Malingering during forensic evaluation: Antisocial, pathological or cost-benefit analysis?
Roos Paris
Studentnummer: 10352139
Content
Abstract pag. 4
Introduction pag. 5
Forensic evaluations pag. 5
Malingering pag. 6
Assessing malingering pag. 7
Why do forensic clients malinger? pag. 8
Empirical evaluation of the explanatory models pag. 9
Current study pag. 11
Method pag. 14
Participants pag. 14
Instruments pag. 18
Procedure pag. 20
Results pag. 24
Cost-benefit hypothesis pag. 24
Criminological hypothesis pag. 26
Pathogenic hypothesis pag. 28
Discussion pag. 33
Cost-benefit hypothesis pag. 33
Criminological hypothesis pag. 35
Pathogenic hypothesis pag. 36
Limitations pag. 38
Suggestions further research pag. 40
Implications pag. 41
References pag. 43
Abstract
Malingering can occur during forensic evaluation (Mendelson & Mendelson, 2014) and complicates correct psychological assessment, which is essential for successful treatment and safe rehabilitation. It is therefore important to learn more about the nature of malingering and what provokes it. The current study has examined the motivation for malingering
following the three explanatory models of Rogers (1990a, 1990b, 2008), which state that malingering can be explained by (1) cost-benefit analysis, (2) antisocial motives and/or (3) underlying psychopathology. We examined the scores on different symptom validity tests (ASTM, validity scales of the MMPI-2-RF, and SIMS) of 546 evaluees. There was no relation between malingering indices and length of the maximum prison sentence claim, as predicted from the cost-benefit analysis. There was also no relationship between malingering indices and antisocial personality disorder (APD), as predicted from antisocial motives. We only found partial support for Pathogenic model: Malingering was more likely for borderline personality disorder (BPD) or a depressive disorder (DD).
Introduction
Mrs. Z. is admitted for a clinical forensic evaluation on a second-degree arson charge. Although she is 31 years old, she behaves like a small child: she walks around with a doll, likes dressing up like a princess and has trouble completing relatively simple tasks like preparing breakfast. She also says that she continuously hears voices. The research team conducting the forensic evaluation has doubts about the sincerity of the symptoms reported by Mrs. Z. They scent that Mrs. Z. might be malingering. To explore this hypothesis, multiple symptom validity tests are administered. The results indeed indicate that Mrs. Z. is
malingering. The question that is however left unanswered, is why she would demonstrate this behavior.
Malingering – the willful, deliberate and fraudulent feigning or exaggeration of illness – can occur during forensic evaluations (Mendelson & Mendelson, 2014) and distorts the diagnostic conclusions reached during this assessment, which are essential for successful treatment and safe rehabilitation. It is therefore important to learn more about the nature of malingering and what provokes it. The current study has examined the motivation for malingering following the three explanatory models of Rogers (1990a, 1990b, 2008), which state that malingering can be explained by (1) cost-benefit analysis, (2) antisocial motives and/or (3) underlying psychopathology.
Forensic evaluations
In the Netherlands a forensic evaluation can be requested by the court or public prosecutor when thought a mental health assessment of a defendant is needed. The Nederlands Instituut voor Forensische Psychiatrie en Psychologie (NIFP) is responsible for forensic evaluations and assigns independent behavioral scientist to evaluate a defendant before he or she goes to
offence. For a clinical evaluation a defendant is admitted to the Pieter Baan Centrum (PBC), the observation clinic of the NIFP. However, most evaluations are done ambulatory. In a forensic evaluation the behavioral scientist attempt to answer the following questions:
- Did the defendant have a psychiatric disorder at the time of committing the offence? - Did the psychiatric disorder influence or let to committing the offence?
- Is there risk of reoffending based on the psychiatric disorder?
- Based on the established psychiatric disorder and risk of reoffending, the behavioral scientists give a recommendation regarding the possibilities of reducing recidivism. The conclusions reached in the forensic evaluation are taken into consideration during the judges’ deliberation about the liability of a person at the time of the offence and the verdict.
Malingering
Clinicians indicate a prevalence of malingering in the forensic population between 15-19% (Roger et al., 1998; Rogers, Sewell & Goldstein, 1994; Mittenberg et al., 2002). During a forensic evaluation malingering of symptoms can be used to diminish criminal
responsibility (Mendelson & Mendelson, 2014). When done successfully, it complicates correct assessment of possible psychiatric disorders (Cornell & Hawk, 2014): the distorted symptom presentation can lead to a faulty diagnosis, while symptoms that are actually present go unnoticed, which impacts possible treatment. When a person with a faulty diagnosis ends up in forensic psychiatric treatment and gets therapy for a psychiatric disorder he or she does not have, treatment will not be effective. It is however also possible that a person with malingered symptoms stops malingering as soon as he or she enters the treatment program. This way treatment will look successful while the symptoms that may actually be present stay untreated. Risk assessment of recidivism neither will be reliable when based on a faulty symptoms. When malingering goes unnoticed and wrongful interventions and risk
assessments are done, it can endanger the safety of the society.
Assessing malingering
Symptom validity tests aim to detect malingering (Wygant et al., 2007; Schmand, De Sterke & Lindeboom, 1999; Tombaugh, 1996; Smith, 1997) and for this reason a minimum of two symptom validity tests are administered during forensic evaluations (Heilbrun, 2002). Often used for this purpose are the Amsterdam Short-Term Memory test (ASTM; Schmand, De Sterke & Lindeboom, 1999), the Structured Inventory of Malingered Symptomatology (SIMS; Smith, 1997) and the validity scales of the Minnesota Multiphasic Personality Inventory (MMPI; Butcher, 1998). The ASTM identifies memory malingering using a test that is presented to the participant as a memory test but for which the actual memory load is minimal, making is a fairly easy test (Schmand, De Sterke & Lindeboom, 1999). A low score is an indication of memory malingering. According to the research of Hoogstraten and Kemperman (2005) the ASTM is successful in detecting memory malingerers, they found a specificity of .89 and sensitivity of .75. The SIMS and validity scales of the MMPI detect the feigning or exaggeration of symptoms. Both are self-report questionnaires with a variety of strategies to detect malingering, like endorsement of bizarre, unlikely or rare symptoms and very unusual answer combinations. A high score on either of these tests is an indication of malingering. Wygant et al. (2009) found good sensitivity and specificity for the validity scales of the MMPI-2-RF (between .70 and .90 for every scale) in the civil forensic setting and Merckelbach & Smith (2003) found a Cronbach’s alpha of .72, a specificity of .98 and a sensitivity of .93 for the total score of the SIMS.
Why do forensic clients malinger?
