Full text of DOI:10.1348/135532500167985

Legal and Criminological Psychology (2000), 5, 71–81 Printed in Great Britain
Ó 2000 The British Psychological Society
71
Debasement and faking bad on the Millon
Clinical Multi-axial Inventory III: An
examination of characteristics, circumstances
and motives of forensic patients
Brian A. Thomas-Peter*
Reaside Clinic and School of Psychology, University of Birmingham, UK
Jason Jones
Clinical Psychology Department, University of Birmingham, UK
Stephen Campbell
Caroline Oliver
and
Reaside Clinic, Birmingham, UK
Purpose. This article constitutes the Ž rst e ort to consider debasement employ-
V
ing the Millon Clinical Multi-axial Inventory III (MCMI-III) in a forensic clinical
population and incorporating the recommendations of Van Denburg and Choca
(1997). It considers the proposition (Rogers, 1990a) that the pathogenic and
criminological models of malingering can be rejected in favour of the adaptional
model.
Methods. A total of 142 MCMI-III proŽ les were collected over a 17-month
period by a forensic psychology service. Of these, 47 had Debasement scale scores
greater than 74. From the remaining 75, a group of 47 were randomly selected for
comparative purposes. Of the total sample there were 75 males and 19 females, 68
were out-patients and 26 were in-patients, the latter having been detained under
the Mental Health Act 1983, applicable to England and Wales. The two groups
were compared on personality variables, o ending history, presenting problems
V
and referral circumstances.
Results. There is tentative support for Rogers (1990a) in that situational variables
appear to have a signiŽ cant impact on debasement as would be predicted by the
adaptional model of malingering. However, this was insu cient evidence to reject
Y
the pathogenic and criminological models of malingering.
Conclusions. It is suggested that a longitudinal study examining the impact of
di erent stages of the mental health or criminal justice system on debasement, and
V
other validity scales, would illuminate the adaptional model further. Factors that
in uence the inclination to exaggerate psychopathology and which could be
manipulated experimentally are also noted. Some speculation about the possible
combined use of Debasement and Histrionic, Narcissistic and Compulsive scales
in distinguishing between malingerers is o ered.
V
*Requests for reprints should be addresed to Dr Brian Thomas-Peter, Director of Psychological Services, Reaside
Clinic, Birmingham Great Park, Bristol Road South, Birmingham B45 9BE, UK.

72
Brian A. Thomas-Peter et al.
The use of the Minnesota Multiphasic Personality Inventory (MMPI) to detect
‘faking bad’ has a long history. Schretlen (1988) reviewed 14 studies examining
this subject, dating back to 1946. Schretlen concluded that normal partici-
pants can fake mental abnormality on the MMPI but ‘malingered proŽ les can
be accurately distinguished from non-pathological ones’ (p. 473). However, it
was more diYcult to distinguish faked proŽ les from those o ered by the
V
genuinely mentally disordered. Heilbrun, Bennett, White, and Kelly (1990) built
on this body of MMPI research to develop the MMPI-based Empirical Response
Style (MERS) procedure to detect false proŽ les. A review of assessments of
malingering and defensiveness using the MMPI and other objective personality
inventories is available in Greene (1988). In Greene’s chapter, the use of the Ž rst
Millon Clinical Multi-axial Inventory (MCMI) in the detection of malingering,
which until then had not featured in the literature, is referred to. The validity
scales of the MCMI, and the assessment itself, have since developed to the
extent that their role in the detection of malingering has been subject to
signiŽ cant research, some of which is reviewed below. One of the validity scales,
used to modify the scores of the MCMI-III, is known as debasement (known as
the Z scale). It is this scale and its place in detecting faking bad that is the
particular interest of this article.
