Assessment of Malingering with Simulation Designs: Threats to External Validity

Law and Human Behavior, Vol. 22. No. 3. 1998

Assessment of Malingering with Simulation Designs:Threats to External Validity

Richard Rogers1 and Keith R. Cruise1

Comprehensive forensic evaluations are predicated on the accurate appraisal ofresponse styles that may affect evaluatees' clinical presentation and experts' conclusionsassociated with psycholegal issues. In the assessment of malingering, forensic expertsoften rely heavily on standardized measures that have been validated exclusively viaanalogue research. While such research augments internal validity, the threats to externalvalidity are readily apparent. As the first study of these threats, type of incentive (positiveversus negative), context (a familiar versus unfamiliar scenario), and relevance to theparticipants was investigated systematically with a between-subjects factorial design. Asample of 231 undergraduates was asked to either (a) feign major depression and givenon easily understood description of this disorder or (b) serve as controls respondinghonestly. They were administered a brief measure of psychopathology (HopkinsSymptom Checklist; Demgatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974) and arecent screen for malingering (Screening Inventory of Malingered Symptoms or SIMS;Smith, 1992) in 1 of 18 experimental conditions. Results suggested that incentive hada main effect on the SIMS. More specifically, simulators under negative incentivesappeared more focused in their feigning; they produced more bogus depressedsymptoms, but fewer symptoms unrelated to depression. Interactions were also observedbetween context and incentive, and context and relevance. Implications of these resultsare explored for both analogue research on malingering and current forensic practice.

A cornerstone of forensic evaluations is the effective evaluation of malingeringand related response styles. Given the adversarial nature of most forensic assess-ments, psychologists and other mental health professionals must address the po-tential for malingering for nearly all psycholegal issues (Rogers & Mitchell, 1991).Recent survey data (Rogers, Salekin, Sewell, & Goldstein, 1996; Rogers, Sewell, &Goldstein, in press) suggest that the prevalence of malingering varies widely acrossforensic settings, but is likely to compose approximately one sixth of all forensiccases. With malingering as both a critical and common issue in forensic assessments,the empirical bases of these determinations are crucial.

1University of North Texas, Dept. of Psychology, P.O. Box 311280, Demon, TX 76203-1280.

273

OI47-7307/<18,«M)0.0273$I500/I D 1998 American Psychology-Law Society/Division 41 of the American Psychological Association

During the last decade, the clinical literature on malingering has burgeonedwith hundreds of studies on the detection of dissimulation (for reviews, see Berry,Wetter, & Baer, 1995; Franzen, Iverson, & McCracken, 1990; Rogers, Harrell, &Liff, 1993; Schretlen, 1988). A unifying thread to this diverse literature is the almostexclusive reliance on analogue designs. Commonly described as "simulation" stud-ies, research participants are randomly assigned to either feigning or control con-ditions. Their results are then compared to clinical comparison samples to parallelreal-world applications: feigning versus genuine disorders. Studies vary with respectto their forensic application; some research addresses specific legal issues, such ascomperency to stand trial (Gothard, Viglione, Meloy, & Sherman, 1995), personalinjury Lees-Haley, English, & Glenn, 1991) and insanity (Schretlen & Arkowitz,1990). while others (Berry et al., 1996; Graham, Watts, & Timbrook, 1991) haveonly indirect relevance to forensic issues.

Rogers and Cavanaugh (1983) voiced the first concerns regarding our over-reliance on the simulation design and its generalizability to clinical settings andactual malingerers. They described the "simulation-malingering" paradox as a situ-ation in which persons are asked to comply with instructions to fake in order tostudy persons who fake when asked to comply. Ensuing investigations of the MTest (Beaber, Marston, Michelli, & Mills, 1985) have sought to underscore the limi-tations of simulation research. While employing a simulation design, studies of theM Test generally have yielded positive results (Rogers et al., 1993). However, whenapplied to actual malingerers in a known-groups comparison, the results are com-paratively modest with classification rates from 30.8% to 72.8% (M of four studieswas 53.3%; Smith, 1997) with poor generalizability to forensic practice. The con-vergence between simulation and known-groups comparisons has yet to be exam-ined for most psychometric measures.

