Deterring Malingered Psychopathology
1
Running head: DETERRING MALINGERED PSYCHOPATHOLOGY
Deterring malingered psychopathology: The effect of warning simulating malingerers.
Joanne King1. & Karen A Sullivan1.
1. School of Psychology and Counselling, Queensland University of Technology
2. Corresponding author: Karen Sullivan, School of Psychology and Counselling, Queensland
University of Technology, Carseldine Campus, Carseldine, Q 4034, AUSTRALIA. Email:
ka.sullivan@qut.edu.au Telephone: 0011 617 3864 4625 Fax: 0011 617 3864 4660.
Acknowledgements
The Human Research Ethics Committee of Queensland University of Technology (QUT3975H) approved all research documented in this report. Preliminary results from this study
were presented as a poster at the Australian Psychological Society College of Clinical
Neuropsychologists Annual Conference, Melbourne 2005. The authors wish to acknowledge
the generous assistance of Kate Ryan who provided help with the collection of data for this
project and the financial support of School of Psychology and Counselling, Queensland
University of Technology.
Deterring Malingered Psychopathology
2
Abstract
The utility of a warning based on deterrence theory to deter malingering on measures of
personality and psychopathology was examined. Sixty-seven first year psychology students
were randomly assigned to one of three conditions: unwarned malingerers, warned
malingerers and controls. Participants in the two malingering groups were given a financial
incentive to simulate believable psychological impairment. Warned malingerers received an
additional warning that the tests could detect malingering and that detection would result in
loss of course credit. Controls received standardised test instructions. It was hypothesized
that the malingering incentive would be sufficient to induce malingering, but that a deterrence
theory warning would have a subsequent deterrent effect. Between-groups analyses indicated
that the warning used in this study significantly altered test performance on the Personality
Assessment Inventory (PAI) and revised Symptom Checklist 90 (SCL-90-R), such that
warned malingerers scored significantly lower (faked less) than unwarned malingerers on the
majority of the psychopathology scales and frequently approximated control group
performances. These results support the effectiveness of a warning to complement existing
malingering detection methods.
Keywords: malingering, psychopathology, personality, faking bad, dissimulation.
Deterring Malingered Psychopathology
3
Deterring malingered psychopathology: The effect of warning simulated malingerers.
Malingering research to date has mostly focused on the construction and validation of
detection methods (see Haines & Norris, 1995; Iverson & Binder, 2000; Slick, Sherman, &
Iverson, 1999) with relatively little research into deterrent methods. However, malingering
deterrents are important for several reasons. First, completely effective malingering detection
methods are unlikely to be developed, despite the advent of increasingly sophisticated
methods (Edens et al., 2001). Second, increased community awareness of the symptoms of
common mental disorders (Lees-Haley & Dunn, 1994; Ruiz, Drake, Marcottee, Glass, & van
Gorp, 2002; Steffan, Clopton, & Morgan, 2003) and the ease with which formal diagnostic
criteria can be accessed, increase the risk of successful malingering. Furthermore, the
possibility that some attorneys may undermine malingering detection by coaching their clients
in taking psychological tests (Victor & Abeles, 2004; Wetter & Corrigan, 1995) also suggests
that methods additional to detection may be needed to address malingering.
Among studies of malingering deterrents, the most widely researched is the low-cost
intervention of malingering warnings embedded within test instructions. Advancing our
understanding of the efficacy of malingering warnings is important because verbal warnings
are already used in some practical settings (Slick, Tan, Strauss, & Hultsch, 2004) and
psychologists arguably have an ethical obligation to ensure clients have sufficient
understanding of the nature and purpose of test procedures (Australian Psychological Society
[APS], 2005), and this may include malingering assessment and the use of detection methods.
Studies of warning methods have mostly concerned malingering of brain injury on
neuropsychological tests (e.g., Erdal, 2004; Gunstad & Suhr, 2001; Johnson & LesniakKarpiak, 1997; Suhr & Gunstad, 2000). For example, Johnson and Lesniak-Karpiak (1997)
assessed differences on the revised Wechsler memory scale (WMS-R) and speeded motor
tasks, between malingerers with and without warning with respect to controls who were told
Deterring Malingered Psychopathology
4
to perform optimally. Malingering was induced by a vignette describing participants’
involvement in a car accident and instructing them to believably fake head injury for a large
financial gain. Warned simulators received an additional warning about simulation detection
techniques. This group malingered less than simulators without warning on the WMS-R
indexes of verbal memory, general memory and delayed recall, while no differences were
found for visual memory and attention-concentration. Warned malingerers approximated
control performance on speeded motor tasks.
The findings of Johnson and colleagues, demonstrating selective warning effects, have
been replicated in some (Erdal, 2004; Gunstad & Suhr, 2001; Johnson & Lesniak-Karpiak,
1997; Suhr & Gunstad, 2000) but not all warning studies. Other simulation studies have
failed to produce any warning effect using measures of memory (Gunstad & Suhr, 2004;
Slick, Hopp, Strauss, & Hunter, 1994; Sullivan, Keane, & Deffenti, 2001; Wong, LernerPoppen, & Durham, 1998), motor function (Wong et al., 1998), intelligence (Johnson, Bellah,
Dodge, Kelley, & Livingston, 1998), or subjective complaints (Sullivan & Richer, 2002).