In the manual of his most recent of clinical inventories, Millon (1994) deŽ nes
high scores on Debasement as ‘an inclination to deprecate or devalue oneself by
presenting more troublesome emotional and personal diYculties than are likely to
be unconvered upon objective review’ (p. 198). This scale has become most closely
associated with the phenomenon of ‘faking bad’ or ‘crying for help’. Faking bad on
the original form of the MCMI (McNiel & Meyer, 1990; Van Gorp & Meyer, 1986)
and MCMI-II (Bagby, Gillis, & Dickens, 1990; Retzla , Sheehan, & Fiel, 1991) has
V
received some attention in the literature, typically by contriving circumstances
experimentally. For example, Retzla et al. asked 250 student participants to answer
V
the questionnaire with one of Ž ve instructions such as ‘as if you were in a great deal
of distress and wanted to get admitted to hospital’ (p. 468). Other instructions
asked participants to fake good. In addition three computer control groups of 50
proŽ les each were generated. As the MCMI is a true/false questionnaire and
requires predominantly positive endorsement to score on scales, the proportion of
positively endorsed items was varied in the random computer proŽ les.
Faking bad proŽ les were easily distinguished from randomly generated, 95% true
proŽ les, as the former had lower response bias and greater variability. However
random 50% true proŽ les could only be distinguished from fake bad proŽ les with
the help of Desirability and Debasement validity scales. The important Ž nding here
is that fake bad proŽ les can be distinguished from random or response biased
proŽ les.
Two other Ž ndings from this study are also interesting. It was observed that
elevated Histrionic (H), Compulsive (C) and Narcissistic (N) characteristics are
most notable on ‘fake good’ proŽ les, suggesting that there may be correlates from
the personality pattern scales that operate inversely to Debasement. Secondly, 92%
of proŽ les from clinical samples that had been told to fake bad had high
Debasement scores. There remains a question of whether Histrionic, Compulsive

Debasement and faking bad on the MCMI-III
73
and Narcissistic characteristics among this group were abnormally low. In response
to these contrived experimental studies Van Denburg and Choca (1997) have called
for an empirical evaluation of the correlation between Debasement scale scores and
idiosyncratic patient presentations.
One of the problems with identifying those who fake bad is the prevalence of
high Z scores among some populations. In clinical populations it has been
estimated that as many as 17% of those administered the MCMI-II had Debase-
ment scores greater than BR84 (Wetzler & Marlow, 1990), suggesting that either the
prevalence of ‘faking bad’ is high, or it is a generic index of psychological
disturbance.
Among forensic populations Lees-Haley (1992) has noted a signiŽ cant pro-
portion of high Debasement scorers and that elevated scores on this scale readily
distinguish between forensic and non-forensic populations. One possible interpret-
ation is that as forensic populations, who may be more willing to falsify their
responses, have high Z scores, it may be used to identify malingerers in other
populations.
Others have questioned this, arguing that the inclination to score highly on the
debasement scale is a characterological issue, perhaps one that is commonly found
among forensic patients, rather than evidence of malingering (Price & Lees-Haley,
1994). Evidence in support of this comes from an earlier study (Bagby, Gillis,
Toner, & Goldberg, 1991) where debasement and disclosure correlated with
measures of dysthymia and major depression with coeYcients as high as .98,
although this can be interpreted in several ways. It seems likely that there will be
some individuals subject to compulsory detention or forensic scrutiny who would
wish to represent themselves in a very positive light. Di erences between response
V
styles may emerge in the comparison of those who are being required to undergo
psychological evaluation and those who seek evaluation, perhaps as a means to
increase their liberty.
A further problem is that deŽ ning malingering is not as straightforward as it
might seem. DSM-IV (American Psychiatric Association, 1994, p. 683) o ers the
V
following deŽ nition, which gives weight to criminal intent:
. . . intentional production of false or grossly exaggerated physical or psychological symptoms,
motivated by external incentives such as avoiding military duty, avoiding work, obtaining
Ž nancial compensation, evading criminal prosecution or, obtaining drugs. Under some
circumstances, Malingering may represent adaptive behaviour—for example, feigning illness
while a captive of the enemy during war time.
This di ers from the description of debasement o ered by Millon, in that it
V
V
emphasizes the participant’s motive. In doing this, it implies that the circumstances
of an assessment will have a bearing on the participant’s response style or the
propensity to dissimulate. For those who work in forensic psychology, this is a
statement of the obvious.
Millon (1997, p. 166) would appear to concur, although he infers personological
and pathological in uences in those who exaggerate their disturbance:
Perhaps there is no greater challenge in the forensic area than in detecting the true malingerer.