More knowledge about the motives behind malingering would be helpful to better
understand the nature of malingering and how it gets provoked, making it easier for clinicians to recognize and to integrate in their forensic evaluation. Rogers (1990a, 1990b, 2008)
proposed three explanatory models for understanding malingering: the Pathogenic, the
Criminological, and the Cost-Benefit model. The Pathogenic model, states that malingering is induced by underlying psychopathology. Patients with psychiatric disorders would, by
feigning symptoms, experience a sense of control over their disorder. This because their feigned symptoms are controllable as opposed to their actual symptoms, which are not. The Criminological model states that defendants malinger out of antisocial motives. This model is based on the assumption that people with a antisocial personality disorder (APD), a disorder characterized by deceit and manipulation, tend to malingering when embroiled in forensic evaluation. This may derive from their personality (they simply cannot not manipulate) or may be induced for the purpose of their personal pleasure. A relationship between APD and malingering can indeed be expected when revising the diagnostic criteria of the disorder (DSM-IV, 2000). The DSM-IV (2000) states that APD is, amongst other symptoms, characterized by: “deceitfulness, as indicated by repeated lying, use of aliases, or conning others for personal profit or pleasure”. This may include malingering. Even though deceit for personal profit is mentioned as a symptom of APD, this is not part of the Criminological model. The Criminological model focusses on behavior deriving from the tendency to manipulate and the personal pleasure gotten from deceiving others. Malingering for personal profit falls under Rogers’s last model, the Cost-Benefit model. This model states that would-be malingerers make a cost-would-benefit analysis and consciously chose for malingering if this lead to the most advantageous outcome, for example because admission to a psychiatric hospital seems more appealing than being locked up in detention. Approaching the Cost-Benefit
model from a juridical perspective, malingering would only lead to the most advantageous outcome when the anticipated time spend in a forensic psychiatric hospital would transcend the anticipated time spend in detention.
Empirical evaluation of the explanatory models of malingering in the forensic context
The Cost-Benefit model found support in various contexts, as pointed out in the literature review by Rogers (1990b). For example, an increased incidence of malingering has been found during wartime. Pretending to be mentally ill is possibly a way of avoiding being called to the front to fight. In another study the costs of malingering for insurance companies in the United Kingdom was estimated approximately 2.1 billion pounds a year (NFA, 2013; Cartwright et al, 2016). When a person applies for an insurance claim, feigning or
exaggeration symptoms may result in a higher financial compensation. Rogers (1990b) also found that malingering occurs more often when a defendant is faced with a very serious charge as opposed to a comparatively benign offence. There needs to be noted that all these studies are correlational rather than experimental, and therefore a causal relationship can not be assumed. The Cost-Benefit model has yet to be studied for the forensic evaluations in the Netherlands.
Several studies have examined the Criminological model, particularly focusing on the assumed relationship between APD and misleading response styles. The results however are mixed (Nietsen et al., 2015). Mills and Kroner (2006) for example, found a positive
relationship between malingering and APD, while Pierson et al. (2011) did not found any relationship between the two. However, both studies have their own methodological
shortcomings; Mills and Kroner’s study (2006) did not take place in a forensic context, where conclusions drawn from the measurements have consequences and the results in the study of
(2015) expressed some critical notes about shortcomings in the theory behind the
criminological model. They stated that even though the model assumes a relationship between APD and ‘faking bad’, it does not go into detail about the nature of the relationship. The model does not clarify if people with APD tend to ‘fake bad’ more often than people without APD or if people with APD are that good in being deceitful that their malingering goes unnoticed. Furthermore the model focusses mainly on ‘faking bad’ while in the forensic context ‘faking good’ may also be relevant. Finally it does not pay attention to the possibility that malingering of symptoms may be accompanied by the experience of sincere symptoms.
The Pathogenic model and the possible relationship between a specific mental disorder and malingering has yet to be studied closely. Rogers (1990a, 1990b, 2008) states that patients with mental disorders malingering symptoms to experience a sense of control over their disorder. The feeling of control is however hard to operationalize and this makes it difficult to base a testable hypothesis on his reasoning. Hopwood, Morey, Rogers and Sewell (2007) hold a different interpretation of the Pathogenic model. They suppose that malingering may occur in mental disorders known with cognitive distortions and give borderline
personality disorder (BPD) and depression as examples. These distortions cause patients to overrepresent the negative aspects of themselves, their environment and their future, which may induce over-reporting of symptoms. Even though this hypothesis is easier to
operationalize, it has not been examined before. Although the relationship between
BPS/depression and cognitive distortions and biases is established in multiple studies (Moritz et al., 2011; Giesen-Bloo & Arntz, 2005; Gonda et al., 2015), one might even assume that every psychiatric disorder comes with cognitive distortions, which makes the reasoning of Hopwood, Morey, Rogers and Sewell less discriminating.
The current study
Given the importance to learn more about malingering during forensic evaluations, it is worth examining which model of Rogers finds the most (if any) support within this context. The current study therefore examines the relationship between malingering and the three models of Rogers in the context of clinical forensic evaluations in the Netherlands (Table 1). For the Cost-Benefit model, the relationship between the prevalence of malingering and ascending length of the maximum prison sentence is examined. From a cost-benefit point of view one would expect to find more malingering when the length of the maximum prison sentence considerably transcend the anticipated length of admission to a forensic hospital. In the Netherlands the average time spend in a forensic psychiatric hospital has decreased from almost ten years (Nagtegaal, van der Horst & Schönberger, 2011) to seven to eight years (van Schijndel, Linckens, Bagchus & Gordeau, 2017). If a defendant gets charged with an offence with a maximum prison sentence of more than eight years admission to a forensic psychiatric hospital may be more beneficial, which can provoke malingering. Betting on this is however not without risk. The maximal length in a forensic psychiatric hospital is not set in advance and can be extended for an unlimited amount of time. Even though this can either turn out in the advantage or disadvantage of a defendant, the insecurity about the length of the sanction (that will continue to exist throughout the admission) can be experienced as dragging and nerve-racking. To avoid being sentenced to a combination of prison time and admission to a forensic psychiatric hospital, a defendant must be completely dismissed for want of
prosecution by being found not guilty by reason of insanity. Malingering can contribute to this. Because of all the risks that come with choosing malingering in this context, we decided to use a higher average of time spend in a forensic hospital in current study, namely ten years. Based on the Cost-Benefit model, a positive relationship is expected between malingering and
during forensic evaluation to avoid being send to a forensic hospital (Muis & Van der Geest, 2010), one may say that in practice admission to such an institution does not seem to be experienced as the most advantageous outcome.
For the Criminological model defendants with APD will be compared to defendants without this disorder. As described earlier, it is possible that people with APD have a
tendency to deceive and manipulate due to their personality or do it for their personal pleasure and therefore might malinger more. Therefore a positive relationship is expected between malingering and APD, based on the Criminological model. Deceit concerning symptom presentation can however be done either by ‘faking bad’ or ‘faking good’. Considering the fact that lawyers tend to advice their clients to avoid being diagnosed with a psychiatric disorder, ‘faking good’ seems in practice like the more logical choice in this forensic context. This way a defendant with APD catches two birds with one stone: he or she can give in to the tendency to deceive and follows his or her lawyers advice.
For the Pathogenic model defendants with either BPD or a DD will be compared to defendants without these disorders. BPD and a DD seem to be intertwined with the already described cognitive distortions, which cause overrepresentation of the negative aspects of the self, the environment and the future. These distortions may influence the symptom
presentation and provoke malingering. This is why a positive relationship is expected between BPD/DD and malingering.
Table 1
Overview of the explanatory models of malingering, their operationalization and expected outcome used in current study.
Model Content Operationalization Hypothesis
Pathogenic Model
Patients with BPD/DD malinger possibly due to cognitive distortions.
BPD/DD vs. [PD cluster A or C or Axis I disorder] and vs. No disorder Patients in the BPD/DD condition malinger more compared to the other conditions. Criminological Model
Patients with APD malinger due to a tendency derived from their personality.
APD vs. [PD cluster A or C] and vs. No PD
Patients in the APD condition malinger more compared to the other conditions.
Cost-Benefit Model
People malinger when it leads to the most
advantageous outcome. Maximum prison sentence of £ 10 years vs. [10-20 years] and vs. ³ 20 years Ascending maximum prison sentence increases malingering
Method
Participants
All participants were aged between 18 and 67 years old and were admitted to the PBC between 2008-2016 for a clinical forensic evaluation of six or seven weeks. The total sample in current study consisted of 546 participants (Table 2; M age = 37.01, SD = 11.09; 91% men).
