years).

Results

The overall alpha coefficient for the IBQ was .62, and for the MCMI-III

subscales alphas were as follows: Personality Patterns (.66), Severe Personality

Patterns (.80), Clinical Syndromes (.84), and Severe Clinical Syndromes (.67).

Evidently, the high alpha coefficients for the Severe Personality Patterns and

Clinical Syndromes may suggest the possibility of some item redundancy in these

scales (cf. Boyle, 1991). Means and standard deviations for the undergraduate

students, general community, and clinical outpatient groups on the IBQ and

MCMI-III scales are presented in Tables 1 and 2 (also Figs. 1 and 2), respectively.

A one-way between-groups ANOVA revealed that age differed significantly

across groups, F(2,107) = 16.92, p < .001. Two-tailed t-test comparisons

(Bonferroni corrected) indicated that the student sample was significantly younger

than both the general adult community sample, t(106) = 5.41, p < .05, and the

clinical sample, t(106) = 4.50, p < .05. In view of the significant main effect, age

was treated as a covariate in the separate between-subjects MANCOVAs (cf.

Huberty & Morris, 1989) performed on the IBQ scales and Modifying Indices,

Personality Patterns, Severe Personality Pathologies, Clinical Syndromes, and

Severe Clinical Syndromes of the MCMI-III. No significant gender differences

were apparent between the three samples.

After adjustment for age differences, the combined scores on both the IBQ and

MCMI-III instruments differed significantly between groups. The multivariate

effects (Bonferroni corrected) were: IBQ, F(14,200) = 6.48, p < .05; MCMI-III:

Modifying Indices, F(6,208) = 8.98, p < .05; Personality Patterns, F(22,192) =

3.38, p < .05; Severe Personality Patterns, F(6,208) = 4.82, p = .05; Clinical

Syndromes, F(l4,200) = 6.83, p < .05; and Severe Clinical Syndromes, F(6,208)

= 12.31, p < .05, respectively. For the IBQ instrument, ANOVAs (Bonferroni

corrected) suggested that Disease Conviction, F(2,106) = 32.45, p < .05; Affective

Disturbance, F(1,106) = 23.72, p < .05; and Irritability, F(2,106) = 23.83, p < .05,

contributed to the overall IBQ main effect. Clinical outpatients scored more highly

than both college students and individuals from the community at large on Disease

Conviction (M = 51.81, M = 15.55, and M = 19.22,

Table 1

IBQ Group Means and Standard Deviations

Students

Community

Outpatients

Scale

Mean

(SD)

Mean

(SD)

Mean

(SD)

General Hypochondriasis

15.34

(14.59)

15.94

(19.37)

28.83

(25.74)

Disease Conviction 15.55 (12.45) 19.22 (18.48) 51.81 (30.25)

Psychological vs. Somatic Perception 46.32 (19.79) 44.44 (16.64) 47.78 (20.59)

Affective Inhibition 40.00 (32.22) 49.44 (34.64) 62.78 (34.19)

Affective Disturbance (Dysphoria) 28.42 (32.76) 37.78 (32.70) 76.11 (26.97)

Denial 45.26 (26.58) 40.47 (32.65) 47.78 (30.34)

Irritability 33.79 (20.59) 32.50 (26.08) 61.97 (31.71)

Table 2

MCMI-Ill Group Means and Standard Deviations

Students (N = 38) Community (N = 36) Outpatients (N = 36)

Scale Mean (SD) Mean (SD) Mean (SD)

Modifying Indices

Disclosure 41.42 (17.51) 53.97 (19.50) 61.11 (19.20)

Desirability 74.50 (13.09) 65.72 (17.65) 58.89 (16.15)

Debasement 31.53 (24.04) 48.50 (23.63) 67.75 (17.78)

Personality Patterns

Schizoid 33.55 (23.46) 48.50 (23.41) 60.31 (22.85)

Avoidant 29.89 (26.04) 42.86 (30.24) 57.78 (24.85)

Depressive 21.58 (25.32) 37.31 (31.12) 57.92 (25.70)

Dependent 27.29 (23.06) 46.00 (29.28) 63.31 (26.70)

Histrionic 74.79 (18.78) 6o.64 (24.27) 47.75 (22.57)