There are several reasons that may account for the inconsistency in warning effects;
the first of which relates to use of different warning methods across studies. Warning has
previously been construed as a caution that tests may detect malingering. However, some
studies have incorporated symptom coaching into warnings (e.g., Erdal, 2004) whilst others
warn in the absence of coaching (e.g., Wong et al., 1998). Sullivan and Richer (2002) used
warning terminology that included a context-specific deterrent motivator (i.e., prosecution for
insurance fraud). Other methods have included specific advice on evading detection, such as
avoiding major exaggeration (Erdal, 2004), or by encouraging detection avoidance such as
warning one to be careful as the test is designed to catch fakers (i.e., Suhr & Gunstad, 2000).
Providing a warning about a test’s ability to detect malingering and its resulting sanctions,
and cautioning against detection or portraying evasion methods, may not produce the same
Deterring Malingered Psychopathology
5
effects. In addition, use of warnings has also been largely restricted to a written format (e.g.,
Erdal, 2004; Gunstad & Suhr, 2001; Johnson & Lesniak-Karpiak, 1997; Suhr & Gunstad,
2000), even though this may differ from what occurs in clinical practice (Slick et al., 2004),
and no previous warning studies have used bona fide sanctions against malingering.
With one exception (Sullivan & Richer, 2002), warnings have not been designed using
an explicit theoretical framework. This exception used a warning against malingering based
on deterrence theory, which asserts that criminal behavior is preceded by a conscious
evaluation of the costs and benefits derived from the behavior (Pogarsky, Piquero, &
Paternoster, 2004). The theory states that the perceived criminal gain is compared to the
perceived certainty, severity and celerity of the crime’s sanctions (Briscoe, 2004; Gibbs,
1975; Tittle, 1980; Zimring & Hawkins, 1973). Individuals are expected to refrain from
offending if they perceive a high likelihood of detection and punishment, especially if severe
and swiftly delivered (Gibbs, 1975; Zimring & Hawkins, 1973). Support for deterrence
theory is found in studies of drunk-driving (Shore & Maguin, 1988); tax evasion (Klepper &
Nagin, 1989); petty theft (Carmichael, Langton, Pendell, Reitzel, & Piquero, 2005;
Paternoster, Saltzman, Waldo, & Chiricos, 1985); bad check-writing (Paternoster et al., 1985);
vandalism and marijuana use (Carmichael et al., 2005). The rationale for using deterrence
theory as the basis of this study, is that malingering is regarded in some contexts as fraud
(Insurance Council of Australia, 2007) and fraud is a criminal act.
Therefore, using a simulation design, the aim of this study was to investigate warning
efficacy using tests of personality and psychopathology, a domain of functioning in which
warning effects have not been examined previously, using a bone fide sanction delivered both
verbally and in writing, that encouraged perceptions of high sanction certainty, severity and
celerity, to encourage cost-benefit analyses of malingering and hence, deterrence. Perceptions
of sanction certainty were further manipulated by the inclusion of factors intended to increase
Deterring Malingered Psychopathology
6
malingerers’ confidence in their ability to malinger. Specifically, the use of measures of
personality and psychopathology, which are widely regarded as more vulnerable to faking
because of their greater reliance on self-report than performance-based neuropsychological
tests (Aubrey, Dobbs, & Rule, 1989), and the requirement that simulators fake impairments
with which they could be presumed to be familiar given the relatively high level of
community awareness of symptoms associated with psychological disorders (Bagby,
Nicholson, Bacchiochi, Ryder, & Bury, 2002; Lees-Haley, 1997), should increase simulating
malingerers’ confidence in their ability to successfully fake mental illness and thereby
increase the probability of significant deterrent effects. Operationally, it was expected that
unwarned malingerers would score higher than both warned malingerers and controls on the
clinical scales and malingering indexes of the Personality Assessment Inventory (PAI; Morey,
1991) and the revised Symptom Checklist 90 (SCL-90-R; Derogatis, 1992). No differences in
clinical and malingering scores were expected between warned malingerers and controls.
Method
Participants
Participants were 70 first-year psychology students enrolled at the Queensland
University of Technology who received course credit for their participation. Of the 70
recruited participants, data from three of these was excluded because they did not follow
malingering instructions as assessed by manipulation checks. Manipulation checks included a
closed-ended question directly asking participants if they malingered and participants who
failed to malinger when instructed, were excluded from the study, reducing the sample size to
67. A further two participants from the group of 21 warned malingerers, failed manipulation
checks for one psychopathology measure only (i.e., only faked one measure but not the other
when instructed to fake both). We retained their data for analyses on the test they faked, but
otherwise excluded them. This exclusion resulted in a sample size of 20 unwarned
Deterring Malingered Psychopathology
7
malingerers for analyses based on one test only, and a sample size of 19 for those analyses
incorporating data from both tests1.
The final sample comprised 48 (71.6%) females and 19 (28.4%) males. The mean age
of participants was 26 years (SD = 10; range = 17 – 56 years). The majority of participants
was from English-speaking backgrounds (86.6%) and had no history of mental illness
(70.1%). There were no significant differences between experimental groups by age, F (2,
64) = 1.601, p >.05, sex, 2 (2, N = 67) = .519, p >.05, ethnicity, 2 (2, N = 67) = 2.447, p
>.05, or psychological history, 2 (2, N = 67) = 2.251, p >.05.