Because of different selection criteria used to test each model, the number of included participants differs per tested model. For the Cost-Benefit hypothesis 546 participants were included (Table 2), for the Criminological hypothesis 468 participants were used (Table 3) and for the Pathogenic model 448 participants were included (Table 4).
General exclusion criteria for assessment were active symptoms of psychosis or objection against the use of their personal information for scientific research. General inclusion criteria for this study were that a participant was admitted to the PBC between 2008-2016 (see ‘data collection’ for further explanation of this specific interval) and had done at least one symptom validity test during his or her stay.
Criminological and Pathogenic model an extra inclusion criteria was added: The
behavioral scientists came in consensus to a diagnostic conclusion about the participant. This conclusion could also be that there was no psychiatric disorder. There was also a third
exclusion criteria for the Criminological and Pathogenic model: Participants diagnosed with both BPD/DD and APD (so BPD and APD or DD and APD). This because differentiating between Criminological model and the Pathogenic model is not possible when a combination of these disorders is present.
Table 2
Demographic information of the 546 participants of the Cost-Benefit hypothesis and corresponding (significant) differences.
Cost-Benefit hypothesis Condition 1: £ 10 years (n=131) Condition 2: 10-20 years (n=225) Condition 3: ³ 20 years (n=187) p-value Gender (men) 92% 93% 88% .05* Age (M) 36.56 (11.08) 36.74 (11.22) 37.86 (11.04) .46 Ethnicity Native 84% 75% 71% .01* Western immigrant 2% 7% 5% Non-western immigrant 14% 18% 24% Level of Education Low 79% 85% 82% .18 Middle 16% 13% 13% High 5% 2% 5% * p ≤ 0.05
Table 3
Demographic information of the 468 participants of the Criminological hypothesis and corresponding (significant) differences.
Criminological hypothesis Condition 1: APD (n=49) Condition 2: Cluster A or C (n=164) Condition 3: No PD (n=225) p-value Gender (men) 98% 92% 93% .15 Age (M) 33.39 (9.41) 39.13 (10.28) 36.12 (11.35) < .01* Ethnicity Native 61% 84% 74% < .01* Western immigrant 12% 4% 5% Non-western immigrant 27% 12% 21% Level of Education Low 100% 87% 76% < .01* Middle 0% 10% 18% High 0% 3% 6% * p ≤ 0.05
Table 4
Demographic information of the 448 participants of the Pathogenic hypothesis and corresponding (significant) differences.
Pathogenic hypothesis Condition 1: BPD or DD (n=60) Condition 2: Cluster A or C or other Axis I disorder (n=311) Condition 3: No disorder (n=77) p-value Gender (men) 73% 94% 88% < .01* Age (M) 40.37 (13.12) 36.50 (10.34) 37.70 (11.91) .18 Ethnicity Native 83% 79% 68% .06 Western immigrant 7% 4% 6% Non-western immigrant 10% 17% 26% Level of Education Low 68% 82% 82% < .01* Middle 18% 15% 13% High 14% 3% 5% * p ≤ 0.05
Instruments
Classification of psychiatric disorders –
Participants in this study were diagnosed by two qualified behavioral scientists (a psychologist and a psychiatrist), which means that these behavioral scientists are trained in doing forensic evaluations and are connected to the NIFP. Their diagnoses are based on the information (observation reports, interviews, psychological tests, and a forensic
environmental examination) that is gained by the research team during the six or seven weeks that a participant is admitted in the PBC for observation, using the DSM-IV-TR criteria. The research team consists of a forensic environmental researcher, a lawyer, a test psychologist, a group mentor, a process-supervisor (a senior psychologist of psychiatrist) and the assigned psychologist and psychiatrist. All the information gained by the different team members is shared in set meetings during the observation period. The assigned psychologist and
psychiatrist are ultimately responsible for the conclusions reached in the forensic evaluation and can be summoned by the court to answer questions about these conclusions.
Symptom validity tests –
ASTM – The Amsterdam Short-Term Memory test (ASTM) is developed by Schmand, De Sterke & Lindeboom (1999) and detects feigned memory deficits. The test consists of two practice items and 30 test items. For each item the participant is asked to read five response words from a common semantic category out loud (for example: salt, sugar, pepper, cinnamon, vinegar) and memorize them. Then the subject is distracted with a simple addition or subtraction task (for example: 56 + 15). Next the subject is presented with five words from the same semantic category as before (for example: mustard, sugar, paprika, salt, pepper), from which three are the same as in the previous trial and two are new. The subject must indicate which three words appeared in both series. Feedback is given on the number of
correct words. The test score is the number of correct responses. The maximum score is 90 and a score below 85 is an indication for memory malingering (for subjects without evident cognitive disorders for which this test is not suitable). According to the research of
Hoogstraten and Kemperman (2005) the ASTM is successful in detecting memory malingerers, they found a specificity of .89 and sensitivity of .75.
MMPI-2-RF – The Minnesota Multiphasic Personality Inventory is originally developed by Butcher (1998). The validity scales of the MMPI-2-RF are often used to indicate impression management or distortion (Ben-Porath & Tellegen, 2008). The eight validity scales of the MMPI-2-RF can be divided into three categories: scales that indicate inconsistent answering (CNS-, VRIN-r and TRIN-r-scale), scales that indicate ‘faking bad’ (F-r, Fp-r, Fs- and FBS-r-scale) and scales that indicate ‘faking good’ (L-r- and K-r-scale). Because current study focusses on ‘faking bad’ only the four validity scales that indicate this will be used. For every scale the items are dichotomous (true/false) and a high scores is indicated by a t-score > 80. The F-r-scale consists of 32 items about rare psychopathological, somatic and cognitive symptoms. A high score on this scale indicates over-reporting of symptoms (Ben-Porath & Tellegen, 2008). An example item of this scale is: “I have nightmares every few nights”. The Fp-r-scale (infrequent psychopathology responses) consists of 21 items and points out over-reporting of psychopathological symptom or a ‘cry for help’ (Derkens, 2006). An example of an items from this scale is: “Everything tastes the same”. This scale is used to differentiate between subjects with actual serious
psychopathology and subjects who pretend to suffer from serious symptoms: The items from this scale are about psychiatric symptoms that are rarely reported by people with (even severe) mental disorders (Ben-Porath & Tellegen, 2008). The Fs-scale (infrequent somatic responses) consists of 16 items about somatic symptoms rarely seen in medical patients, for
somatic symptoms (Ben-Porath & Tellegen, 2008). The FBS-r-scale consists of 30 items and indicates noncredible somatic and neurocognitive complaints (Sellbom ,Toomey, Wygant, Kucharski & Duncan, 2010) and a high score indicates over-reporting of these symptoms. An example item of this scales is: “I am troubled by discomfort in the pit of my stomach every few days or oftener”. Wygant et al. (2009) found good sensitivity and specificity for the validity scales of the MMPI-2-RF (between .70 and .90 for every scale) in the civil forensic setting, which indicates that the over-reporting validity scales are effective at detecting symptom over-reporting in this context.