Narcissistic 81.29 (20.29) 65.69 (20.60) 54.28 (19.49)

Antisocial 41.37 (21.23) 52.11 (22.53) 47.64 (23.77)

Aggressive (Sadistic) 44.16 (22.49) 51.53 (24.56) 51.00 (23.95)

Compulsive 60.18 (15.67) 55.31 (19.29) 52.56 (18.95)

Passive-Aggressive (Negativistic) 32.61 (22.02) 52.47 (27.08) 56.75 (28.11)

Self-Defeating 27.16 (29.92) 35.58 (32.55) 47.97 (32.12)

Severe Personality Patterns

Schizotypal 29.61 (28.12) 41.94 (27.33) 52.53 (24.74)

Borderline 27.08 (24.96) 41.17 (28.33) 54.81 (25.76)

Paranoid 35.18 (26.77) 49.44 (28.00) 51.67 (25.25)

Clinical Syndromes

Anxiety 34.68 (28.96) 48.42 (30.62) 79.03 (23.76)

Somatoform 16.89 (21.41) 32.75 (29.47) 61.69 (26.98)

Bipolar: Manic 50.61 (23.33) 53.72 (20.29) 55.03 (20.92)

Dysthymia 13.61 (20.79) 32.47 (33.51) 65.53 (24.62)

Alcohol Dependence 36.47 (29.40) 47.64 (27.82) 44.42 (30.27)

Drug Dependence 45.87 (21.50) 50.03 (25.29) 43.83 (24.48)

Post-Traumatic Stress Disorder 21.79 (22.74) 29.81 (26.34) 62.67 (15.92)

Severe Clinical Syndromes

Thought Disorder 28.97 (24.19) 37.97 (28.07) 59.81 (22.07)

Major Depression 12.71 (18.60) 30.06 (31.91) 64.14 (27.92)

Delusional Disorder 27.66 (29.25) 38.00 (29.92) 31.92 (29.42)

50

0

100

90

80

70

"' 60

--Students

_ Community Sample

40

30

20

10

0

GH DC PS AI AD D

IBQ Scales

--- Outpatients

Fig. I. Mean scale profiles on the IBQ. OH: General Hypochondriasis; DC:

Disease Conviction; PS: Psychological versus Somatic Perception; AI: Affective

Inhibition; AD: Affective Disturbance (Dysphoria); D: Denial; I: Irritability.

120

110

100

90

" 80

70 )

---< -- Students

- Community Sample

--Outpatients

60

50

13

MCMI-Scales

Fig. 2. Mean scale profiles on the MCMI-III. X: Disclosure; Y: Desirability; Z: Debasement; 1:

Schizoid; 2A: Avoidant; 2B: Depressive; 3: Dependent; 4: Histrionic; 5: Narcissistic; 6A:

Antisocial; 6B: Aggressive (Sadistic); 7: Compulsive; 8A: Passive-Aggressive (Negativistic); 8B:

Self-Defeating; S: Schizotypal; C: Borderline; P: Paranoid; A: Anxiety; H: Somatoform; N: Bipolar

(Manic); D: Dysthymia; B: Alcohol Dependence; T: Drug Dependence; R: Post-Traumatic Stress

Disorder; SS: Thought Disorder; CC: Major Depression; PP: Delusional Disorder.

respectively); Affective Disturbance (M = 76.11, M = 28.42, and M =

37.78,(respectively); and Irritability (M = 61.97, M = 33.79, and M = 26.08,

respectively).

For the MCMI-Ill instrument, the variable that contributed most to the

Modifying Indices main effect was Debasement, F(2,106) = 27.81, p < .05. Clinical

patients (M =67.75) scored more highly on Debasement than both students (M =

31.53) and community individuals (M = 48.50). The univariate main effect for

Disclosure was F(2,106) = 16.46, p < .05, and for Desirability F(2, 106) = 8.25, p <

.05. Clinical outpatients scored more highly than did students on Disclosure (M =

61.11 and M = 41.42), and lower on Desirability (M = 58.89 and M = 74.50). In

contrast, community individuals scored more highly than students on Debasement and

on Disclosure (M = 53.97), and lower on Desirability (M = 65.72).