Measures
Participants completed two measures of personality and psychopathology, the
Personality Assessment Inventory (PAI; Morey, 1991) and the revised Symptom Checklist 90
(SCL-90-R; Derogatis, 1992). The PAI is a self-report inventory that measures a number of
clinical and personality variables (Morey, 1991). The PAI includes 344-items, which are
presented using a four-point nominal scale ranging from F (false, not at all true) to VT (very
true). These items are used to generate twenty-two standard, non-overlapping scales
comprising eleven clinical, two interpersonal, five treatment-related, and four validity scales.
In addition to this, several research based indices have been developed for this test and two of
these, both of which assess malingering (the Rogers Discriminant Function, RDF, Rogers et
al., 1996 and the Malingering index (MAL), Morey, 1996), were also used in this study.
The SCL-90-R is self-report screening test used to assess current psychopathology in
psychiatric and medical patients (Derogatis, 1992). Ninety items are used to generate nine
subscales comprising: somatization, obsessive-compulsive, interpersonal sensitivity,
depression, anxiety, hostility, phobic anxiety, paranoid ideation and psychoticism. SCL-90-R
items are presented on a five-point Likert scale ranging from 1 (not at all) to 5 (extremely),
1
Note: full exclusion of these two participants from all analyses did not alter the pattern of results.
Deterring Malingered Psychopathology
8
such that higher scores indicate greater psychopathology. One of its three global distress
indexes, the Positive Symptom Total (PST; Derogatis, 1992), developed to detect a
dramatizing response style indicative of faking bad, was used in this study as an index of
malingering.
Procedure
Participants were randomly allocated to one of three groups: Unwarned malingerers,
warned malingerers, or controls. After providing informed consent and completing a
demographic questionnaire, participants read an instructional-set specific to group
membership. Unwarned malingerers were instructed to believably fake psychological
impairment on the PAI and SCL-90-R for a chance to win $100. To facilitate this, they were
given a list of characteristic disorder symptoms and were asked to familiarize themselves with
these for a few minutes prior to testing. Rather than request participants fake specific
disorders, we allowed participants to use symptom information in whatever way they wished
to assist their faking. Participants were verbally reminded of the malingering incentive ($100)
immediately prior to test-taking. Warned malingerers received identical instructions to
unwarned malingerers with an additional warning that the measures could detect malingering
and individuals detected would have their course credit for study participation revoked.
These participants were verbally reminded of the malingering incentive immediately prior to
test-taking, preceded by a verbal warning identical to that of the written warning. Control
group participants received standard test instructions to perform optimally with compliance
affording them a chance to win $1002. Participants then completed the PAI and SCL-90-R
which were counterbalanced to control for order effects. Malingering index scores on the
SCL-90-R and PAI were evaluated against relevant published standards (i.e., MAL, Morey,
1996; RDF, Rogers et al., 1996; SCL-90-R PST, Derogatis, 1992).
2
Note: In fact the $100 prize incentive was administered randomly from a pool comprising all participants.
Despite participant instructions, the prize was not distributed based according to particular criteria; a fact about
which participants were informed at debriefing.
Deterring Malingered Psychopathology
9
Consistent with recommendations for research in this area (Edens et al., 2001; Nies &
Sweet, 1994; Sullivan & Richer, 2002), participants completed a post-experimental
questionnaire specific to group membership to assess understanding and compliance with
instructions. Malingering outcome was assessed by a closed question directly asking
participants if they malingered. Malingering strategy use was assessed qualitatively by asking
both warned and unwarned participants to indicate strategies they used to fake believable
deficits from a given list of strategies, in addition to an open-ended question affording them
the opportunity to report their own alternative strategies. Participants were informed they
were able to report more than one malingering strategy.
The Human Research Ethics Committee of Queensland University of Technology
(QUT-3975H) approved all research documented in this report. Because this study involved
deception (i.e., there was the threat that warned malingerers, if detected, would have their
course credit revoked), participants were debriefed. This debriefing included explanation of
the deception. Course credit was not actually revoked from any of the study participants.
Results
Group Differences on the PAI
Differences among the groups (i.e., unwarned malingerers, warned malingerers and
controls) on the primary clinical scales of the PAI were examined using multivariate analysis
of variance (MANOVA). A significant multivariate effect was found between groups on the
11 clinical scales, Pillai’s trace = .781, F (22, 108) = 3.15, p < .001, ղ2 = .391. Table 1
displays the PAI clinical scale means, standard deviations and results from univariate tests
and pairwise comparisons for each individual clinical scale as a function of group
membership. Univariate analyses (with Bonferoni correction to p <.004) demonstrated
significant group differences on all clinical scales with the exception of mania, and antisocial
features. Unwarned malingerers made little attempt to fake mania and antisocial features.
Deterring Malingered Psychopathology 10
Pairwise comparisons indicated that warned malingerers produced significantly lower
psychopathology scores than unwarned malingerers, and frequently approximating that of
controls. There was an exception; warned malingerers scored significantly lower than
unwarned malingerers but higher than controls on the depression scale. Together, these
results largely established the effectiveness of the malingering instructional-set to induce
malingering, and the utility of a warning to reduce malingering.
[Table 1 about here]
Group Differences on the SCL-90-R
Differences among the three groups on the clinical scales of the SCL-90-R were
examined using MANOVA and revealed significant multivariate effects, Pillai’s trace = .532,
F (18, 112) = 2.25, p < .01, ղ2 = .266. Table 2 shows the SCL-90-R clinical scale means,
standard deviations and results for each individual clinical scale as a function of group
membership. Univariate analyses (with Bonferoni correction to p < .005) revealed significant
group differences on all nine scales. Pairwise comparisons indicated warned malingerers
scored significantly lower than unwarned malingerers, and no differently to controls on all
clinical scales with the exception of obsessive-compulsive disorder (OCD). Although no
significant differences emerged between warned and unwarned simulators on OCD,
differences were in the expected direction, demonstrating the efficacy of the malingering
incentive to induce malingering and the efficacy of a subsequent warning to reduce faking.