SIMS – The Structured Inventory of Malingered Symptomatology is developed by Smith (1997) and detects malingering of symptoms of five different conditions: low
intelligence, affective disorders, neurological impairment, psychosis and amnestic disorders. The test is a self-report questionnaire with 75 dichotomous (correct/incorrect) items. The test incorporates a variety of strategies to detect malingering, including endorsement of bizarre symptoms (for example: “Flowers have magical powers like the ability to talk to people”) and unlikely complaints (for example: “Sometimes I lose all feeling in my hand so that it is as if I have a glove on”). A total score of 17 or higher indicates malingering. Merckelbach & Smith (2003) found a Cronbach’s alpha of .72, a specificity of .98 and a sensitivity of .93 for the total score of the SIMS.
Procedure
Power analysis –
A power analysis was done, using the software G*Power, to estimate the minimum number of participants needed to be able to pick up the expected effects. Since the results of this study rely on multiple ANCOVA’s, the alpha used in the analysis was corrected for
ANCOVA’s done (3 hypotheses times 6 test scores = 18). For every hypothesis an alpha of 0.0028 (0.05/18), a power of .8 and a medium effect size of f = .25 were used. The possibility of multiple confounders (gender, age, ethnicity and level of education) was also taken into account by indicating the use of four covariates in the analysis. A minimum of 280
participants was needed for the intended power, 94 per condition.
Data collection -
Initially 620 participants wanted to cooperate with psychological testing during their stay in the PBC in the period 2008-2016. A total of 67 participants were excluded due to active symptoms of psychosis, no administration of a symptom validity test or objections against the use of personal information for scientific research, leaving 553 participants. The corresponding forensic evaluation report needed to code demographic information,
psychiatric diagnoses and charged offence was found for 516 of these participants.
The PBC uses a standard psychological test battery since 2008 (Spaans et al., 2014). The MMPI-2 (1998), ASTM (1999) and SIMS (1997) are part of this battery. This means that without contra-indications (low IQ, illiteracy or does not speak the Dutch language) all these tests are administered for every defendant who cooperates during the forensic evaluation. There was however a varying number of missing data for each test. For the ASTM, we retrieved data from 412 of the 553 participants (28 not completed; 13 completed but not found). For the MMPI-2, we retrieved data from 374 of the 553 participants (128 not completed; 51 completed but not found). All of the 374 participants answered enough
questions to make their scores on the MMPI-2 valid and reliable (none of the participants left more than 19 questions unanswered). For the SIMS, we retrieved data from 407 of the 553 participants (131 not completed; 15 completed but not found). Because not every participant
For a participant to be included in the sample of the Cost-Benefit hypothesis, besides having done at least one symptom validity test, only the charged offence needed to be known. To be included in the Criminological hypothesis a participant had to be diagnosed with either APD, a PD of cluster A or C, or no PD. For a participant to be included in the Pathogenic hypothesis, he or she had to be diagnosed with either BPD, a DD, a PD of cluster A or C, another disorder from Axis I or no disorder.
Data-analyses –
To make sure that the possible found differences in malingering between conditions could be attributed to the independent variable and not to a difference between the conditions on one of the demographic variables, we first examined our demographic data. We compared the variables gender, age, ethnicity and level of education between the different conditions for each hypotheses, using the Chi-Square and the Kruskal-Wallis Test (Table 2-4). For ethnicity the country of birth of the participant was used and level of education was categorized using the general guidelines of the Centraal Bureau voor Statistiek (CBS, 2016). When a difference on a demographic variables was found (a confounder), it was controlled for in the main analyses to make sure the difference did not influence the found effect. For the Cost-Benefit model the condition ‘£ 10 years’ was compared to the condition ‘10-20 years’ and ‘³ 20 years’ and the conditions ‘10-20 years’ and ‘³20 years’ were compared. Note that the category ‘³ 20 years’ also includes life sentence, which in the Netherlands actually means being imprisoned for life. For the Criminological model the condition APS was compared to the clinical condition (PD from cluster A or C) and to the non-clinical condition (no PD). The clinical and non-clinical condition were also compared. Finally for the Pathogenic model the BPD/DD condition was compared to the clinical condition (PD from cluster A or C or other disorder from Axis I) and to the clinical condition (no disorder). The clinical and
non-clinical condition were also compared.
For every condition the average scores and standard error were calculated (Table 5-7). ANCOVA’s were used to compare the average scores on the symptom validity tests (ASMT, SIMS and validity scales of the MMPI-2-RF) between the different conditions, after
controlling for possible confounders found in the analysis of the demographic variables. Simple contrasts were used to further explore significant effects. For the Cost-Benefit
hypothesis the variables gender and ethnicity were controlled (Table 2). The contrasts for this hypothesis were: ‘£ 10 years’ vs. ’10-20 years’ (contrast 1) and ‘£ 10 years’ vs. ’ ³ 20 years’ (contrast 2). For the Criminological hypothesis the demographic variables age, ethnicity and level of education were controlled (Table 3). Here the contrasts were: APD vs. PD from cluster A or C (contrast 1) and APD vs. no PD (contrast 2). In the Pathogenic hypothesis the demographic variables gender and level of education were controlled (Table 4). The contrasts of the Pathogenic hypothesis were: BPD/DD vs. PD from cluster A or C/other Axis I disorder (contrast 1) and BPD/DD vs. no disorder (contrast 2).
We decided to use partial eta squared (partial η²) to calculate the effect size based on the spreadsheet of Daniel Lakens (Lakens, 2013). Partial η² = .01 indicates a small effect,
partial η² = .10 indicates a medium effect and partial η² = .25 indicates a large effect (Cohen,
1988). For the interpretation of the results of our main analyses, we focussed on the effect sizes rather than on the p-values, as recommended by Cumming (2014).
Because the analysis of the data involves multiple ANCOVA’s the problem of multiplicity arises. For this reason the level of significance will post-hoc be corrected using the Holm’s step-down procedure. We chose for this posted-hoc correction as opposed to altering the alpha beforehand (as we did in de power analysis) because the Holm’s step-down
Results
Cost-Benefit hypothesis –
The analyses of the demographic variables, using the Chi-Square and the Kruskall-Wallis Test (Table 2), revealed that there was no significant difference found in gender between the conditions ‘£ 10 years’ and ‘10-20 years’, X 2 (1) = .12, p = .73, nor between the conditions ‘£ 10 years’ and ‘³ 20 years’, X 2 (1) = 1.80, p = .18. There was a significant difference found in gender between the conditions ‘10-20 years’ and ‘³ 20 years’, X 2 (1) = 3.87, p = .05. There was also a significant difference found in ethnicity between the ‘£ 10 years’ and ‘10-20 years’ conditions, X 2 (1) = 6.39, p = .01 and between the ‘£ 10 years’ and ‘³ 20 years’ conditions, X 2 (1) = 4.68, p = .03 but not between the ‘£ 10 years’ and ‘10-20 years’ conditions, X 2 (1) = 1.92, p = .17. There was no significant difference in level of education between the ‘£ 10 years’ and ‘10-20 years’ conditions, X 2 (1) = 1.51, p = .22, the conditions ‘£ 10 years’ and ‘³ 20 years’, X 2 (1) = .41, p = .53, nor the conditions ‘10-20 years’ and ‘³ 20 years’, X 2 (1) = 1.78, p = .18. There was also no significant difference in average age between the different conditions, H (2) = 1.56, p = .46.
There was no effect of maximum prison sentence found on the average scores on any of the symptom validity tests (Table 5).
Table 5
The effect of maximum prison sentence on the average test scores of the ASMT, MMPI-2-RF and SIMS, controlled for gender and ethnicity.