For the MCMI-III Personality Patterns significant (Bonferroni corrected) main

effects occurred for the Schizoid, F(2,106) = 12.64, p < .05; Avoidant, F(2,106) =

10.79, p <.05; Depressive, F(2,106) = 16.31, p < .05; Dependent, F(2,106) = 16.05,

p < .05; Histrionic, F(2,106) = 14.14, p < .05; Narcissistic, F(2,106) = 14.59, p <

.05; and Passive-Aggressive, F(2, 106) = 13.71, p < .05, personality scales. Simple

contrasts showed that clinical outpatients scored significantly lower than did the

student sample on Histrionic (M = 47.75 and M = 74.79), Narcissistic (M = 54.28

and M = 81.29), and Compulsive (M = 52.56 and M = 60.18) personality scales,

and significantly higher on most of the remaining personality scales. Community

individuals scored lower than did college students on Histrionic (M = 60.64) and

Narcissistic (M = 65.69) personality scales, and more highly on the remaining scales.

14

Clinical outpatients scored higher than did individuals from the community at large on

Depressive (M = 57.92 and M = 37.31), Dependent (M = 63.31 and M = 46.00),

Avoidant (M = 57.78 and M = 42.86), and Schizoid (M = 60.31 and M = 48.50)

personality scales, and lower on Histrionic (M = 47.75 and M = 60.64) and

Narcissistic (M = 54.28 and M = 65.69) personality scales, respectively.

Two of the Severe Personality scales of the MCMI-Ill contributed

significantly to group differences (Bonferroni corrected). The Borderline personality

scale appeared particularly important, F(2, 106) = 12.63, p < .05, as did the

Schizotypal scale, F(2,106) = 8.82, p < .05. Simple contrasts showed that college

students scored significantly lower than both clinical outpatients and community

individuals on all three scales. Community individuals scored lower than clinical

outpatients on the Borderline personality scale only (M = 41.17 and M = 54.81).

Four out of the seven Clinical Syndromes scales contributed significantly to the

MCMI-III main effect. The univariate effects (Bonferroni corrected) were: Anxiety,

F(2, 106) = 26.43, p < .05; Somatoform Complaints, F(2,106) = 28.67, p < .05;

Dysthymia, F(2,106) = 37.56, p < .05; and Post-Traumatic Stress Disorder, F(2,

106) =

37.56, p < .05, respectively. Clinical outpatients obtained significantly higher scores

than did community individuals and college students on Dysthymia (M = 65.53, M =

37.47, and M = 13.61); Post-Traumatic Stress Disorder (M = 62.67, M = 29.81, and

M = 21.79); Somatoform Complaints (M = 61.69, M = 32.75, and M == 16.89); and

Anxiety (M = 79.03, M = 48.42, and M = 34.68). Individuals from the community at

large scored significantly higher than did students on Somatoform, Bipolar, and

Dysthymia scales.

The significant MCMI-III Severe Clinical Syndrome main effect was mainly

due to differences in Major Depression, F(2,106) = 36.97, p < .05, and Thought

Disorder, F(2,106) = 17.19, p < .05. Clinical outpatients scored more highly than

15

both community members and college students on Major Depression (M = 64.14, M =

30.06, and M = 12.71) and Thought Disorder (M = 59.81, M = 37.97, and M =

28.97).

Discussion

The three groups obtained scores that differed markedly on several of the IBQ

and MCMI- III scales, lending partial support to Hl. Clinical outpatients reported

significantly more elevated scores on MCMI-III scales labelled Debasement,

Schizoid, Avoidant, Depressive, Dependent, Self-Defeating, Schizotypal,

Borderline, Anxiety, Somatoform, Dysthymia, Post-Traumatic Stress Disorder,

Thought Disorder, and Major Depression, than did both nonclinical samples,

suggesting greater emotional and personal difficulties, as would be expected.

Similarly, clinical patients exhibited elevated scores on Disease Conviction,

Affective Disturbance (Dysphoria), Irritability, General Hypochondriasis, and

Affective Inhibition on the IBQ. Apparently, such individuals are convinced that

they suffer from an illness; they exhibit high levels of anxiety, depression, tension,

and sadness; and they report resentment and interpersonal difficulties (cf.