This is always a possibility in a forensic case. There are multiple factors which can be operating
other than the patient’s intent to deceive.

74
Brian A. Thomas-Peter et al.
Pollock (1996) makes this point when reporting that those with acquiescent
response styles who are liable to interrogative suggestibility are readily classiŽ ed as
malingering when examined in a closed interview situation using verbally presented
questions, such as that developed by Rogers, Bagby, and Dickens (1992).
Rogers (1990a, 1990b) o ers three models of malingering to assist one’s thinking
V
about those who fake bad. The pathogenic model implies internal processes
designed to cope with psychic tensions, and is closely associated with what has
become known as factitious disorders. Rogers, Bagby, and Rector (1989) and
Rogers (1990a) have o ered critiques of the pathogenic model which, they argue, is
V
based on tenuous logic and is without empirical support. The criminological model
of malingering re ects the underpinning of DSM-III, -III-R and -IV implied
explanatory processes of greed, dishonesty and deception for personal gain. It
seems likely that no one would dispute that the criminal motivation does exist;
however, Rogers argues that it is insuYcient to explain all types of malingering.
Rogers (1990a) reports Ž ndings that indicate that logical corollaries of this model,
such as the expected prevalence of antisocial personality disorder among malinger-
ing groups, do not stand scrutiny. As a consequence of these limitations, Rogers
o ers the adaptional model, which draws on decision theory to postulate ‘that
V
malingerers perceive their circumstances as both adversarial and risky. In response
to this, would-be malingerers seek the most advantageous choice to maximize their
chances of success . . . Malingering is one of many options. The process is seen as
adaptive . . . although the end result may not always be so’ (p. 327).
This article arises from the work of a forensic psychology service within a
regional secure unit. The authors observed a surprising number of elevated
debasement (Z) scores among in-patients and out-patients and sought to explore
this phenomenon. From a service point of view, they were interested to know the
impact on Z of the purpose of their involvement. From a theoretical point of view
it was thought important to explore debasement among clinical populations
without experimental instructions. The article constitutes the Ž rst e ort to consider
V
debasement in this way, employing the MCMI-III and incorporating the recom-
mendations of Van Denburg and Choca (1997).
It was anticipated that there could be evidence of the adaptational model arising
from the data, in that circumstances of the participants’ contact with the
department would be signiŽ cant in predicting Z.
Method
Sample
This study examined the MCMI-III proŽ les of 94 mentally disordered o enders referred to and
V
assessed by a clinical psychology department in a regional secure unit between May 1996 and
September 1997. The sample comprised 47 proŽ les with MCMI-III Debasement scores of over 74.
These were identiŽed among a total of 142 available MCMI-III proŽ les. In addition, a comparison
group was created by randomly selecting a further 47 proŽ les from the remaining 95, where all scores
on the Debasement scale fell below 75. It was important to determine a clinically meaningful
discrimination between the groups. The di cult judgments to make are rarely those that involve
Y

Debasement and faking bad on the MCMI-III
75
Table 1. Summary of primary psychiatric diagnosis for each group
High
Z group
(>74)
Not abnormal
Z group
Primary psychiatric
diagnosis
(<75)
Schizophrenia/psychosis
Personality disorder
Depressive illness
Othera
7
19
3
15
9
6
8
3
None
10
14
aOther category included diagnoses of PTSD, OCD, anxiety, other neurotic mental disorders, epilepsy, and other organic mental illness.
participants who fall in the upper or lower quartile of any distribution. The question was how are
those who score BR75 and greater di erent from those who score less than BR75, in terms of their
V
characteristics and circumstances?
The total sample comprised 75 males and 19 females. Mean age was 32.58 years (SD = 9.27,
min. = 17.97, max. = 57.19). Of the 94 participants, 68 were out-patients at the time of their
assessment and 26 were in-patients at the regional secure unit, having been detained under the Mental
Health Act 1983.
A formal diagnosis was not always stated within out-patient case records because of the
administrative procedures falling into disuse. Clinicians failing to record the absence of a diagnosis
partly explain these omissions in case notes. To remedy this problem, two raters allocated participants
to one of the broad diagnostic categories indicated in Table 1. Having reviewed these case records 24
out-patient participants did not have a diagnosis and were allocated the status of None. A further 24
omissions in the remainder of the sample were identiŽed. Omissions of all types were equally
distributed in the two groups. The inter-rater reliability of this process was assessed by random
selection of 20% of these participants who were categorized by two raters, which resulted in a mean
percentage agreement of 90% (j = .80), indicating good inter-rater reliability.