Recent research has revealed an important consideration regarding the clinicalusefulness of cutting scores derived from simulation studies of malingering. Withthe original MMPI, Berry, Baer, and Harris (1991) found wide-ranging cut scoresfor malingering. Likewise, Rogers, Sewell, and Salekin (1994) examined 14 MMPI-2studies that produced highly disparate results both for optimum cutting scores andthe resulting hit rates. For example, F-scale cut scores ranged from 7 to 30,2 withhit rates from 72% to 95%. The primary cause for these marked fluctuations ap-pears to be differences in samples and design. Of particular interest, cut scores onforensic issues ranged widely from the lowest (i.e., a simulation study with personalinjury claimants; Lees-Haley, 1991) to average (e.g., inmates asked to simulate men-tal illness as part of a pretrial evaluation; Iverson, Franzen, & Hammond, 1995)and moderately high (e.g., students faking insanity or fraudulent disability; Bagby,Rogers, Buis, & Kalemba, 1994). With respect to design, studies varied widely intheir instructional sets. Some studies offer only very general instructions withoutpresenting a specific scenario or providing any incentive for a convincing presen-tation (e.g., Timbrook, Graham, Keiller, & Watts, 1993). Among those studies with

2Interpretation of these cut scores is confounded by an equally marked range on F for a normativesample of patients with mental disorders; Greene (1997) found that 90% of clinical samples fell between2 and 24.

274 Rogers and Cruise

specific instructions, marked differences were observed with respect to both sce-narios and incentives.

The purpose of the current study was to investigate how differences in simu-lation design might constitute threats to external validity3 (Kazdin, 1992). Partici-pants are unlikely to put forth a concerted effort if they cannot identify with thescenario, do not appreciate the potential for negative incentives, or are disengagedfrom the research study. Given the myriad possibilities, we selected three designissues to investigate: context, incentive, and relevance. Rogers (1997) hypothesizedthat the contexts (scenarios) for feigning vary in familiarity to participants and thatthis would likely affect performance. In addition, nearly all simulation studies haverelied exclusively on positive incentives (e.g., extra credit) as an analogue for suc-cessful malingering (for a rare exception, see Patrick & Iacono, 1986). In contrast,most real-world settings include very negative consequences for unsuccessful feign-ing that may entail the cessation of current services and the imposition of legalsanctions. Finally, few studies have addressed the relevance of the research to par-ticipants. Simply put, the external validity of simulation research is in jeopardy ifparticipants do not understand how its results are related to them or their families.

METHOD

Participants

Participants were recruited from three undergraduate psychology classes atthe University of North Texas. In accordance with Institutional Review Board re-quirements, all participants gave written informed consent. The sample was com-posed of 231 participants (71 men and 160 women) with a mean age of 22.54 (SD= 6.32). With respect to racial composition, the sample was 17 (7.4%) AfricanAmerican, 167 (72.3%) Anglo American, 19 (8.2%) Hispanic American, and 28(12.1%) other/unreported.

Design

The study is a 3 (context) x 2 (incentive) x 3 (relevance) between-subjectsfactorial design. The three conditions for context are no description, applicable (po-tential flunking from school), and inapplicable (arrested on a felony charge). Thetwo conditions for incentive are positive (class credit and possible financial reward)and negative (loss of class credit and public posting of unsatisfactory performance).The three conditions for relevance are none, real-world implications (financial bur-

3As observed by an anonymous reviewer, "threats to external validity" in simulation designs could alsobe construed as either "ecological validity" or "limits to construct validity." It may also be referred toas "validity generalization" (American Psychological Association, 1985). We selected "threats to externalvalidity" in keeping with past malingering research. According to Kazdin (1992, pp. 25-33), experimentalinstructions and other stimulus characteristics may limit the generality of results. In this light, veryatypical scenarios or use of incentives that do not correspond with clinical realities pose threats toexternal validity. In addition, actual malingerers are strongly invested in the outcome of psychologicalassessments. If participants do not appreciate the relevance of their involvement, the generality ofsubsequent findings is severely constrained.