Pilowsky et al., 1987; Pilowsky & Spence, 1994). They also tend to display

exaggerated health concerns, being overly preoccupied with physical symptoms, as

well as difficulties in expressing emotions (Pilowsky & Katsikitis, 1994). Likewise,

Vollrath et al. (1994b) reported that clinical outpatients have less emotional support,

and tend to disengage more from goals, and to focus on and vent more emotions than

nonclinical samples.

The Depressive personality scale of the MCMI-III was sensitive in

differentiating between clinical and nonclinical individuals. Similarly, the

Dependent personality scale was a critical indicator of psychological

maladjustment. Earlier studies have found the Dependent personality scale to be

16

sensitive and specific in the assessment of dependency (Miller, Streiner, &

Parkinson, 1992), and to display good convergent validity (Morey & Levine, 1988;

Torgersen & Almes, 1990), as well as satisfactory concurrent and predictive validity

(Overholser & Freiheit, 1994). The finding that the nonclinical groups obtained

higher scores on narcissistic personality suggests that self-esteem tends to be lower

in a group of mixed psychiatric outpatients than in members of the general adult

population at large.

It is not surprising that these measures would discriminate between outpatient

and nonclinical samples. What is surprising and interesting (and from a practical

point of view raises questions regarding test validity) is the lack of discrimination

between some of the MCMI-III personality scales, clinical syndrome scales, and

the Delusional Disorder scale. In terms of clinical utility, this apparent lack of

discriminant validity may suggest that only some items in these scales can

discriminate adequately, and that some items may be poorly written and may not be

measuring the syndromes they are purported to measure. These questions must

await further research on larger samples than used in the present study, given that item

analyses involving small sample sizes are notoriously unreliable (Boyle, Stankov, &

Cattell, 1995).

Little support for H2 (that the community group would obtain scores midway

between those obtained by the clinical and student sample) was found in relation to

the IBQ subscales. In support of H2 though, virtually all of the MCMI-III

subscales differed significantly between the two nonclinical groups, suggesting

(contrary to Millon's assertion) that this multidimensional instrument may be useful for

assessing nonclinical samples.

However, severe personality pathologies measured by the MCMI-III failed to

discriminate clinical outpatients from the sample of community adults. While in

accord with H2, students differed from the other two groups on each of the scales, only

17

Borderline personality discriminated between the clinical and community adults.

Surprisingly, previous studies have found these scales to have good convergent

validity, and to be quite sensitive to psychological maladjustment (Craig &

Weinberg, 1993; Millon, 1994; Piersma, 1993; Torgersen & Alnres, 1990; Widiger &

Corbitt, 1993). One possibility is that the present study used clinical outpatients rather

than inpatients who might have displayed more severe maladaptive features. However,

the proposition by Millon and Davis (1996) and others (e.g., Clarke, Minas, & Stuart,

1991; .Mayou & Hawton, 1986) that psychiatric morbidity in the community may be

more widespread than expected cannot be discounted.

Only half of the personality patterns and none of the severe personality patterns

differed significantly between the community and clinical groups, while half of the

clinical syndromes and two of the three severe clinical syndromes exhibited

corresponding significant differences. Consequently, the differential sensitivity of the

personality and symptoms subscales of the MCMI-III instrument may be deficient.

Summary and Conclusions

The profile of clinical outpatients on the IBQ was consistent with previous

studies which found that psychiatric patients tend to report more abnormal illness

behaviour than non-clinical groups. The IBQ discriminated efficiently between the

clinical sample and non-clinical samples. The scales labelled Disease Conviction,

Affective Disturbance, and Irritability appeared very sensitive to the report of

abnormal illness behaviour by clinical outpatients. General Hypochondriasis and

Affective Inhibition also appeared to play a role, although to a lesser extent, in

discriminating clinical from nonclinical individuals.

The profile of clinical outpatients on the MCMI-Ill also seemed fairly

consistent with earlier findings. The Debasement modifying index, the Anxiety,

Somatoform, Dysthymia, and Post-Traumatic Stress Disorder clinical-syndrome scales, as

18

well as the Thought Disorder and Major Depression severe-clinical-syndrome scales

appeared to be particularly sensitive to psychological maladjustment.