Measure
The MCMI-III is a 175 item true–false self-report inventory. Each item contributes to one or more
of 27 scales. The MCMI-III comprises 11 basic personality scales (Schizoid; Avoidant; Depressive;
Dependent; Histrionic; Narcissistic; Antisocial; Aggressive; Compulsive; Passive–Aggressive; Self-
defeating); three severe personality scales (Schizotypal; Borderline; Paranoid); seven clinical syndrome
scales (Anxiety; Somatoform; Bipolar; Manic; Dysthymia; Alcohol Dependence; Drug Dependence;
Post-traumatic Stress Disorder); and three severe syndrome scales (Thought Disorder; Major
Depression; Delusional Disorder). In addition, the MCMI-III incorporates three modifying indices
(Disclosure, Desirability and Debasement scales) which serve to adjust scores on the aforementioned
scales according to the patient’s response style. The MCMI-III is scored by converting raw scores to
base rate (BR) scores, which take into account the prevalence of personality and clinical syndromes
within clinical populations. Based on these data, a BR score of over 74 indicates the presence of
personality or symptom features, whereas a BR score of over 84 suggests that the syndrome is
prominent (Millon, 1994).
Procedure
Having obtained prior permission from the clinicians concerned, the corresponding case notes for
each of the 94 MCMI-III proŽ les were examined. For each individual, additional information
pertaining to diagnosis, referral reason, presenting di culties, index o ence(s), previous conviction(s),
Y
V
and previous therapeutic input was collated.

76
Brian A. Thomas-Peter et al.
Results
The MCMI-III modifying indices and personality scales of the two groups of high
and not abnormal Debasement participants were compared using MANOVA
followed by a series of univariate F tests. Comparing the modifying indices of the
two groups using MANOVA (d.f. = 1,2) yielded a signiŽcant main e ect supported
V
by Wilk’s Lambda .38 ( p< .0000). Comparison of the groups in respect of
combined clinical and severe personality scales indicate a similar Ž nding of
signiŽ cant main e ect supported by Wilk’s Lambda .36 ( p< .0000).
V
The results of the univariate analysis for both comparisons are shown in
Table 2, which shows how the two groups compare in terms of the modifying
indices (Disclosure, Desirability and Debasement), and in terms of the clinical and
severe personality characteristics. From Table 2 it is evident that the high
Debasement group had signiŽ cantly higher scores on Disclosure and Debasement
and signiŽcantly lower scores on the Desirability scale.
In terms of the clinical and severe personality scales (Table 2), the high
Debasement group had signiŽcantly higher scores on all scales except for
Histrionic, Narcissistic and Compulsive, on which the not abnormal Debasement
group had signiŽ cantly higher scores. This latter Ž nding concurs with those of
Retzla et al. (1991), as described in the introductory text above.
V
Table 2.
Di erences (univariate F test) between high and not abnormal Debasement
V
groups on the modifying indices, clinical and severe personality characteristics
High Z group (>74) Not abnormal Z group (<75)
p level
Mean
SD
Mean
SD
F
d.f.=1,92
Disclosure
Desirability
Debasement (z)
85.55
33.36
85.10
9.33
18.07
6.86
58.27
64.46
53.06
16.07
18.49
17.40
2816.27
672.84
2565.36
.0000
.0000
.0000
Schizoid
Avoidant
Depressive
Dependent
Histrionic
Narcissistic
Antisocial
Aggressive
Compulsive
Passive–aggressive
Self defeating
Schizotypal
Borderline
Paranoid
76.25
79.11
87.89
73.76
21.64
36.15
68.25
66.64
26.51
82.98
78.06
78.34
84.32
77.66
17.58
11.14
10.77
15.54
17.44
17.70
15.42
14.75
17.61
10.48
9.78
48.19
46.40
53.17
55.94
50.72
57.38
61.85
53.96
49.55
56.47
51.51
44.45
54.87
50.15
27.67
29.22
30.00
26.81
20.32
19.20
23.35
19.74
19.46
26.87
26.74
28.48
22.91
27.51
677.21
757.38
920.66
823.25
343.28
602.90
1015.84
1125.66
394.85
1098.34
973.46
730.81
1471.61
789.99
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
12.60
9.68
14.67

Debasement and faking bad on the MCMI-III
Groups compared in terms of age and number of o ences
77
Table 3.