Cost-Benefit hypothesis Condition 1: £ 10 years (n=79-91) Condition 2: 10-20 years (n=138-147) Condition 3: ³ 20 years (n=114-134)
F-value p-value partial η²
ASMT1 85.81 (4.06) 85.67 (5.29) 86.05 (4.68) .53 .59 .00 MMPI-2-RF F-r-scale Fp-r-scale Fs-scale FBS-scale 79.05 (13.39) 59.52 (15.52) 52.87 (16.2) 57.81 (11.55) 80.82 (13.49) 57.04 (15.33) 53.13 (13.78) 57.49 (13.09) 82.58 (16.59) 60.72 (16.85) 56.95 (16.98) 60.42 (13.33) .88 1.29 1.51 .73 .41 .27 .22 .48 .00 .00 .00 .00 SIMS 6.64 (5.64) 7.34 (6.19) 8.86 (7.89) 2.49 .09 .01
Criminological hypothesis –
The analyses of the demographic variables, using the Chi-Square and the Kruskall-Wallis Test (Table 3), revealed that there was no significant difference in gender between the APD condition and the clinical control condition, X 2 (1) = 2.13, p = .15, nor between APD and the non-clinical control condition, X 2 (1) = 1.76, p = .18, or between the two control conditions, X 2 (1) = .11, p = .74. There was a significant difference in ethnicity between the APD condition and the clinical control condition, X 2 (2) = 12.46, p < .01 and between the two control conditions, X 2 (2) = 6.32, p = .04. There was no significant difference found in
ethnicity between the APD condition and the non-clinical control condition, X 2 (2) = 4.37, p = .11. There was a significant difference in level of education between the APD condition and the clinical control condition, X 2 (2) = 6.97, p = .03, and the non-clinical condition, X 2 (2) = 14.45, p < .01. The two control conditions also significantly differed in level of education, X 2 (2) = 7.35, p = .03. There was also a significant difference found in age between the
conditions, H (1) = 17.01, p < .01.
The results for the Criminological hypothesis can be found in Table 6. There was no effect of APD on the average scores of the ASMT, the F-r-scale, the F-p-scale, or on the F-s-scale of the MMPI-2-RF. There was a small effect of APD found on the FBS-F-s-scale of the MMPI-2-RF. Contrasts revealed that having an APD did decrease the average score on the FBS-scale compared to having a PD from cluster A or C, a medium effect was found, t (292) = 2.66, p < .01 (1-tailed) , d = .64, but not compared to having no PD, where a small effect was found, t (292) = -1.36, p = .18 (1-tailed) , d = .36. There was also a small effect of APD on the average score of the SIMS. Contrasts revealed that having an APD did decrease the average score on the SIMS compared to having a PD from cluster A or C, a medium to large effect was found, t (324) = 3.4, p < .01 (1-tailed) , d = .79, and compared to having no PD,
All found effects stayed significant after correcting for multiplicity with the Holm-Bonferroni method except for the difference found on the FBS-scale of the MMPI-2-RF.
Table 6
The effect of APD on the average test scores of the ASMT, MMPI-2-RF and SIMS, controlled for age, ethnicity and level of education.
Crimino-logical hypothesis Condition 1: APD (n=31-37) Condition 2: Cluster A or C (n=118-122) Condition 3: No PD (n=143-177)
F-value p-value partial η²
ASMT2 86.00 (2.85) 85.93 (4.47) 85.87 (5.13) .44 .65 .00 MMPI-2-RF F-r-scale Fp-r-scale Fs-scale FBS-scale 78.41 (10.77) 57.03 (13.61) 52.68 (13.29) 53.57 (8.32) 82.37 (15.81) 58.89 (17.29) 56.11 (17.79) 60.57 (13.17) 79.96 (14.06) 59.27 (15.59) 53.09 (14.33) 57.41 (12.49) 1.19 .63 1.24 4.22 .31 .53 .29 .02 .00 .00 .00 .03
SIMS 4.09 (2.84) 8.66 (7.61) 7.29 (6.23) 5.79 < .01* .03 Pathogenic hypothesis –
The analyses of the demographic variables, using the Chi-Square and the Kruskall-Wallis Test, revealed that the there was a significant difference in gender between the BPD/DD condition and the clinical control condition, X 2 (1) = 24.88, p < .01, and between the BPD/DD and the non-clinical control condition, X 2 (1) = 5.07, p = .02. There was no
significant difference in gender between the control conditions, X 2 (1) = 2.87, p = .09. There was also no significant difference found in ethnicity between the BPD/DD condition and the clinical control condition, X 2 (2) = 2.12, p = .35, or non-clinical control condition, X 2 (2) = 5.67, p = .06. The control conditions neither differed in ethnicity, X 2 (2) = 4.94, p = .09. There was a significant difference found in level of education between the BPD/DD condition and the clinical control condition, X 2 (2) = 14.57, p < .01 but not between the BPD/DD condition and the non-clinical control condition, X 2 (2) = 3.62, p = .16, or between the two control conditions, X 2 (2) = 1.74, p = .42. There was no significant difference found in age between the conditions, H (4) = 6.29, p = .18.
The results for the Pathogenic hypothesis can be found in Table 7. There was no effect of BPD/DD on the average score of the ASMT. There was a small to medium effect of BPD/DD on the average score of the F-r-scale of the MMPI-2-RF. Contrasts revealed that having BPD/DD increased the average score on the F-r-scale compared to having a PD from cluster A or C or having another Axis I disorder, a medium effect was found, t (273) = -2.43, p < .01 (1-tailed) , d = .5, and compared to having no psychiatric disorder, where a large effect was found, t (273) = 2.27, p < .01 (1-tailed) , d = .83. There was a small effect found on the
average score of the FBS-scale. Contrasts revealed that having BPD/DD increased the average score on the FBS-scale compared to having a PD from cluster A or C or having another Axis I disorder, a medium effect was found, t (273) = -2.09, p < .05 (1-tailed) , d = .51, and compared to having no psychiatric disorder, where also a medium effect was found, t (273) = 2.4, p < .05 (1-tailed) , d = .59. There was no effect of BPD/DD on the average score of the F-p-scale, or on the F-s-scale of the MMPI-2-RF. There was a small to medium effect of BPD/DD on the average score of the SIMS. Contrasts revealed that having BPD/DD increased the average score on the SIMS compared to having a PD from cluster A or C or having another Axis I disorder, a medium effect was found, t (317) = -3.5, p < .01 (1-tailed) ,
d = 56, and compared to having no psychiatric disorder, where a large effect was found, t
(317) = 3.91, p < .01 (1-tailed) , d = 89.
All found effects stayed significant after correcting for multiplicity with the Holm-Bonferroni method except for the difference found on the FBS-scale of the MMPI-2-RF.
Table 7
The effect of BPD or a DD on the average test scores of the ASMT, MMPI-2-RF and SIMS, controlled for gender and level of education.