Although the present study found limited support for discriminant validity, the

utility of the MCMI-III personality scales remains to be further explored. For the

severe personality scales, only the Borderline scale differentiated clinical outpatients

from either of the nonclinical groups. Similarly, only half of the personality patterns

(and none of the severe personality patterns) discriminated between community adults

and clinical outpatients. Since the present results provide only partial support for the

MCMI-III clinical syndrome scales in the screening of psychopathology, the

discriminative validity of the instrument necessarily remains somewhat uncertain.

Consequently, the MCMI-III in its present form should be regarded more as a

research instrument which has only limited utility for clinical practice and diagnostic

considerations.

References

Boyle, G.J. (1991). Does item homogeneity indicate internal consistency or item

redundancy in psychometric scales? Personality and Individual Differences, 11,

291-294.

Boyle, G.J., Stankov, L., & Cattell, R.B. (1995). Measurement and statistical models in

the study of personality and intelligence. In D.H. Saklofske & M: Zeidner (Eds.),

International hand- book of personality and intelligence. New York: Plenum.

Cantrell, J.D., & Dana, R.H. (1987). Use of the Millon Clinical Multiaxial Inventory

(MCMI) as a screening instrument at a community mental health center. Journal

of Clinical Psychology, 43, 366_:375.

Carson, R.C. (1991). Dilemmas in the pathway of the DSM-IV. Journal of Abnormal

Psychology, 100, 302-307.

19

Chick, D., Martin, S.K., Nevels, R., & Cotton, C.R. (1994). Relationship between

personality disorders and clinical symptoms in psychiatric inpatients as measured

by the Millon Clinical Multiaxial Inventory. Psychological Reports, 74, 331-336.

Choca, J.P., Shanley, L.A., & VanDenburg, E. (1993). Interpretative guide to the

Millon Clinical Multiaxial Inventory. Washington, DC: APA.

Clark, L.A., & Watson, D. (1995). Constructing validity: Basic issues in objective scale

development. Psychological Assessment, 7, 309-319.

Clark, L.A., Watson, D., & Reynolds, S. (1995). Diagnosis and classification of

psychopathology: Challenges to the current system and future directions. Annual

Review of Psychology, 46, 121-153.

Clarke, D.M., Minas, I.H., & Stuart, G.W. (1991). The prevalence of psychiatric

morbidity in general hospital inpatients. Australian and New Zealand Journal of

Psychiatry, 25, 322-329.

Cortina, J.M. (1993). What is coefficient alpha? An examination of theory and

applications. Journal of Applied Psychology, 78, 98-104.

Craig, R.J., & Olson, R.E. (1992). Relationship between MCMI-ll scales and normal

personality traits. Psychological Reports, 71, 699-705.

Craig, R.J., & Weinberg, D. (1993). MCMI: Review of the literature. In R. J. Craig

(Ed.), The Millon ·clinical Multiaxial Inventory: A clinical research information

synthesis (pp. 23-70). Hillsdale, NJ: Erlbaum.

Dubro, A., Wetzler, S., & Kahn, M.A. (1988). A comparison of three tests for Axis II

diagnosis. Journal of Personality Disorders, 2, 256-266.

Dyer, F.J. (1994). Factorial trait variance and response bias in MCMI-ll personality

disorder scale scores. Journal of Personality Disorders, 8, 121-130.

Gabrys, J.B., Utendale, K.A., Schumph, D., Phillips, N., Peters, K., Robertson, G.,

Sherwood, G., O'Haire, T., Allard, I., Clark, M., & Laye, R.C. (1988). Two

20

inventories for the measurement of psychopathology: Dimensions and common

factorial space on Millon's clinical and Eysenck's general personality scales.

Psychological Reports, 62, 591-:601.

Gibertini, M., & Retzlaff, P.D. (1988). Factor invariance of the Millon Clinical

Multiaxia1 Inventory. Journal of Psychopathology and Behavioral Assessment, 10,

65-74.

Grassi, L., Rosti, G., Albierti, G., & Marangolo, M. (1989). Depression and abnormal

illness behaviour in cancer patients. General Hospital Psychiatry, 11, 404-411.

Harkins, S.W., Price, D.D., Braith, J. (1989). Effects of extraversion and neuroticism

on experimental pain, clinical pain, and illness behavior. Pain, 36, 209-218.