V
High Z group (>74)
Not abnormal Z group (<75)
Mean
SD
Mean
SD
t
d.f.
p*
Age
32.79
0.83
2.34
3.17
8.82
1.13
3.21
3.67
32.35
1.23
2.45
3.68
9.85
1.72
4.15
4.74
0.22
21.35
2.14
87
92
92
92
.8259
.1817
.8898
.5609
No. index o ences
No. prev. o ences
Total no. o ences
V
V
V
2.58
*Bonferroni adjustment indicated p value of .002 as acceptable.
Table 4.
Summary of o ence types for each group
V
High
Z group
(>74)
Not abnormal
Z group
(<75)
Type of o ence
N
N
V
Acquisitive
Homicide
Against the person
Sexual
25
1
17
4
18
5
19
8
Destructive
Public order
Miscellaneous
14
8
8
13
10
11
Note. N is the number in each group with at least one oVence of the given oVence type. The oVence categories are not
mutually exclusive: an individual can have committed oVences of more than one type.
The high Debasement group consisted of 37 males and 10 females, and the not
abnormal debasement group of 38 males and 9 females ( 2(1) = .00, p>.05, n.s.)
v
v2
(Pearson with Yates’ correction is used as the statistical measure of signiŽcance).
The continuous variables of age and number of o ences for the two groups were
V
compared using independent samples t tests. The results of these tests are
presented in Table 3. There were no signiŽ cant di erences between the groups
V
either in terms of the number of index o ences or number of previous o ences,
V
V
although the mean for both measures was higher in the not abnormal Debasement
group. Similarly, Table 4 shows the distribution of o ence types in both groups,
V
between whom no signiŽ cant di erences were found.
V
The groups were found to di er in only one diagnostic category (Table 1).
V
Individuals in the high Debasement group are more likely to have a primary
psychiatric diagnosis of Personality Disorder ( 2(1) = 4.12, p< .05).
v
From Table 5, it is evident that individuals who are referred for routine in-patient
clinical assessment tend to fall into the not abnormal Debasement group

78
Brian A. Thomas-Peter et al.
Table 5.
Groups compared in terms of reason for referral
High
Not abnormal
Z group
(<75)
Z group
(>74)
N
Reason for referral
N
Routine admission/assessment
Initial assessment for treatment or risk management
Anger management assessment
Assessment for admission to regional secure unit
Child protection issues
Court assessment
Suggestibility
5
14
16
1
4
5
16
7
3
0
10
10
1
1
Table 6.
Summary of reason for referral
High
Z group
(>74)
N
Not abnormal
Z group
(<75)
Reason for referral
N
Required assessment (imposed)
Desired assessment (patient-driven)
15
31
36
11
2(1) = 5.88, p< .05). In contrast, individuals who are referred for anger
v
(
management assessments tend to have high Debasement scores ( 2(1) = 9.85,
p< .01).
v
A further analysis of presenting diYculties was attempted, but the large number
of categories meant that cell size was too small for meaningful comparisons. Some
potential di erences are indicated though. All individuals referred for child
V
protection issues, sexually deviant behaviour, substance abuse and disturbed
behaviour (total 14) fall into the not abnormal Debasement group. This is also true
of all except one of the individuals presenting as acutely psychotic (total 7). It might
be argued that these participants represent those who have no desire to be seen as
disturbed, in order to secure release or access to their family. Conversely, all but
one individual presenting with self-harming behaviour, and all of those presenting
with depressed mood, disturbing ideation or confusion (total 14) fall into the high
Debasement group.
Table 6 summarizes the data in Table 5 and shows the number of individuals in
each group referred for imposed and patient-driven assessments. It emerged during
the analysis that individuals referred for routine admission assessments, child
protection issues and court assessments could be regarded here as having these

Debasement and faking bad on the MCMI-III
79
assessments imposed upon them, whilst the remainder of the sample might be
classiŽed as participating in patient-driven assessments.