Pathogenic hypothesis Condition 1: BPD or DD (n=36-49) Condition 2: Cluster A/C or other Axis I disorder (n=196-229) Condition 3: No disorder (n=44-46)
F-value p-value partial η²
ASMT3 85.36 (5.36) 85.79 (5.01) 86.26 (4.67) .96 .39 .00 MMPI-2-RF F-r-scale Fp-r-scale Fs-scale FBS-scale 89.33 (17.69) 64.17 (17.16) 59.92 (16.27) 65.67 (14.98) 81.29 (14.43) 59.63 (16.85) 54.56 (16.37) 58.54 (12.88) 76.39 (12.96) 55.22 (13.09) 52.35 (13.58) 57.48 (12.64) 6.84 2.75 1.69 3.05 <.01* .07 .19 .05 .05 .02 .01 .02
3 Found effects did not chance after controlling for gender and level of education, ANOVA: ASTM: F (4, 381) =
SIMS 11.88 (7.69) 7.75 (6.93) 6.11 (5.00) 8.52 < .01* .05
Since we found indications for a relationship between malingering and BPD/DD, we decided to further explore our data for this hypothesis. We examined what percentage of the participants in each condition actually scored above the cut-off score for each symptom validity test (Table 8) since, looking at the average scores in Table 7, only on the F-r-scale the BPD/DD and clinical control conditions average score is above the cut-off. The results in Table 8 show that on every scale participants in the BPD/DD condition scored above the cut-off score more often than participants in the control conditions. This difference was the biggest on the SIMS and F-r-scale of the MMPI-2-RF, which were also the tests/scales on which a relationship between BPD/DD and malingering was found. There was no substantial difference between the sexes in average scores within the BPD/DD condition on the ASTM (M men = 85.28, SD = 4.89, M women = 85.17, SD = 6.51), F-r-scale (M men = 71.36, SD = 26.37, M women = 70.31, SD = 29.01), Fp-r-scale (M men = 54.82, SD = 16.31, M women = 56.63, SD = 19.85), Fs-r-scale (M men = 53.18, SD = 15.57, M women = 52.56, SD = 14.98), FBS-scale (M men = 49.70, SD = 21.45, M women = 51.25, SD = 26.02), or SIMS (M men = 11.21, SD = 6.94, M women = 13.00, SD = 9.30). Neither was there a substantial difference found in average scores between the participants with BPD (n = 28) and the participants with DD (n = 32) within the BPD/DD condition on the ASTM (M BPD = 84.35, SD = 6.51, M DD = 86.08, SD = 3.72), F-r-scale (M BPD = 70.39, SD = 26.17, M DD = 71.69, SD = 27.85), Fp-r-scale (M BPD = 54.39, SD = 16.34, M DD = 56.09, SD = 18.09), Fs-Fp-r-scale (M BPD =53.50,
Table 8
Percentage of participants in the Pathogenic hypothesis that scored above the cut-off of the ASTM and SIMS, indicating malingering. For the MMPI-2-RF a T > 80 was used, indicating a high score. Pathogenic hypothesis Condition 1: BPD or DD (n=36-49) Condition 2: Cluster A/C or other Axis I (n=196-229) Condition 3: No disorder (n=44-46) Chi-Square (p-value) ASTM 31% 21% 14% 4.25 (p = .12) MMPI-2-RF F-r-scale Fp-r-scale Fs-scale 62% 14% 14% 35% 10% 8% 28% 6% 14% 12.41 (p < .01) 1.49 (p = .48) 2.01 (p = .37)
FBS-scale 16% 7% 10% 3.71 (p = .16)
SIMS 23% 10% 2% 11.65
(p < .01)
Discussion
Malingering can occur during forensic evaluation (Mendelson & Mendelson, 2014) and complicates correct psychological assessment, which is essential for successful treatment and safe rehabilitation. The current study has examined the motivation for malingering following the three explanatory models of Rogers (1990a, 1990b, 2008), which state that malingering can be explained by (1) cost-benefit analysis, (2) antisocial motives and/or (3) underlying psychopathology. In general we found no support for the cost-benefit analysis or antisocial motives, and some support for underlying psychopathology. We will discuss the results in depth per explanatory model.
Cost-Benefit hypothesis
There was no relationship found between the ascending length of maximum prison sentence and the prevalence of malingering, the average scores on the ASTM, validity scales of the MMPI-2-RF and the SIMS did not differ per condition. This is not in line with the expected outcome based on the Cost-Benefit hypothesis, we expected to find more
of admission to a forensic hospital. Where one could argue that the difference between the ‘£ 10 years’ condition and the ’10-20 years’ condition can not be considered substantial (which would explain why there was no difference found in prevalence of malingering between these conditions), the difference between the ‘£ 10 years’ condition and the ‘³ 20 years’ condition (which includes life sentence) can be considered substantial. It is therefore notable that even between these two conditions no difference in prevalence of malingering was found.
A possible explanation for this is that the anticipated length of the sentence (either being send to prison or being send to a forensic psychiatric hospital) is not the decisive factor when a defendant weights the costs and benefits of either outcome. Factors like the uncertainty of the length of the sanction, having to follow treatment programs and being admitted with psychiatric patients are all elements that can make admission to a forensic psychiatric hospital seem less advantageous. We tried to compensate for this beforehand, adding two more years to the average time spend in a forensic hospital, but this does not seem to outweigh the disadvantages. What also needs to be considered is the fact that a defendants own subjective estimate of his or her possible sentence may not be in line with the general jurisprudence, making the objective measure ‘maximum prison sentence’ less representative for the
expectations of a defendant. These factors (and probably many more) need to be considered and examined before ruling out the Cost-Benefit hypothesis, since it is still possible a defendant weights pro’s and con’s and choses the (in his or her eyes) most advantageous outcome. The findings in current study do therefore not reject the Cost-Benefit hypothesis completely, but rather rule out the possibility that the decision about which outcome is more advantageous is merely based on the anticipated length of the sanction.
Criminological hypothesis
In contrast to the hypothesis based on the Criminological model, participants with APD scored lower on the SIMS compared to the control conditions. There was no relationship found between the other symptom validity scales and APD. The Criminological model predicted that participants with APD would malinger more that participants without APD, based on the assumption that people with APD tend to malingering when embroiled in forensic evaluation. The results in current study are not in line with this hypothesis.
A possible explanation for this is that, as suggested earlier, the possible need to deceive others would rather be reflected in ‘faking good’ instead of in ‘faking bad’ in this specific context. By ‘faking good’ a defendant with APD can give in to the possible tendency to deceive and follow his or her lawyers advice to avoid being diagnosed with a psychiatric disorder. Note that it can even be expected that most defendants admitted to the PBC
demonstrate at least some impression management, knowing that the results of their tests will end up at the judge determining their sentence. This makes it therefore notable that
participants with APD scored lower on the SIMS compared to the control conditions. However, it needs to be emphasized that a low score on the SIMS does not indicate ‘faking good’ but solely indicates no malingering. Several studies have examined the relationship between psychopathy and ‘faking good’, but either found no (Watts et al., 2015) or a negative relationship (Verschuere et al., 2014). We explored the relationship between APD and the scores on the L-r- and K-r-scale of the MMPI-2-RF4, scales that indicate ‘faking good’, but found no relationship as well. Thus, in current study patients with APD do not seem display a distorted symptom presentation in either direction, even though impression management might be expected in the context of a forensic evaluation. Verschuere et al. (2014) explained
of psychopathy, namely low social desirability. This might also go for patients with APD and would explain the findings in current study. Patients with APD would possible not be
concerned about how others see them and would therefore not adjust their symptom presentation to manage the image other people have of them. Further research is needed to learn more about the relationship between APD and low social desirability and how this may affect the response style of patients with APD within the context of forensic evaluations.
It conclusion current study found no support for the Criminological model, patients with APD do not seem to malinger during forensic evaluations. Also, they neither seem to
demonstrate impression management in this context.
