I am honored to be invited to provide this editorial for the Journal of the American Academy of Psychiatry and the Law. Since the mid-1980s, I have immersed myself in the study of malingering and related response styles. Less publicized have been my extensive consultations on malingering cases, often complex and frequently contested. This editorial provides a constructive opportunity to enumerate four enduring pitfalls in assessments and determinations of malingering in hopes of improving forensic practice.
For examinees markedly overreporting their psychological impairment, “malingering”1,2 must be clearly distinguished from “feigning.” Both clinical constructs share the intentional gross exaggeration or fabrication of symptoms (i.e., mental and physical disorders, and cognitive impairments). They differ fundamentally, however, with respect to motivation. Malingering requires the person to be “motivated by external incentives” (Ref. 2, p 726). In contrast, feigning is a broader construct that includes malingering, factitious disorders without an external incentive, and other forms of dissimulation with unknown motivation.3
Reliance on Fatally Flawed Indicators
The Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III),1 revolutionized the diagnostic nosology with its formal establishment of explicit inclusion and exclusion criteria. It also introduced malingering as a V code. In its most recent revision, DSM-5 provided four indicators for when “malingering should be strongly suspected” (Ref. 2, p 727): medicolegal context, antisocial personality disorder, uncooperativeness, and marked discrepancies. Despite their superficial similarities, these V code indicators do not represent formal inclusion criteria. More than three decades ago, the limitations of these indicators were easily recognized and were first published in The Journal (then The Bulletin).4 Although the indicator of marked discrepancies was slightly useful, the other three indicators were completely unhelpful. Uncooperativeness, in fact, is an erroneous indicator of malingering. Using the only empirical data prior to DSM-III, genuine, mostly psychotic patients were more than twice as likely to be uncooperative than malingerers.4
The DSM indicators were doomed to failure from the onset because they confused common characteristics with distinguishing characteristics.5,–,7 Medicolegal evaluations include all forensic examinees, whereas antisocial personality disorder remains commonplace in criminal-forensic referrals. To underscore the fundamental flaw of applying common characteristics, virtually all malingerers have opposable thumbs, but it would be a senseless exercise to rely on this common variable as a malingering indicator.8
Forensic practitioners do have a useful and easily accessible alternative. For feigned mental disorders, the Miller Forensic Assessment of Symptoms Test (M-FAST)9 should be considered as a time-efficient, 25-item screen. The M-FAST is well-validated and can be easily administered with minimal preparation.10 As an initial screen, its cut score is intentionally set low to retain many genuine patients as well as possible feigners; therefore, it should not be used by itself as indicative of feigning.9 Moreover, practitioners should focus on the M-FAST total score,10 which is much more effective than its specific scales.11
Believing in Malingering Tests
Many seasoned forensic practitioners have been convinced inaccurately that psychological measures directly assess malingering. This misapprehension is entirely understandable, given that a recent combined literature search of the Medline and PsycInfo databases (both accessed May 4, 2021) identified 46 peer-reviewed articles citing “malingering test” or “malingering scale.” Moreover, the term malingering is actually included in several test names, such as the widely used Test of Memory Malingering (TOMM),12 the Malingering Probability Scale (MPS),13 and the Structured Inventory of Malingered Symptomatology (SIMS).14
No measure has been developed, or likely will be developed, to capture the often complex motivations underlying decisions to malinger. As illustrated by Rogers and Pan,8 many factors could be considered by examinees in deciding whether to malinger insanity. For example, they may consider the odds of deceptively garnering favorable expert evidence that supports an insanity acquittal by concealing their malingering while appearing extremely impaired. Failed efforts to malinger insanity typically involve a full admission to the crime(s) and an intentional effort to foil justice, both likely factors in the almost certain convictions and subsequent sentencing. Other motivations for malingered insanity may reflect an examinee's estimation of the practitioner's expertise (e.g., potentially gullible) or just pure desperation without rational thinking. No psychometric measure can begin to wrestle with these complex motivations, which may continue to evolve due to changing circumstances (e.g., newly discovered DNA evidence).
To re-emphasize a critical point, no psychometric measures or scales can assess the varied motivations that may underlie malingering. Effective feigning measures have been developed and validated, but they should never be misconstrued as measures of malingering. The motivation to malinger must be assessed for each examinee and simply not assumed. One line of inquiry can focus on examinees' perceptions of their options and their estimations of achievable goals.
All Numbers Reflect Equal Precision
Numbers appear to reflect accuracy and certainty, which is certainly not always the case. For more than a century, Heinz touted “57 varieties” as an utterly false, but apparently effective, marketing tool. As recently described in Smithsonian Magazine,15 the numbers “5” and “7” simply represented Heinz's and his wife's lucky numbers. Unlike the “57” varieties, numbers on feigning measures are not simply conjured, but are based on empirical research, albeit displaying a wide range of scientific rigor. That said, the replication crisis in social sciences (i.e., the frequent failures at cross-validation) highlights the sobering reality16 that has been described as a “crisis of credibility” (Ref. 17, p 1084). Both positively and importantly, the implementation of rigorous standards for differentiating feigned from genuine responding (i.e., technically known as “effect sizes”) represents a major step forward over the last decade in ensuring the accuracy of feigning classifications.5,7
Regarding the matter of certainty, any particular number on psychometric measures should not be interpreted as a single score but rather as a likely range of scores. This point has been specifically affirmed by the American Academy of Psychiatry and the Law (AAPL) in an amicus brief provided to the U.S. Supreme Court18 in Hall v. Florida.19 From this brief, the Court cited with approval, “[I]t is standard psychometric practice to report the ‘estimates of relevant reliabilities and standard errors of measurement’ when reporting a test score” (Ref. 19, p 722). Again, the Court relied on this brief in determining that the standard error of measurement must be applied. It ruled that the 95 percent confidence interval must be considered in evaluating intellectual abilities in death penalty cases, relying largely on “the unanimous professional consensus” (Ref. 19, p 722). Whereas unanimity in such matters is often unattainable, the brief, authored by the American Psychological Association, was fully endorsed by the American Psychiatric Association, AAPL, and the National Association of Social Workers.