This distinction is not ideal because it infers the nature of motivation on the
basis of circumstances without taking into account the participants’ individual
psychological motivation. Nevertheless it may be the case that those on whom such
assessments are made have a vested interest in participating in order to achieve
their goals, that typically involve access to their family, or discharge from secure
provision. Hence these two broad categories of required and desired assessment
groups would seem to be a potentially useful distinction to make. Table 6 supports
this conclusion by indicating a predominance of desired assessments in the high
debasement group and required assessments in the not abnormal Debasement
group ( 2(1) = 16.43, p<.01). Closer examination of this distinction in a subse-
v
quent study where the motivations of individuals have been more speciŽ cally
ascertained may be useful.
Discussion
This statistical re ection on this clinical sample would appear to be congruent with
Rogers’ (1990a) support of the adaptional model of malingering. Situational
variables (i.e. the circumstances of the assessment) appear to have an impact on Z
as would be predicted by the adaptional model of malingering.
It would seem pertinent to consider a more systematic appraisal of the perceived
functional utility/secondary gain of being assessed in respect of avoiding prison,
discharge from security, access to children/family and so on, in each case. Closely
associated with this is the phenomenology of need. That is, to what extent does a
particular individual value the potential outcomes of being assessed? It is only in
understanding the participant’s attributions about the assessment process and the
meaning of the potential outcomes that one can begin to understand their
motivation to dissimulate. Varying these characteristics would seem to o er one
V
means of exploring the parameters of malingering and one’s capacity to predict,
identify and manage it. It could also be hypothesized that a longitudinal study
examining the changing Z and other validity scales at di ering stages of the mental
V
health or criminal justice system would illuminate the adaptional model further.
Criminal motives have not been speciŽ cally examined in this sample, although
there is no evidence of di erences between the groups in respect of o ending
V
V
behaviour or numbers of o ences committed. There remains a question about the
V
role of pathology in those who ‘fake bad’. Tables 1 and 2 clearly indicate di erences
V
between high debasement participants and those with no signiŽ cant elevation of
Debasement in respect of their diagnosis and personality. It is therefore obvious
that pathology is evident in the di erences between the groups. Several issues are
V
raised by this observation.
One question is to consider whether these diagnostic and personality di erences
V
explain the variation of Z, or whether such variations predict diagnosis and
personality. It was observed that those with a diagnosis of schizophrenia/psychosis
are more than twice as likely to be in the not abnormal Debasement group (15/7).

80
Brian A. Thomas-Peter et al.
The overwhelming number of these schizophrenic patients are also inpatients
(14/15). It may be reasonable to infer that those diagnosed as su ering from
V
schizophrenia/psychosis, being cared for in hospital and who are well enough to
complete psychometric assessments are generally concerned not to be seen as
worse than they are, and hence do not exaggerate their psychopathology. It may be
that some participants of this subgroup are keen to portray themselves in a negative
way as a means of sustaining their dependence on supportive care. This proposition
is supported by a supplementary t test comparing dependent scores of the Ž ve
schizophrenia/psychotic in-patients in the high Debasement group (M = 89.2) with
the 14 similar participants in the not abnormal group (M = 64.3). This reveals
signiŽ cantly higher scores among the Ž ve high Debasement group in-patient
participants (t = 2.32, p = .033). In general terms, for the majority it can be argued
that it is not their diagnosis that predicts Z but their circumstances. However there
appears to be a minority who are likely to display a strength of motivation to meet
psychological needs through representing themselves negatively, o ering some
V
support for the pathogenic model of malingering rejected by Rogers (1990b).
A second and related issue is to distinguish between those who fake bad as a
result of an adaptive motive to do so, and those who characteristically express
themselves and describe their experience in an exaggerated form. One possible way
of distinguishing those who are actively seeking to ‘fake bad’ from those who
perceive themselves to be su ering greatly, regardless of objective review, may be
V
the combination of Z and Histrionic, Narcissistic and Compulsive scales. The not
abnormal Debasement group had signiŽcantly higher scores on these three scales,
a Ž nding which concurs with research into faking good proŽ les on the MCMI-II
(Retzla et al., 1991).