Malingering with Simulation Designs 275

den and decrease in health services), and challenge (questioning participants' abilityto "beat" the test)

Instructional Sets

All participants in simulation conditions received a general set of instructionsthat described their goal of feigning major depression and provided in nontechnicallanguage a description of its major features. General simulation instructions arereproduced:

You are asked to complete the following tests and pretend that you have a majordepression. By major depression, we mean much more than just being sad, lonely or blue.Instead, major depression is a serious mental disorder in which the person almost alwaysfeels depressed, down, or hopeless. Many times, persons with major depression experienceno pleasure, have lost interest in nearly everything, feel worthless, and have thoughts abouttheir own death. Many also have problems with eating, sleeping, concentration, and energylevel.

There are many mental disorders. Remember your job is to appear as if you have amajor depression.

Context

Specific instructions were presented to provide either a generally applicableor inapplicable scenario for the simulation. For college students, academic failureappeared to constitute a crisis that was potentially applicable to many students. Forthe inapplicable scenario, we deliberately choose a forensic case (i.e., a sentencingevaluation with treatment options). Even with the inapplicable scenario, we did notopt for extreme cases that would likely be devoid of relevance for most students(e.g., pleading insanity; Bagby, Gillis, & Dickens, 1990). Instructions on contextconsisted of the following:

Applicable. For this study, I would like you to imagine that you worked very hard in acourse but are having trouble with the exams. Just before the midterm exam, you brokeup with your boyfriend or girlfriend. You felt so lousy you didn't even go to the exam.Now, it looks like you might flunk the class. One way out of this mess would be if theuniversity psychologist reported that you had a major depression that prevented you fromtaking the exam.

Inapplicable. For this study, 1 would like you to imagine that you were arrested forpossession of marijuana and dangerous driving. You were stopped by the police for drivingerratically on Eagle Drive, nearly hitting a pedestrian. When stopped, you were found tobe smoking a joint. Given the evidence against you, there is no way you can beat thecharges. One way out of this mess would be if the court psychologist reported that youhad a major depression and that you needed treatment more than prosecution.

Incentives

Participants were provided with either positive or negative incentives. Theseinstructions are reproduced. As a method of not compromising future dissimulationresearch, yet not causing unnecessary embarrassment, fictitious names were postedfor the "failures."

276 Rogers and Cruise

Positive. If you succeed at faking depression, you will have several rewards. First andforemost, you will be given extra credit. Even though this study was done on class time,we will give you an hour of extra credit (only if you are successful). Second, we will postthe names of the 10 most successful students. Third, the most successful student will begiven a $50 reward.

Negative. Even though this study was done on class time, we will give you an hour of extracredit for your participation. The only exception is if you are unsuccessful at faking. Inthis case, you won't get the extra credit. In addition, students that are really bad at fakingwill be identified. We will post the names of the 10 least successful students. All students,except these 1O, will be eligible for a $50 drawing.

Relevance

Research participants either received no information about relevance or oneof these two instructional sets. According to Rogers (1997), actual malingerers arepoignantly aware of how successful feigning is likely to affect their lives. In mostsimulation studies, participants are distanced from the relevance of the research.As an exploratory investigation, we sought to increase the relevance of the researchby emphasizing either its real-world implication or challenging participants to "beat"the tests:

Real-world Implications. We need you to take this study very seriously. Billions ofhealth-care dollars are wasted each year on persons that are freeloading on the system.Right now, your family's taxes are being squandered on healthy persons that are living offyour money. Tax dollars that could be spent on your education and reduced tuition arebeing blown on free loaders. Help us to help you through this study.