Pathogenic hypothesis
Participants with BPD/DD scored higher on the SIMS and the F-r-scale compared to the control conditions, but not on the other validity scales of the MMPI-2-RF or the ASTM. Based on the Pathogenic hypothesis, we expected to find a higher score on every validity scale. To better understand these findings we need to look at each scale in more detail and specify what they measure and how they differ from each other. First of all the F-s- and FBS-scale are more specifically focused on somatic and/or neurocognitive complaints, whereas the F-r- and Fp-r scale are more focused on psychopathological complaints (Sellbom, Toomey, Wygant, Kucharski & Duncan, 2010). This might indicate that patients with BPD/DD do not malinger somatic or neurocognitive complaints. This assumption finds support in the fact that there was also no relationship found between BPD/DD and the average score on the ASTM, which detects a specific feigned neurocognitive complaints, namely memory deficits.
This still does not explain the fact that we found an effect on the F-r-scale and SIMS, but not on the Fp-r-scale, while each indicates over-reporting of psychopathological complaints.
The difference between the two scales is that where the Fp-r-scale focusses on complaints that are rare among patients with genuine psychopathological disorders, the F-r-scale indicates complaints which are rare among the general population (Sellbom, Toomey, Wygant, Kucharski & Duncan, 2010). In other words the Fp-r-scale is designed to differentiate between actual symptoms and malingered ones within the clinical population, where the symptoms of the F-r-scale are not per definition rare in the clinical population (Sellbom, Toomey, Wygant, Kucharski & Duncan, 2010). This means that the high score on the latter might just be an indication for pathology as opposed to malingering. The same goes for a high score on the SIMS. Because of the absence of genuine symptoms in the SIMS, it makes is difficult to determine whether a high score may be related to genuine pathology or to malingering (van Impelen et al, 2014). However, despite these limitations the BPD/DD condition scored higher on the F-r-scale and the SIMS even compared to the clinical control condition, where pathology is also present. Taking for granted that part of the reported symptoms on these measures are indeed caused by genuine pathology, participants with BPD/DD still report more rare symptoms compared to other psychiatric disorders, indicating that they report more psychopathological symptoms than are genuinely present. If this is the case, why did we not find an effect on the Fp-r-scale? A possible explanation for this is that patients with BPD/DD more often embrace symptoms that are rare within the general population and less often adopt symptoms that are even atypical in the clinical population. This because perhaps the former is more relatable than the latter, even to patients with
BPD/DD. In conclusion the findings in current study partly support the Pathogenic hypothesis within the context of forensic evaluation: patients with BPD/DD seem to mainly malinger symptoms that are rare within the general population but do not seem to malinger somatic, cognitive of psychopathological symptoms that are rare within the clinical population.
cut-off scores on all the symptom validity tests compared to the control conditions. This indicates a response style that goes further than a ‘cry for help’ and actually indicates malingering. Still the questions remains why patients with BPD/DD demonstrate this
behavior. A possible reason could be for the extra attention that is received through the extra hours of treatment treating these malingered symptoms. A patient may also receive more empathy from others, having to bear so many complaints. If this is true, malingering behavior would vanish when a patient with BPD/DD is alone and no one can respond to it. Another possibility is that malingering gets provoked by underlying cognitive distortions, as described by Hopwood, Morey, Rogers and Sewell (2007). These cognitive distortions would have to provoke the need to malinger but not let the patient actually believe he or she suffers from these made up symptoms, since the awareness that the symptoms are feigned is a necessity for malingering. If specific cognitive distortions indeed induce the need to malinger, patients with BPD/DD would demonstrate malingering unregarded the absence of other people, since the need to do so would continue to persist either way. Further and innovative research is needed to see which of these suggestions finds (if any) support, for example by exploring if patients with BPD/DD demonstrate malingering behavior in the absence of others.
Limitations
A limitation of current study is the method used for classifying psychiatric disorders. In the PBC defendants are diagnosed by a psychologist and psychiatrist who base their
diagnoses on information gathered by the research team. Although the work method of the research team is thorough, clinical judgement plays a significant role in this process. Some argue that clinical judgement is more flexible and can take new or powerful information into account, which leads to better decision-making (Montgomery, 2005; Skeem et al., 2005). However, evidence found in several studies indicate that empirical actuarial methods are more
valid and reliable than clinical judgement in determining a diagnosis (Jensen-Doss et al., 2014; Croskerry, P., 2003; Dawes, Faust, & Meehl, 1989). The assignment to the conditions of the Criminological and Pathogenic hypotheses are based on the diagnoses of the behavioral experts and the conditions would therefore have been more valid and reliable if these
diagnoses were based on empirical actuarial methods.
Another limitation that needs to be taken into consideration when interpreting the findings is the fact that current research only considered cooperating, Dutch-speaking, nonpsychotic defendants who had been admitted to the PBC. This limits the generalizability to the whole Dutch forensic population because of a possible selections bias. Generally, defendants are admitted to the PBC because of the severity of the charged offence or the suspected
complexity of the pathology, which logically makes the forensic population in the PBC less representative for the whole Dutch forensic population. The findings of current study should therefore be interpreted with caution when generalized to the Dutch forensic population as a whole. It would be interesting to conduct a similar study for ambulatory forensic evaluations to see if the found effects in current study are also found in an ambulatory context.
The final limitation is that even though required number of participants is met for the total sample, this is not the case for the number of participants in each condition. The number of participants in the experimental condition of the Criminological hypothesis and the
experimental and non-clinical condition of the Pathogenic hypothesis is lower than the required number of participants given in the power analysis. A low power decreases the chance of discovering an effect and also reduces the likelihood that significant results reflect a true effect (Button et al., 2013). It is possible that we would have found more significant effects had our the sample of our conditions been the required size. The effects that were
al., 2013). A suggestion for further research is to replicate current study in a couple of years (when more defendants have been admitted to the PBC) to make sure all conditions consist of the required number of participants.
Suggestions further research
Hopwood, Morey, Rogers and Sewell (2007) supposed that malingering occurs in mental disorders known with cognitive distortions and gave BPD and DD as examples. Since
findings in current study partly support the relationship between BPD/DD and malingering, a suggestion for further research is to examine the possible role of cognitive distortions in malingering, and more specifically which cognitive distortions can be associated with this. One can assume that almost every mental disorder comes with certain cognitive distortions, which makes the explanation of Hopwood, Morey, Rogers and Sewell (2007) not very discriminative. Since the exploratory analyses of current data suggest that prevalence of malingering does not differ between patients with BPD and DD, the cognitive distortions present in both these disorders could therefore be used as a starting point. Hopwood, Morey, Rogers and Sewell (2007) described these cognitive distortions as overrepresenation of the negative aspects of a patient him- or herself, their environment and their future.
Operationalizing cognitive distortions could for example be done using the Automatic Thoughts Questionnaire-Revised (ATQ-R; Kendall, Howard, & Hays, 1989) or the more recent Cognitive Distortions Questionnaire (CD-Quest; Morrison et al., 2015). Another possibility is to measure cognitive distortions like Arntz and ten Haaf (2012) did, using BPD-patients their evaluations of others to determine if their way of thinking could be
characterized by splitting, dichotomous thinking or negativity.
for malingering of somatic or cognitive symptoms and it would be interesting to examine if similar results can be found in further research. This can be examined by, for example,
comparing the scores of malingerers on the different subscales of the SIMS or on the different validity scales of the MMPI-2RF. It is possible that the general population malingers different symptoms than the psychiatric population or that it differs per context which sort of
symptoms get malingered. More knowledge about this could be used to improve symptom validity tests and makes distinguishing genuine symptoms from malingered ones easier.