Huberty, C.J., & Morris, J.D. (1989). Multivariate analysis versus multiple univariate

analysis. Psychological Bulletin, 105, 302-308. ·

Joffe, R., & Regan, J. (1988). Personality and depression. Journal of Psychiatric

Research, 22, 279-286.

Leaf, R.C., Alington, D.E., Ellis, A., DiGiuseppe, R., & Mass, R. (1992). Personality

disorders, underlying problems, and clinical syndromes. Journal of Personality

Disorders, 6, 134-152. Leventhal, H. (1983). Behavioral medicine: Psychology in

health care. In D. Mechanic (Ed.),

Handbook of health, health care, and the health professions (pp. 709-743). New York:

Free Press.

Lorr, M. (1993). Dimensional structure of the Millon Clinical Multiaxial Inventory. In

R.J. Craig (Ed.), The Millon Clinical Multiaxial Inventory: A clinical research

information synthesis (pp. 81-89). Hillsdale, NJ: Erlbaum.

Lorr, M., Retzlaff, P.D., & Tarr, H.C. (1989). An analysis of the MCMI-I at the item

level. Journal of Clinical Psychology, 45, 884-890.

21

Lorr, M., Strack, S., Campbell, L., & Larnnin, A. (1990). Personality and symptom

dimensions of the MCMI-II: An item factor analysis. Journal of Clinical

Psychology, 46, 749-754.

Lumsden, E.A. (1988). The impact of shared items on the internal-structural validity of

the MCMI. Educational and Psychological Measurement, 48, 669-678.

Main, C.J., & Waddell, G. (1987). Psychometric construction and validity of the

Pilowsky Illness Behaviour Questionnaire in British patients with chronic low

back pain. Pain, 28, 13-25.

Mayou, R., & Hawton, K. (1986). Psychiatric disorder in the general hospital. British

Journal of Psychiatry, 149, 172-190.

McCann, J.T. (1991). Convergent and discriminant validity of the MCMI-11 and MMPI

personality disorder scales. Personality Assessment, 3, 9-18.

McMahon, R.C., Flynn, P.M., & Davidson, R.S. (1985). Stability of the personality

and symptoms scales of the Millon Clinical Multiaxial Inventory. Journal of

Personality Assessment, 49, 231-234.

Miller, H.R., Streiner, D.L., & Parkinson, A. (1992). Maximum likelihood estimates of

the ability of the MMPI and MCMI personality disorder scales and the SIDP to

identify personality disorders. Journal of Personality Assessment, 59, 1-13.

Millon, T. (1994). Millon Clinical Multiaxial Inventory (3rd ed.). Minneapolis, MN:

National Computer Systems.

Millon, T., & Davis, R.D. (1996). An evolutionary theory of personality disorders. In

J.F. Clarkin & M.F. Lenzenweger (Eds.), Major theories of personality disorder

(pp. 221-346). New York: Guilford.

Montag, I., & Cornrey, A.L. (1987). Millon MCMI scales factor analyzed and

correlated with MMPI and CPS scales. Multivariate Behavioral Research, 22, 401-

413.

22

Morey, L.C., Blashfield, R.K., Webb, W.W., & Jewell, J. (1988). MMPI scales for

DSM-ill personality disorders: A preliminary validation study. Journal of Clinical

Psychology, 44, 47-50. Morey, L.C., & Levine, D.J. (1988). A multitrait-

multimethod examination of Minnesota Multiphasic Personality Inventory

(MMPI) and Millon Clinical Multiaxial Inventory (MCMI). Journal of

Psychopathology and Behavioral Assessment, 10, 333-344.

Nelson-Gray, R.O. (1991). DSM-IV: Empirical guidelines from psychometrics. Journal

of Abnormal Psychology, 100, 308-315.

Overholser, J.C., & Freiheit, S.R. (1994). Assessment of interpersonal dependency

using the Millon Clinical Multiaxial Inventory-IT (MCMI-11) and the Depressive

Experience Questionnaire. Personality and Individual Differences, 17, 71-78.

Overholser, J.C., Kabakoff, R., & Norman, W.H. (1989). The assessment of

personality characteristics in depressed and dependent psychiatric inpatients.

Journal of Personality Assessment, 53, 40-50.