Challenge. Can you beat the test? The test is set up to catch people trying to fake a mentaldisorder. Only students who are smart enough and shrewd enough are likely to avoidbeing caught. We know that most students won't be able to do it. We are challenging youto see if you can one of those few who are skillful enough to beat the test.

Dependent Measures

One methodological constraint in the use of large numbers of undergraduatestudents was the need to complete data collection in a single class period (50 min).With this constraint in mind, two relatively brief measures were selected: the HopkinsSymptom Checklist (HSCL; Derogatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974)and the Structured Interview of Malingered Symptomatology (SIMS; Smith, 1992).

HSCL

The HSCL is a 52-item symptom checklist; variations of the HSCL have beenwidely used in clinical research (e.g., Castaneda, Lifshutz, Galanter, & Franco, 1994;Green, Walkey, Taylor, & McCormick, 1989; Matt, Dean, Wang, & Wood, 1992).As a predecessor to the Symptom Checklist-90-Revised (SCL-90-R; Derogatis,1977), the HSCL is representative of the SCL-90-R and similar symptom checklistsutilized in clinical practice. Because of the large number of research participants,the affordability of the HSCL as a nonproprietary measure was also a considerationin its selection.

Malingering with Simulation Designs 277

SIMS

The SIMS is a 75-item true-false test designed to screen for malingering ofseveral conditions, including mood disorders. Initial research (Smith, 1992) has sug-gested its potential utility for identifying persons feigning both psychopathology andcognitive deficits. Recently, Rogers, Hinds, and Sewell (1996) examined its clinicalutility in a forensic setting with dually diagnosed adolescents and concluded thatits total score (>16) was useful in identifying cases of potential malingering, witha positive predictive power of .87, but less useful in classifying genuine patients,with a negative predictive power of .62.

Procedure

Participants were recruited from undergraduate classes in psychology. Thegeneral purpose of the study was explained to potential participants. With writteninformed consent, participants were given packets with the instructions and experi-mental measures to complete. Packets with experimental or control conditions wereinterleaved so that each condition was represented equally; packets were distributedsequentially to minimize any potential bias in group assignment. Packets were rec-ollected upon completion.

RESULTS

Prior to any investigation of the hypotheses, the data were examined withrespect to statistical power. As a factorial design, the cell frequencies averaged11.55. One consideration was whether larger cell frequencies (>20; Kraemer &Thiemann, 1987) are more likely to yield significant results. In this regard, posthoc power analysis revealed for the SIMS factors .52 for relevance, .87 for incentive,and .47 for context.

As a manipulation check, we compared simulators across all conditions to acontrol group receiving standard (i.e., honest instructions) on the HSCL. As ex-pected, simulators produced higher HSCL scale elevations than controls, Wilks'lambda =.71, F(6, 212) = 14.38, p = .0005.

Multivariate F tests ( 3 x 2 x 3 MANOVA) using context, incentive, and relevanceas independent variables were computed separately on test scales from the HSCL andSIMS. Multivariate F tests consistently were nonsignificant for HSCL without any evi-dence of nonsignificant trends (i.e., all ps > .30). Inspection of the HSCL clinical scalesrevealed marked elevations across all simulation conditions. Overall, M score for de-pression was 3.24 (SD = .64) on a 4-point scale; the remaining scales were less elevated,ranging from 2.31 (SD = .59) to 2.98 (SD = .60). In contrast, multivariate F tests forthe SIMS revealed one significant main effect (incentive) and one significant interaction(incentive x relevance). For the incentive condition, Wilks' lambda = .89, F = 2.24, p= .01,n|2 = .11. For the interaction of incentive x relevance, Wilks' lambda = .85, F= 1.63, p = .05, n2 = .15. Cell means and standard deviations are presented in