[Table 2 about here]
Group Differences on the Malingering Indexes
Group differences for the three malingering indexes of the PAI (NIM, MAL, RDF)
were examined using MANOVA. An omnibus multivariate effect was present, Pillai’s trace
= .488, F (6, 122) = 6.57, p < .001, ղ2 = .24. Table 3 displays the means, standard deviations
and univariate results for the PAI and SCL-90-R individual malingering index scales as a
Deterring Malingered Psychopathology 11
function of group membership. Univariate analyses (with Bonferoni correction for PAI scales
to p <.0167) revealed significant group differences on all malingering indexes. Warned
malingerers scored significantly lower than unwarned malingerers, and no differently to
controls on all indexes.
[Table 3 about here]
The percent correct classification across groups using standard cut-offs for the NIM,
MAL, RDF, and PST was also examined. The results of this analysis (see Table 4) showed
that most controls were correctly classified as non-malingerers (>% 80 on three of the four
malingering indices). Most warned malingerers were classified as non-malingerers but
classifications were variable, ranging from 67% for the PST to 96% for the MAL. The RDF
correctly classified almost 90% of participants in the malingerer condition as malingerers, but
the performance of other malingering indices was much less than this number.
[Table 4 about here]
As a further check on results, the analyses of the effect of group on PAI and SCL-90R scores was reexamined using the most accurate of the standard cut-offs to define groups
rather than experimental conditions (the RDF). The results of these analyses were largely
identical to those reported above. That is, the overall PAI and SCL-90-R MANOVAs with
group defined by RDF (malingering versus non-malingering) were significant, and the pattern
of subtest differences was identical in 10 of the 11 PAI subtests, and all of the SCL-90-R
subtests. The subtest that was different was the paranoia subscale of the PAI. When
experimental group was used as the independent variable in the PAI MANOVA, significant
differences between the three conditions (controls, warned and unwarned) were noted as
shown in Table 1, but no such differences were evident when groups were defined by RDF
status (as malingerer versus non-malingerer).
Malingering Strategies
Deterring Malingered Psychopathology 12
Qualitative results pertaining to strategy use are shown in Figure 1. Where
participants reported multiple strategies (N = 11), each reported strategy was counted
separately. Reliance on disorder knowledge was the most common strategy used by
malingerers to fake believable deficits. Other less reported strategies included not reporting
extreme, bizarre or unusual symptoms, attempting to report symptoms that appeared normal,
and generally taking a conservative approach to symptom reporting. In general the strategies
reported by unwarned malingerers and those that faked despite the warning were similar: the
most popular strategy was the same for both groups; and, this ‘popular’ strategy was endorsed
by a similar and relatively large proportion (approximately 70%) of the participants in each
group.
[Figure 1 about here]
Discussion
This simulation study assessed the utility of a warning based on deterrence theory to
deter malingering on measures of personality and psychopathology. The hypothesis that a
relatively weak incentive and malingering instructional set would be sufficient to induce
malingering was supported. Unwarned participants instructed to fake psychological
impairment for the opportunity to win a financial reward, reported significantly greater
psychopathology than controls. This finding is consistent with previous research
demonstrating the ease with which malingering may be induced based on hypothetical
incentives (Erdal, 2004; Sullivan & Richer, 2002; Wong et al., 1998) or bona fide financial
inducements (Sullivan, Keane, & Deffenti, 2001). These results reinforce the notion that real
external incentives offered to bona fide malingerers would, in comparison, be strong
malingering motivators. These results also highlight the vulnerability of personality and
psychopathology measures to malingering and emphasize the need for both deterrence and
detection methods in conjunction with these measures.
Deterring Malingered Psychopathology 13
The Utility of Warning to Deter Malingering
A warning based on deterrence theory and utilizing a bona fide sanction, reduced
malingering on measures of personality and psychopathology. The hypothesis that warned
malingerers would score significantly lower than unwarned malingerers, and no differently to
controls on the PAI, SCL-90-R and their respective malingering indexes, was largely
supported. All malingering indexes, and 17 of the 18 psychopathology scales where
malingering was successfully induced, showed an impact of warning, with one scale
demonstrating partial support for a warning effect. This finding is consistent with studies
supporting the utility of warning to reduce malingering on neuropsychological measures
(Erdal, 2004; Gunstad & Suhr, 2001; Johnson & Lesniak-Karpiak, 1997; Suhr & Gunstad,
2000), and faking good on personality inventories (Braun & Faro, 1968; Nias, 1972). The use
of a theoretically-informed warning method and a bona fide sanction in this study may help
explain why previous research has sometimes failed to find a warning effect (Johnson et al.,
1998; Slick et al., 1994; Suhr et al., 2004; Sullivan et al., 2001; Wong et al., 1998).