Many readers may be questioning the relevance of Hall to the accuracy of scores on feigning measures. In this regard, the brief applied generally to professional practice, openly acknowledging that “every standardized test score has a ‘standard error of measurement’ (‘SEM’) that reflects the reliability (precision) of scores from the test” (Ref. 19, p 6). As noted, the brief asserted, and the Court ruled, that a 95 percent confidence interval must be applied.
Practitioners without extensive research training may be justifiably skeptical about specific numbers and the requisite need for 95 percent confidence intervals. On this point, multiscale inventories, such as the Minnesota Multiphasic Personality Inventory (MMPI) in its various versions, are particularly vulnerable to imprecise estimates of feigning and clinical scales, perhaps mostly due to their diverse professional applications and the problematic expectation that a single test can accurately cover different response styles, core personality features, and patterns of psychopathology. As a result, it might be best to conceptualize feigning indicators for these inventories as potentially useful but error-prone estimates rather than precise measurements. Because it usefully reported data on a large clinical sample including inpatients, the MMPI–2–Restructured Form (MMPI–2-RF)20 is an instructive example. On several feigning scales (i.e., F-r and Fp-r), the 95 percent confidence intervals are approximately 20 points (i.e., 19.6); thus, a moderate elevation of 70 could be just average (i.e., 50 or the 50th percentile) or markedly elevated (i.e., 90 or virtually the 100th percentile).
Much more precise estimates of malingering can be easily achieved with interview-based, single-purpose feigning measures. The two reasons for this finding are straightforward. First, the answers and scoring are completed by experienced forensic practitioners. Second, the scales are specialized; they focus primarily, if not exclusively, on response styles. A well-validated combination of two measures for feigned mental disorders includes the M-FAST9 as a screen and the Structured Interview of Reported Symptoms-2 (SIRS-2)21,22 as a comprehensive measure. For the latter, no further description is provided to minimize any potential conflict of interest.
Betting All on a Single Method
The poker metaphor of “going all in,” while having appreciable merit, lacks the same outcome as mistaken conclusions about malingering. In poker, if wrong, the players shoulder their own personal losses. In forensic evaluations, however, the examinee may suffer potentially life-altering consequences because of the faulty but well-meaning actions of a forensic practitioner. This gambling metaphor is provided as a pointed counter-argument to address a growing concern about “cutting corners” in malingering assessments, especially in high-volume settings.8
To establish better practices, Rogers and Pan8 strongly recommend a multi-method, multi-strategy approach to the assessment of malingering. Multi-method evaluations typically involve three valuable components: clinical interviews, interview-based specialized feigning measures, and self-administered measures (e.g., inventories) with embedded feigning scales. First, clinical interviews are clearly essential for both evaluating noncredible presentations as well as establishing the motivation for malingering. Second, specialized measures generally provide the most accurate classifications of feigning because they are based on well-validated detection strategies described in the next paragraph. Third, self-administered measures typically afford a broader viewpoint on response styles that include marked inconsistencies and under-reporting of symptoms and impairment. For instance, marked inconsistencies (e.g., a haphazardly completed inventory) often could be mistaken for feigning if this response style is not systematically ruled out.
Detection strategies for feigned mental disorders have been formally operationalized and empirically tested since 1984.23 They represent conceptually based methods of systematically differentiating between two or more response styles (e.g., feigning versus genuine responding) in relevant populations (e.g., forensic patients).5,–,7 For example, a “rare-symptom strategy” utilizes symptoms that are very uncommon in genuine clinical populations but often endorsed by feigners, who are unaware of this. A multi-strategy approach typically incorporates at least four detection strategies, some of which are unlikely to be present (e.g., rare-symptom strategy), whereas other strategies differ in magnitude (e.g., feigners reporting too many symptoms with extreme severity).7,24 With this brief overview, two points are worthy of consideration. First, forensic practitioners should acquire a solid foundation regarding the unlikely and amplified detection strategies for feigned mental disorders. In a matter of a few hours, they will be able to substantiate competently and confidently the bases of their opinions relevant to feigning and malingering. Secondly, this competence will help ensure thorough and balanced assessments of feigned mental disorders.
Conclusions
Forensic psychiatrists and psychologists draw much of their expertise from highly valuable training and mentoring. In a fashion perhaps similar to that of the law,25 the intergenerational continuity of forensic knowledge and methods sometimes hampers the acceptance of more recent, empirically tested advances. Specific to feigned mental disorders, forensic practitioners may wish to ask introspectively how their assessment methods have changed from the earlier generation of their mentors. Focusing on my specialty in forensic psychology, I am deeply concerned that troublesome pockets of professional complacency may further diminish unhurried and comprehensive assessments of feigned mental disorders. In closing, I deeply appreciate this opportunity to examine critically four major professional pitfalls in the rigorous evaluation of forensic examinees for feigned mental disorders, with direct forensic relevance to malingering.
Footnotes
Disclosures of financial or other potential conflicts of interest: Dr. Rogers receives royalties for the SIRS-2 mentioned in this editorial.
- © 2021 American Academy of Psychiatry and the Law