278 Rogers and Cruise

Malingering with Simulation Designs

Table 1. Means and Standard Deviations for SIMS Scales x Factors on Dis-simulation of Depression

279

Condition AF LI AM Total

ContextualityNo scenario

Applicable

Inapplicable

RelevanceNo mention

Real world

Challenge

IncentivePositive

Negative

GeneralControl

6.38(4.04)6.49

(3.35)6.65

(4.02)

6.51(4.02)6.38

(3.52)6.64

(3.91)

6.86(4.01)6.15

(3.51)

2.25(1.42)

11.00(2.18)11.22(1.87)11.20(2.32)

10.80(2.40)11.29(1.62)11.33(2.09)

10.77(2.17)11.51(1.89)

6.25(2.01)

3.85(3.67)4.10

(3.51)4.70

(4.10)

4.39(3.98)4.41

(3.28)3.84

(4.02)

4.56(3.91)3.87

(3.60)

.83(1.19)

8.34(1.95)7.97

(1.68)8.19

(2.02)

8.09(1.97)8.30

(1.82)8.11

(1.89)

8.41(2.02)7.92

(1.71)

6.50(.67)

6.70(4.18)6.83

(3.84)6.48

(4.34)

6.41(4.25)7.07

(3.69)6.54

(4.40)

7.23(4.35)6.11

(3.80)

1.58(.51)

36.27(12.97)36.63

(10.67)37.22

(14.01)

36.19(13.96)37.45

(10.93)36.46

(12.85)

37.83(13.73)35.56

(11.28)

17.42(3.92)

Note. For the SIMS scales, N = Neurological Impairment, AfP = Psychosis, LI = Low Intelligence, Am = Amnesia.

= Affective Disorders,

TABLE 2. Univariate F Tests on the Incentive Main Effect on theScreening Inventory of Malingered Symptoms (SIMS)

Incentive

SIMS Scales Positive Negative Cohen's d

NAFPLIAMTotal

6.8610.774.568.417.23

37.83

6.1611.513.867.926.11

35.56

1.546.891.543.343.641.68

.22

.01

.22

.05

.06

.20

.19

.36

.19

.26

.27

.18

Note. All M values and SDs are summarized in Table 1, selected Afs areprovided here simply to assist interpretation. Degrees of freedom for allF ratios are 1, 189. For the SIMS scales, N = Neurological Impairment,AF = Affective Disorders, P = Psychosis, LI = Low Intelligence, Am =Amnesia. Total = the sum of SIMS scales.

Table 1. To understand the significant findings for incentive, univariate F tests werecomputed for positive and negative incentives. As reported in Table 2, participantswith negative incentives appear to be much more targeted in their presentationthan those with positive incentives. More specifically, the negative incentives appearto be responsible for a greater elevation on mood symptoms (AF) than positiveincentives, but had lower elevations on other scales, with significantly smaller scores

280 Rogers and Cruise

Fig. I. Significant interactions of context by relevance for participants feigning majordepression on Scales AF and LI of the SIMS.

on low intelligence (LI) and a nonsignificant trend (p = .06) on amnesia (AM). Figure1 presents the significant interactions of context by relevance for Scales AF and LI.

The interaction of context x relevance was evaluated by Roy-Bargman step-down analysis. This procedure allows for independent analysis of a theoreticallyimportant dependent variable while treating all other dependent variables as co-variates (Tabachnick & Fidell, 1996). Two scales were significant: (a) Affective Dis-orders (AF), stepdown F (1, 189) = 6.89, p < .01; and (b) Low Intelligence (LI),stepdown F (1, 188) = 5.95, p - .02. Duncan multiple range tests for AF yieldedno significant differences. However, the same analysis for LI indicated that whenchallenged participants were provided an applicable context their scores were lesselevated than those offered no scenario.