The finding that the effects of warning extended to most scales on both measures and
to all malingering indexes, differs somewhat to past neuropsychological warning research that
has found warning to have a more selective and differential effect across measures. For
instance, Johnson and Lesniak-Karpiak (1997) found warning effective on three of five
WMS-R memory indexes, whilst Erdal (2004) found that warning reduced malingering on the
DCT but not on the Rey FIT, even though these tests measure similar constructs. The
relatively robust and less differentiated warning effect found in the present study suggests that
the warning method may account for much of the variation in warning effects found in
previous studies. In addition, warning may be more effective for personality and
psychopathology measures than neuropsychological measures, supporting the notion that
specific tests, abilities or contexts influence warning efficacy (Sullivan & Richer, 2002). It is
Deterring Malingered Psychopathology 14
also possible that the certainty of detection may have been perceived as lower when faking
psychopathology compared to neuropsychological disorders because of greater community
awareness of the former (Bagby et al., 2002; Lees-Haley, 1997), although this requires further
investigation. Furthermore, although speculative, participants may presume that measures of
personality and psychopathology include malingering detection capacities whilst similar
devices may not be assumed to exist in neuropsychological measures because these are less
well known.
These results suggest that a warning method based on deterrence theory, incorporating
a bona fide sanction and delivered both verbally and in writing, can increase the size of the
warning effect. However, it is not clear from this study whether all of these components are
necessary. Hence more work is needed to identify the optimal content and delivery forms for
malingering warnings. Overall the warning effect was robust in this study, yet there were two
anomalies. Warned malingerers faked significantly less depression on the PAI than unwarned
malingerers, yet they faked significantly more than controls. This effect was not observed for
depression on the SCL-90-R. Additionally, warned malingerers scored no differently to
controls or unwarned malingerers for OCD on the SCL-90-R. Again, this effect was not
observed for OCD (anxiety related disorders scale [ARD]) on the PAI. As these effects were
not consistently observed in both the PAI and the SCL-90-R, they may be linked to scale
properties rather than reflect the nature of the underlying psychopathologies.
The finding that potential malingerers were deterred by a relatively weak deterrent,
suggests that warnings which successfully communicate sanctions to bona fide malingerers
may reduce malingering on measures of personality and psychopathology, particularly in
those who lack awareness of detection consequences. These results do not imply that
detection methods be substituted by warnings. It is likely that warnings will never deter all
malingering and detection methods remain the best defense, especially when multiple
Deterring Malingered Psychopathology 15
indicators are used and interpreted carefully (Iverson & Binder, 2000). However, the warning
method investigated here may help to deter individuals from malingering on measures of
personality and psychopathology. Further, without good malingering detection on measures,
warnings would become ineffective over time as individuals learn that tests are unable to
effectively detect malingering. In this respect, the validity and subsequent utility of a warning
remains dependent on the detection capabilities of the measures concerned.
Limitations and conclusions
There are several important limitations of this study. It is unknown whether findings
derived from undergraduate student simulators may generalize to actual malingerers who may
have stronger malingering motivations. The use of samples in which bona fide malingers are
likely to be present, or samples more representative of the populations of interest, such as
inmates, psychiatric patients, compensation claimants, or welfare recipients may provide more
generalizable results. A further limitation of this study was the lack of a clinical control
group. Little is known about how comparing malingerers to well individuals as opposed to a
clinical population, might influence the group differences between malingerers and controls.
Whilst the use of a simulation design is an acknowledged limitation, simulation designs can
provide valuable results because the actual status of malingerers is known. Additionally, the
use of student samples has proven viable with students demonstrating more sophisticated
malingering than clinical simulators (Haines & Norris, 2001).
The idea of incorporating warnings in clinical practice raises several issues. Whilst
such practice is recommended (Iverson, 2006) and North American surveys suggest that
between one quarter (Slick et al., 2004) and one third (Sharland & Gfeller, 2007) of clinicians
already “often or always” use warnings prior to testing, the effect of warnings on genuine
clients must be considered. The present study did not address this issue; however, in previous
work we have shown that warnings do not alter the performance of participants who are trying
Deterring Malingered Psychopathology 16
to perform at the best of their ability, at least when memory tests are used (Sullivan, Deffenti,
& Keane, 2002). On that basis, it seems reasonable to suggest that warnings will not
adversely impact the performance of genuine test-takers, although further research is needed.
On the question of whether or not malingerers become more sophisticated when they received
a warning, we have argued that warnings appear to deter some (up to 70%) but not all
malingerers; and, based on our preliminary data on strategy use, similar strategies were
reported by those who faked whether or not they received a warning. We cannot rule out the
possibility that the few participants that faked despite warning (30% of the warned group) did
so in a more sophisticated manner than if they had they not been warned. Until we are able to
tease out these issues further, perhaps by conducting a closer examination of strategy use, it
remains possible that warnings have both a deterrent effect (perhaps on the majority of people
who receive them) as well as an effect on the strategies used by those who fake despite
warnings.
Deterring Malingered Psychopathology 17
References
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental
disorders (4th ed., text rev.). Washington, DC: Author.
Aubrey, J. B., Dobbs, A. R., & Rule, B. G. (1989). Laypersons’ knowledge about the sequelae
of minor head injury and whiplash. Journal of Neurology, Neurosurgery, and
Psychiatry, 52, 842-846.
Australian Psychological Society. (2005). Code of ethics. Victoria: The Australian
Psychological Society Limited.
Bagby, R., Nicholson, R., Bacchiochi, J., Ryder, A., & Bury, A. (2002). The predictive
capacity of the MMPI-2 and the PAI validity scales and indexes to detect coached and
uncoached feigning. Journal of Personality Assessment, 78, 69-86.