V
The combination of high Z and low H, N and C might prove to be a useful
apparent and subtle means of identifying those who make deliberate e orts at
V
faking bad (high Z and very low H, N and C) and those who experience great need
(high Z but moderate H, N and C). If this is so, then there may be a means of
distinguishing between those with adaptional sources of malingering and those with
other motivations, and managing them accordingly.
References
American Psychiatric Assocation (1994). Diagnostic and statistical manual of mental disorders (4th ed.).
Washington, DC: American Psychiatric Association.
Bagby, R. M., Gillis, J. R., & Dickens, S. (1990). Detection of dissimulation with the new generation
of objective personality measures. Behavioral Sciences and the Law, 8, 93–102.
Bagby, R. M., Gillis, J. R., Toner, B. B., & Goldberg, J. (1991). Detecting fake-good and fake-bad
responding on the Millon Clinical Multiaxial Inventory-II. Psychological Assessment, 3(3), 496–498.
Greene, R. L. (1988). Assessment of malingering and defensiveness by objective personality
inventories. In R. Rogers (Ed.), Clinical assessment of malingering and deception. New York: Guilford.
Heilbrun, K., Bennett, W. S., White, M. A., & Kelly, J. (1990). An MMPI-based empirical model of
malingering and deception. Behavioral Science and the Law, 8, 45–53.
Lees-Haley (1992). E cacy of MPI-2 validity scales and MCMI-II modiŽ er scales for detecting
Y
spurious PTSD claims: F, F–K, fake bad scale, ego strength, subtle–obvious subscales, DIS, and
DEB. Journal of Clinical Psychology, 48, 681–689.

Debasement and faking bad on the MCMI-III
81
McNiel, K., & Meyer, R. G. (1990). Detection of deception on the Millon Multiaxial inventory
(MCMI). Journal of Clinical Psychology, 46, 755–764.
Millon, T. (1994). Millon Clinical Multiaxial Inventory Manual-III manual. Minneapolis, MN: National
Computer Systems.
Millon, T. (1997). The Millon inventories: Clinical and personality assessment. New York: Guilford.
Pollock, P. (1996). Malingering as mental disorder. Forensic Update, 49, 19–22.
Price, J. R., & Lees-Haley, P. R. (1994). Raising doubts about claims of malingering: Implications of
relationships between MCMI-II and MMPI-2 performances. Journal of Clinical Psychology, 50,
651–655.
Retzla , P., Sheehan, E., & Fiel, A. (1991). MCMI-II report scales and bias: proŽ le and validity scales
V
analyses. Journal of Personality Assessment, 56, 466–477.
Rogers, R. (1990a). Development of a new classiŽ catory model of malingering. Bulletin of the American
Academy of Psychiatry and Law, 18, 323–333.
Rogers, R. (1990b). Models of feigned mental illness. Professional Psychology: Research and Practice, 21,
182–188.
Rogers, R., Bagby, R. M., & Dickens, S. E. (1992). SIRS: Structured interview of reported symptoms,
professional manual. Odessa, FL: Psychological Assessment Resources.
Rogers, R., Bagby, R. M., & Rector, N. (1989). Diagnostic legitimacy of factitious disorder with
psychological symptoms. American Journal of Psychiatry, 146, 1312–1314.
Schretlen, D. J. (1988). The use of psychological tests to identify malingered symptoms of mental
disorder. Clinical Psychology Review, 8, 451–476.
Van Denburg, E. J., & Choca, J. P. (1997). Interpretation of the MCMI III. In T. Millon (Ed.),
The Millon Inventories: Clinical and personality assessment (pp. 41–58). New York: Guilford.
Van Gorp, W. G., & Meyer, R. G. (1986). The detection of faking on the Millon Clinical Multiaxial
Inventory (MCMI). Journal of Clinical Psychology, 42, 742–748.
Wetzler, S., & Marlow, D. (1990). ‘Faking bad’ on the MMPI, MMPI-2, and Millon-II. Psychological
Reports, 67, 1117–1118.
Received 19 June 1998; revised version received 5 March 1999