Piersma, H.L. (1986). The Millon Clinical Multiaxial Inventory (MCMI) as a

treatment outcome measure for psychiatric inpatients. Journal of Clinical

Psychology, 42, 493-499.

Pilowsky, I., & Katsikitis, M. (1994). A classification of illness behavior in pain clinic

patients. Pain, 57, 91-94.

Pilowsky, I., Smith, Q., & Katsikitis, M. (1987). Illness behavior and general practice

utilization: A prospective study. Journal of Psychosomatic Research, 37, 177-183.

Pilowsky, I., & Spence, N. (1994). Manual for the Illness Behaviour Questionnaire (3rd

ed.). Department of Psychiatry, University of Adelaide.

Reich, J., & Troughton, E. (1988a). Comparison of DSM-ill personality disorders in

recovered depressed and panic disorder patients. Journal of Nervous and Mental

Disease, 176, 300-304.

23

Reich, J., & Troughton, E. (1988b). Frequency of DSM-ill personality disorders in

patients with panic disorder: Comparison with psychiatric and normal control

subjects. Psychiatry Research, 26, 89-100.

Retzlaff, P.D., & Gibertini, M. (1987). Factor structure of the MCMI basic personality

scales and common-item artifact. Journal of Personality Assessment, 51, 588-594.

Retzlaff, P.D., Lorr, M., Hyer, L., & Ofman, P. (1991). An MCMI-II item-level

component analysis: Personality and clinical factors. Journal of Personality

Assessment, 57, 323-334. Schuler, C.E., Snibbe, J.R., & Buckwalter, J.G.

(1994). Validity of the MMPI personality disorder scale (MMPI-PD). Journal

of Clinical Psychology, 50, 220-227.

Sher, K.J., & Trull, T.J. (1996). Methodological issues in psychopathology

research. Annual Review of Psychology, 47, 371-400.

Stretton, M.S:, Salovey, P., & Mayer, J.D. (1992). Assessing·health concerns.

Imagination, Cognition and Personality, 12, 115-137.

Terpylak, 0., & Schuerger, J.M. (1994). Broad factor scales of the 16PF Fifth Edition

and Millon personality disorder scales: A replication. Psychological Reports,

74, 124-126.

Torgersen, S., & Almes, R. (1990). The relationship between the MCMI

personality scales and DSM-III, Axis II. Journal of Personality Assessment,

55, 698-707.

Turk, D.C., Holzman, A.D., & Kerns, R.D. (1986). Chronic pain. In K.A. Holroyd

& T.L. Creer (Eds.), Self-management approaches to the prevention and

treatment of chronic disease (pp. 441- 472). San Diego, CA: Academic Press.

Turk, D.C., & Salovey, P. (1995). Cognitive-behavioral treatment of illness

behavior. In P.M. Nicassio & T.W. Smith (Eds.), Managing chronic illness: A

biopsychosocial perspective (pp. 245- 284). Washington, DC: APA.

24

Vollrath, M., Alnres, R., & Torgersen, S. (1994a). Coping and MCMI-II personality

disorders. Journal of Personality Disorders, 8, 53-63.

Vollrath, M., Alnres, R., & Torgersen, S. (1994b). Coping and MCMI-II symptom

scales. Journal of Clinical Psychology, 50, 727-736. .

Wetzler, S., Kahn, R.S., Cahn, W., van Pniag, H.M., & Asnis, G.M. (1990).

Psychological test characteristics of depressed and panic patients. Psychiatry

Research, 31, 179-192.

Whyne Berman, S.M., & McCann, J.T. (1995). Defence mechanisms and

personality disorders: An empirical test of Millon's theory. Journal of

Personality Assessment, 64, 132-144.

Widiger, T.A., & Corbitt, E.M. (1993). The MCMI-II personality disorder scales

and their relationship to DSM-III-R diagnosis. In R.J. Craig (Ed.), The Millon

Clinical Multiaxial Inventory: A clinical research information synthesis (pp.

181-201). Hillsdale, NJ: Erlbaum.

Zonderman, A.B., Heft, M.W., & Costa, P.T., Jr. (1985). Does the Illness Behaviour

Questionnaire measure abnormal illness behaviour? Health Psychology, 4,

425-436.