Braun, J., & Faro, D. (1968). Effects of salesman faking instructions on the contact
personality factor test. Psychological Reports, 22, 1245-1248.
Briscoe, S. (2004). Raising the bar: Can increased statutory penalties deter drink-drivers?
Accident Analysis and Prevention, 36, 919-929.
Carmichael, S., Langton, L., Pendell, G., Reitzel, J., & Piquero, A. (2005). Do the experiential
and deterrent effect operate differently across gender? Journal of Criminal Justice, 33,
267-276.
Clifford, D., Byrne, K., & Allan, C. (2004). Getting caught in court: Base rates for
malingering in Australasian litigants. Psychiatry, Psychology and Law, 11, 197-201.
Derogatis, L. (1992). SCL-90-R: Administration, scoring and procedures manual II for the
revised version. Towson, MD: Clinical Psychometric Research.
Edens, J. F., Guy, L. S., Otto, R. K., Buffington, J. K., Tomicic, T. L., & Poythress, N. G.
(2001). Factors differentiating successful versus unsuccessful malingerers. Journal of
Personality Assessment, 77, 333-338.
Deterring Malingered Psychopathology 18
Erdal, K. (2004). The effects of motivation, coaching, and knowledge of neuropsychology on
the simulated malingering of head injury. Archives of Clinical Neuropsychology, 19,
73-88.
Gibbs, J. (1975). Crime, punishment and deterrence. New York: Elsevier Scientific
Publishing Company.
Gunstad, J., & Suhr, J. (2001). Courting the clinician: Efficacy of the full and abbreviated
forms of the Portland digit recognition test: Vulnerability to coaching. The Clinical
Neuropsychologist, 15, 397-404.
Gunstad, J., & Suhr, J. (2004). Use of the abbreviated Portland digit recognition test in
simulated malingering and neurological groups. Journal of Forensic Neuropsychology,
4, 33-47.
Haines, M. E., & Norris, M. P. (1995). Approaches to the detection of malingering of
cognitive deficits: A critical review. Neuropsychology Review, 5, 125-148.
Haines, M., & Norris, M. (2001). Comparing student and patient simulated malingers’
performance on standard neuropsychological measures to detect feigned cognitive
deficits. The Clinical Neuropsychologist, 15, 171-182.
Insurance Australia Group, & Economist Intelligence Unit. (n.d.). Hidden costs: Insurance
fraud in Australia. Retrieved March 21, 2007, from
http://www.iag.com.au/pub/iag/results/submissions/media/20041119a.pdf
Iverson, G. (2006). Ethical Issues Associated With the Assessment of Exaggeration, Poor
Effort, and Malingering. Applied Neuropsychology, 13, 77–90.
Iverson, G. L., & Binder, L. M. (2000). Detecting exaggeration and malingering in
neuropsychological assessment. Journal of Head Trauma Rehabilitation, 15, 829-858.
Johnson, J., Bellah, C., Dodge, T., Kelley, W., & Livingston, M. (1998). Effect of warning on
feigned malingering on the WAIS-R in college samples. Perceptual and Motor Skills,
Deterring Malingered Psychopathology 19
87, 152-154.
Johnson, J., & Lesniak-Karpiack, K. (1997). The effect of warning on malingering on
memory and motor tasks in college samples. Archives of Clinical Neuropsychology,
12, 231-238.
Klepper, S., & Nagin, D. (1989). Tax compliance and perceptions of the risks of detection and
criminal prosecution. Law and Society Review, 23, 209-240.
Lees-Haley, P. R. (1997). MMPI-2 base rates for 492 personal injury plaintiffs: Implications
and challenges for forensic assessment. Journal of Clinical Psychology, 53, 745-755.
Lees-Haley, P. R., & Dunn, J. (1994). The ability of naïve subjects to report symptoms of
mild brain injury, post-traumatic stress disorder, major depression, and generalized
anxiety disorder. Journal of Clinical Psychology, 50, 252-256.
Mittenberg, W., Patton, C., Canyock, E., & Condit, D. (2002). Base rates of malingering and
symptom exaggeration. Journal of Clinical and Experimental Neuropsychology. 24,
1094- 1102.
Morey, L. (1991). Personality assessment inventory: Professional manual. Florida:
Psychological Assessment Resources.
Morey, L. (1996). An interpretive guide to the personality assessment inventory (PAI).
Odessa, FL: Psychological Assessment Resources.
Nias, D. (1972). The effects of providing a warning about the lie scale in a personality
inventory. British Journal of Educational Psychology, 42, 308-312.
Nies, K. J., & Sweet, J. J. (1994). Neuropsychological assessment and malingering: A critical
review of past and present strategies. Archives of Clinical Neuropsychology, 9, 501552.
Paternoster, R., Saltzman, L., Waldo, G., & Chiricos, T. (1985). Assessment of risk and
behavioural experience: An exploratory study of change. Criminology, 23, 417-436.
Deterring Malingered Psychopathology 20
Pogarsky, G., Piquero, A. R., & Paternoster, R. (2004). Modeling change in perceptions about
sanction threats: The neglected linkage in deterrence theory. Journal of Quantitative
Criminology, 20, 343-369.
Rogers, R., Sewell, K., Morey, L., & Ustad, K. (1996). Detection of feigned mental disorders
on the personality assessment inventory: A discriminant analysis. Journal of
Personality Assessment, 67, 629-640.