Implications
The findings of currents study have implications for both the forensic diagnostic and treatment setting. When a clinician comes during a forensic evaluation across a defendant with presumed BPD/DD, he or she should be aware of the increased risk of malingering. Since findings in current study indicate that patients with BPD/DD mainly malinger symptoms that are rare within the general population, clinicians should specifically be skeptical when a BPD/DD patient reports those kind of symptoms. The reported symptoms should be (where possible) verified using heteroanamnesis, observation reports, multiple psychological tests and symptom validation tests, to make the diagnostic conclusions in the forensic evaluation as reliable as possible. Since BPD/DD is associated with malingering of specific rare symptoms, it makes it easier for clinicians to interpret this malingering behavior of patient with BPD/DD, making it easier integrate it in their forensic evaluation. When the symptom validity tests administered during the forensic evaluation actually indicate
malingering, it is important that this gets included in the advice to the judges. This way the judges can also take this into account while reading the forensic evaluation.
disorders. A clinician should always keep an open mind, especially diagnostically speaking, during the treatment process of a patient with BPD/DD. Malingered symptoms may vanish, while actual yet unrecognized symptoms can appear. A clinician needs to be able to recognize these symptoms to make sure all the relevant (recidivism risk increasing) symptoms get monitored. When a patients presents actual symptoms and malingered symptoms, it is always a challenge for clinicians to differentiate between which symptoms are genuine and which are not (Dandachi-Fitzgerald et al., 2017). Bass and Halligan (2014) suggested a number of characteristics that might help medical clinicians to recognize malingered symptoms, that possibly can be generalized to the forensic field: the patient is an inconsistent, selective or misleading informant when asked about the complaints, the course of the complaint is atypical and does not follow the natural history of the presumed associated disorder, and the magnitude of complaint consistently exceeds objective pathology. These red flags might help to discriminate sincere symptoms from malingered ones, making risk assessment based on symptoms and other dynamic factors more reliable.
Conclusions
Mrs. Z. ended up getting diagnosed with BPD and was admitted to a forensic hospital. When she was confronted with the indications for malingering, she said she was happy someone finally saw she was capable of more than she had shown. However, her malingering behavior did not chance, which complicates her treatment and makes reintegration in the near future unlikely.
In conclusion do patients with BPD/DD, when embroiled in forensic evaluation, seem to malinger symptoms that are rare within the general population, and clinicians working in the forensic field should be aware of this during both assessment and treatment. Further research
cognitive distortions. It is essential to make forensic evaluations as reliable as possible and more knowledge about malingering will contribute to this.
References
American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR). Washington, DC: American
PsychiatricAssociation; 2000.
Arntz, A., & ten Haaf, J. (2012). Social cognition in borderline personality disorder: evidence for dichotomous thinking but no evidence for less complex attributions. Behaviour
research and therapy, 50(11), 707-718.
Bass, C., & Halligan, P. (2014). Factitious disorders and malingering: challenges for clinical assessment and management. The Lancet, 383(9926), 1422-1432.
Ben-Porath, Y. S., & Tellegen, A. (2008). MMPI-2-RF: Handleiding voor afname, scoring en
interpretatie. Nijmegen: PEN Psychodiagnostics.
Butcher, J. N. (1989). Minnesota multiphasic personality inventory. Corsini Encyclopedia of
Psychology.
Button, K. S., Ioannidis, J. P., Mokrysz, C., Nosek, B. A., Flint, J., Robinson, E. S., & Munafò, M. R. (2013). Power failure: why small sample size undermines the reliability of neuroscience. Nature Reviews Neuroscience, 14(5), 365-376.
Cartwright, A., Roach, J., Wood, H., & Wood, P. J. (2016). Mental health malingering and the fraudulent motor insurance claimant. Open Access Journal of Forensic Psychology, 8. Centraal Bureau voor Statistiek. (2016). Standaard onderwijsindeling, 2016.
Cornell, D. G., & Hawk, G. L. (1989). Clinical presentation of malingerers diagnosed by experienced forensic psychologists. Law and Human Behavior, 13, 375.
Croskerry, P. (2003). The importance of cognitive errors in diagnosis and strategies to minimize them. Academic medicine, 78(8), 775-780.
Cumming, G. (2014). The new statistics: Why and how. Psychological science, 25(1), 7-29. Dandachi-Fitzgerald, B., Merckelbach, H., & Ponds, R. W. H. M. (2017).
Neuropsychologists’ ability to predict distorted symptom presentation. Journal of Clinical
and Experimental Neuropsychology, 39, 257-264.
Dawes, R. M., Faust, D., & Meehl, P. E. (1989). Clinical versus actuarial judgement. Science,
243, 1668-1674.
Derksen, J., de Mey, H., Sloore, H. & Hellenbosch, H. (2006) MMPI-2 Handleiding voor
afname scoring en interpretatie. Nijmegen: PEN Tests Publishers.
Ellerbroek, R. (2017). Afdeling voor moeilijk onderzoekbaren. DJI Zien.
Giesen-Bloo, J., & Arntz, A. (2005). World assumptions and the role of trauma in borderline personality disorder. Journal of behavior therapy and experimental psychiatry, 36(3), 197- 208.
Gonda, X., Pompili, M., Serafini, G., Carvalho, A. F., Rihmer, Z., & Dome, P. (2015). The role of cognitive dysfunction in the symptoms and remission from depression. Annals
of general psychiatry, 14(1), 27.
Greene, R. L. (2000). The MMPI-2: An interpretive manual. Allyn & Bacon.
Heilbrun, K. (2002). Principles of forensic mental health assessment (Vol. 12). New York: Kluwer Academic Publishers.
Hoogstraten, C. H. J., & Kemperman, C. J. F. (2005). Malingering en onderpresteren bij neuropsychologische expertises. TBV–Tijdschrift voor Bedrijfs-en
Hopwood, C. J., Morey, L. C., Rogers, R., & Sewell, K. (2007). Malingering on the
Personality Assessment Inventory: Identification of specific feigned disorders. Journal
of Personality Assessment, 88(1), 43-48.
van Impelen, A., Merckelbach, H., Jelicic, M., & Merten, T. (2014). The Structured Inventory of Malingered Symptomatology (SIMS): A systematic review and meta-analysis. The
Clinical Neuropsychologist, 28(8), 1336-1365.
Jelicic, M., Merckelbach, H. L. G. J., Giesbrecht, T., & de Ruiter, C. (2008). Structured inventory of malingered symptomatology. Forensisch psychodiagnostisch
gereedschap: malingering psychopathie en andere persoonlijkheidstrekken, 151-160.
Jensen-Doss, A., Youngstrom, E. A., Youngstrom, J. K., Feeny, N. C., & Findling, R. L. (2014). Predictors and moderators of agreement between clinical and research
diagnoses for children and adolescents. Journal of consulting and clinical psychology,
82(6), 1151.
Kendall, P. C., Howard, B. L., & Hays, R. C. (1989). Self-referent speech and
psychopathology: The balance of positive and negative thinking. Cognitive Therapy
and Research, 13, 583–598.
Kuijck, Y. van (2005). De vraag naar en de rechterlijke weging van de gedragskundige expertise. Delikt en delinkwent, 35(6), 627-641.
Lakens, D. (2013). Calculating and reporting effect sizes to facilitate cumulative science: A practical primer for t-tests and ANOVAs. Frontiers in Psychology, 4:863.
Mendelson, G., & Mendelson, D. (2014). Legal and psychiatric aspects of malingering.
Journal of Law and Medicine, 1, 28-34.
Merckelbach, H., & Smith, G. P. (2003). Diagnostic accuracy of the Structured Inventory of Malingered Symptomatology (SIMS) in detecting instructed malingering. Archives of