Ruiz, M., Drake, E., Marcottee, D., Glass, A., & van Gorp, W. (2002). Trying to beat the
system: Misuse of the internet to assist in avoiding the detection of
neuropsychological symptom dissimulation. Archives of Clinical Neuropsychology,
8, 846.
Sbordone, R. J., Syeranian, G. D., & Ruff, R. M. (2000). The use of significant others to
enhance the detection of malingerers from traumatically brain-injured patients.
Archives of Neuropsychology, 15, 465-477.
Sharland, M. J. & Gfeller, J. D. (2007). A survey of neuropsychologists’ beliefs and practices
with respect to the assessment of effort. Archives of Clinical Neuropsychology, 22,
213-223.
Shore, E., & Maguin, E. (1988). Deterrence of drink-driving: The effects of changes in the
Kansas driving under the influence law. Evaluation and Program Planning, 11, 245254.
Slick, D., Hopp, G., Strauss, E., & Hunter, M. (1994). Detecting dissimulation: Profiles of
simulated malingerers, traumatic brain-injury patients, and normal controls on a
revised version of Hiscock and Hiscock’s forced-choice memory test. Journal of
Clinical and Experimental Neuropsychology, 16, 472-481.
Slick, D. J., Sherman, E. M. S., & Iverson, G. L. (1999). Diagnostic criteria for malingered
neurocognitive dysfunction: Proposed standards for clinical practice and research. The
Deterring Malingered Psychopathology 21
Clinical Neuropsychologist, 13, 545-561
Slick, D., Tan, J., Strauss, E., & Hultsch, D. (2004). Detecting malingering: A survey of
experts’ practices. Archives of Clinical Neuropsychology, 19, 465-473.
Steffan, J. S., Clopton, J. R., Morgan, R. D. (2003). An MMPI-2 scale to detect malingered
depression (Md scale). Assessment, 10, 382-392.
Suhr, J., & Gunstad, J. (2000). The effects of coaching on the sensitivity and specificity of
malingering measures. Archives of Clinical Neuropsychology, 15, 415-424.
Sullivan, K., Keane, B., & Deffenti, C. (2001). Malingering on the RAVLT part 1: Deterrence
strategies. Archives of Clinical Neuropsychology, 16, 627-641. Sullivan, K, Lange, R.
T., & Dawes, S. (in press). Symptom exaggeration base rates and detection methods in
Australia, Archives of Clinical Neuropsychology.
Sullivan, K., & Richer, C. (2002). Malingering on subjective complaint tasks. An exploration
of the deterrent effects of warning. Archives of Clinical Neuropsychology, 17, 691708.
Tittle, C. (1980). Sanctions and social deviance. New York: Praeger.
Victor, T., & Abeles, N. (2004). Coaching clients to take psychological and
neuropsychological tests: A clash of ethical obligations. Professional Psychology,
Research and Practice, 35, 373-379.
Wetter, M., & Corrigan, S. (1995). Providing information to clients about psychological tests:
A survey of attorneys’ and law students’ attitudes. Professional Psychology, Research
and Practice, 26, 474-477.
Wong, J., Lerner-Poppen, L., & Durham, J. (1998). Does warning reduce obvious
malingering on memory and motor tasks in college samples? International Journal of
Rehabilitation and Health, 4, 153-165.
Zimring, F., & Hawkins, G. (1973). Deterrence: The legal threat in crime control.
Deterring Malingered Psychopathology 22
Chicago:University of Chicago Press.
Deterring Malingered Psychopathology
23
Table 1. Malingering on the PAI: Means, Standard Deviations, Pooled Within Groups Correlations, and Group Differences
______________________________________________________________________________________________________________________
Controls
Warned
Unwarned
(n = 22)
(n = 24)
(n = 20)
Scale
M
1.SOM
56.38a
11.22 62.00a
15.94 105.84b 34.18 31.35*
2.ANX
51.05a
9.55
58.80a
14.05
82.65 b
17.56 28.84*
3.ARD
51.22a
10.05 58.50a
18.01
82.29 b
19.58 20.50*
4.DEP
50.24a
10.09 66.76b 19.46
92.57 c
23.18 28.41*
5.MAN 51.36a
12.24 45.30a
7.42
56.69 a
18.62
6.PAR
50.42a
9.49
59.39a
20.11
86.38 b
23.00 21.07*
7.SCZ
50.02a
9.79
63.53a
20.42
89.08 b
24.25 22.65*
8.BOR
50.38a
8.19
56.84a
12.61
71.46 b
16.95 14.52*
9.ANT
51.72a
9.63
51.42a
10.58
61.80 a
20.38
3.72
10.ALC
50.90a
8.94
55.95a
13.73
71.95 b
22.87
9.92*
10.57 60.28a
20.37
92.50 b
31.15 18.63*
11.DRG 53.72a
SD
M
SD
M
SD
F(2,63)
4.09
1
2
3
4
5
6
7
8
9
10
11
1.00
.74
.75
.71
.48
.78
.76
.69
.58
.56
.62
1.00
.84
.71
.32
.74
.84
.67
.34
.29
.41
1.00
.75
.34
.80
.83
.69
.48
.37
.46
1.00
.21
.77
.80
.78
.39
.39
.46
1.00
.35
.38
.47
.68
.40
.27
1.00
.87
.75
.56
.42
.45
1.00
.80
.53
.40
.45
1.00
.63
.56
.46
1.00
.59
.59
1.00
.73
1.00
Note. PAI = Personality Assessment Inventory; SOM = Somatic Complaints; ANX = Anxiety; ARD = Anxiety Related Disorders; DEP = Depression; MAN = Mania; PAR =
Paranoia; SCZ = Schizophrenia; BOR = Borderline Features; ANT = Antisocial Features; ALC = Alcohol Problems; DRG = Drug Problems. Means represent T-scores (M =
50, SD = 10) with a T-score at or above 70 representing 2 SDs from the nonclinical college student standardisation sample. Means in the same row that do not share subscripts
differ significantly at p < .004 using Bonferroni correction.
*p < .001.
Deterring Malingered Psychopathology
24
Table 2
Malingering on the SCL-90-R: Means, Standard Deviations, Pooled Within Groups Correlations, and Group Differences
______________________________________________________________________________________________________________________
Scale
Controls
Warned
Unwarned
(n = 22)
(n = 24)
(n = 20)
M
SD
M
SD
M
SD
F(2,63)
1.SOM
60.27a 19.60
61.23a
17.50
91.65b
30.57
12.77*
2.O-C
65.19a 17.84
74.85a&b 21.20
90.10b
22.25
7.84*
3.IS
72.55a 21.55
71.26 a
25.96
102.94b 31.25
9.69**
4.DEP
6603 a
20.18
72.68 a
24.84
96.08 b
27.76
8.72**
5.ANX
60.32a 24.97
67.11 a
22.61
101.79b 30.88
15.09**
6.HOS
63.99a 19.83
64.90 a
20.56
87.62 b
7.PHOB 57.01a 27.79
64.22 a
37.68
104.41b 45.00
9.79**
8.PAR
62.09a 19.29
61.62 a
23.13
89.49 b
31.00
8.74*
9.PSY
72.74a 37.47
75.71 a
37.34
124.82b 49.81
10.34*
31.24
6.45*
1
2
3
4
5
6
7
8
9
1.00
.72
.64
.69
.83
.66
.67
.66
.69
1.00
.75
.76
.79
.73
.61
.74
.74
1.00
.80
.83
.65
.79
.81
.76
1.00
.77
.67
.67
.69
.71
1.00
.72
.87
.81
.85
1.00
.57
.75
.69
1.00
.74
.73
1.00
.80
1.00
Note. SCL-90-R = Symptom Checklist-90 Revised; SOM = Somatisation; O-C = Obsessive-Compulsive; I-S = Interpersonal Sensitivity; DEP = Depression; ANX =
Anxiety; HOS = Hostility; PHOB = Phobic Anxiety; PAR = Paranoid Ideation; PSY = Psychoticism. Means represent scale T-scores (M = 50, SD = 10) with a T-score at or
above 70 representing a pronounced deviation (2 SDs) from the nonclinical standardisation sample. Means in the same row that do not share subscripts differ significantly at
p < .005 using Bonferonni correction.
*p < .005 **p < .001.
Deterring Malingered Psychopathology
25
Table 3
Performance on the Malingering Indexes: Means, Standard Deviations, Pooled Within Groups Correlations, Effect Sizes, and Group Differences
______________________________________________________________________________________________________________________
Scale
Controls
Warned
Unwarned
(n = 22)
(n = 24)
(n = 19a)
M
SD
M
SD
M
SD
F(2,62)
db
1
2
3
4
PAI
1.NIM
52.89a
9.95
63.83 a
21.28
97.97 b
38.61
17.51**
1.70
2.MAL
.82 a
.91
1.00 a
1.50
2.68 b
1.95
9.60**
1.29
3.RDF
.92 a
1.52
-.45 a
1.61
1.76 b
1.43
17.58**
1.86
43.42 a
21.48
66.95 b
25.58
7.05*
.96
1.00
.75
.29
.47
1.00
.31
.39
1.00
.42
SCL-90-R
4.PST
42.55 a
23.44
1.00
Note. NIM = Negative Impression Management Scale (PAI; ≥ 92T = malingering); MAL = Malingering Index (PAI; ≥ 5 = malingering); RDF = Roger’s Discriminant
Function Index (PAI; > .12368 = malingering); PST = Positive Symptom Total (SCL-90-R; > 50 males, > 60 females = malingering). The cutoffs listed are original
recommended cutoffs by test authors. Means in the same row that do not share subscripts differ significantly at p < .0125 using Bonferonni correction.
a
The two partially-excluded participants were omitted from this analysis to permit comparison across malingering indexes.
Cohen’s d = effect sizes for unwarned malingerers versus controls.
b
*p < .01. **p < .001.
Deterring Malingered Psychopathology 26
Table 4
Percent Correct Classification as Non-malingerer Across Three Experimental Groups (Controls,
Warned Malingerers and Malingerers) Using standard Cut-Offs for Four Malingering Indices.
Controls (n = 22)
Warned (n = 24)
Unwarned (n = 19a)
NIM
100
92
52
MAL
100
96
74
RDF
82
71
11
PST
68
67
32
Note. NIM = Negative Impression Management Scale (PAI; ≥ 92T = malingering); MAL = Malingering Index
(PAI; ≥ 5 = malingering); RDF = Roger’s Discriminant Function Index (PAI; > .12368 = malingering); PST =
Positive Symptom Total (SCL-90-R; > 50 males, > 60 females = malingering). The cutoffs listed are original
recommended cutoffs by test authors.
a
The two partially-excluded participants were omitted from this analysis to permit comparison across
malingering indexes.