Duncan multiple range tests (alphaFW = .05) for context x incentive on LIscale (see Figure 2) revealed on the stepdown analysis that participants with positive

281

Fig. 2. Significant interactions of context by incentive for participants feigning majordepression on LI Scale of the SIMS.

incentives produced fewer bogus symptoms with the applicable scenario than thosereceiving no scenario or an inapplicable scenario. The presence of a negative in-centive also produced fewer bogus symptoms for participants given no scenario oran inapplicable scenario. One explanation of these findings is that participants wereinfluenced significantly by some form of "reality check." In the case of the LI scale,either the applicability of the scenario (flunking a class) or the threat of externalsanctions provided this reality check. Conversely, the attraction of even modest re-wards (class credit and the possibility of $50) without any risk (no negative sanc-tions) appeared to encourage less cautious feigning when the scenario wasinapplicable or not given at all.

DISCUSSION

The misclassification of malingering carries incalculable costs in forensic evalu-ations. Sanguine attempts to offer precise estimates of malingering (Mossman &Hart, 1996) belie the terrible tragedy of mistakes (Rogers & Salekin, 1996). Thedevastation to defendants or plaintiffs of being falsely accused of malingering byforensic experts is unimaginable. Conversely, undetected cases of malingering wreaktheir own havoc for survivors of unpunished crimes and bloated insurance premiums.

Rogers (1997) recommended that measures of malingering be validated witha combination of simulation design and known-groups comparison. In the latter,measures developed with analogue research are cross-validated with actual malin-gerers. A convergence of findings between the two research methods offers thestrongest evidence of accurate determinations because of their respective strengths:simulation design (internal validity) and known-groups comparison (external valid-ity). This combined approach has proved successful with the Structured Interviewof Reported Symptoms (SIRS; Rogers, Bagby, & Dickens, 1992) and has promisingresults with the Personality Assessment Inventory (PAI; Morey, 1991) in an initial

Malingering with Simulation Designs

study (Rogers, Ustad, & Salekin, 1998). Importantly, the combined approach mini-mizes the difficult and enduring problem faced by simulation designs, namelythreats to external validity.

The current data represent the first systematic attempt to investigate simul-taneously several potential threats to external validity in the assessment of malin-gering. The most striking finding was that the type of incentive appeared to affectsimulators' presentations. In the face of a very mild negative incentive, simulatorsappeared to become more focused in their presentation. More specifically, theyachieved a higher elevation on the AF scale, but their other SIMS scales weregenerally lower. With a greater opportunity to prepare and the use of more exten-sive measures, more pronounced differences may be observed.

Methodological Considerations

A potential design limitation is that participants were not debriefed regardingtheir knowledge of the instructions (see Rogers, 1997). Our rationale was that ourparticipants, unlike those in many malingering studies, had continued access to theinstructions during the 30-45 min of data collection. Still, it is conceivable thatsmall numbers of participants were disinterested in the study and its incentives;noncompliance would likely obscure significant results.

The selection of applicable and inapplicable scenarios presents an importantdilemma. Although differences in applicability must be substantial, any extreme ma-nipulations (e.g., ax murders for "inapplicable") is likely to undermine the credibilityof the research and distort its findings. We selected as "inapplicable" dangerousdriving under the influence of marijuana. We recognized that a considerable numberof students have experimented with cannaboid substances, but surmised that mostwould circumscribe their use to private residences. As a general rale, applicablescenarios (e.g., course failure) likely cany fewer penalties than inapplicable scenar-ios (e.g., third-class felony).

The most troublesome limitation of analogue research on malingering is theability to approximate negative consequences found in real-world applications. Thenegative incentives applied in this case offered, at worst, mild but temporary em-barrassment. In adversarial contexts, the negative incentives for unsuccessful ma-lingering likely include loss of social services and the imposition of severe sanctions.Ethical principles preclude any approximation of these severe consequences.

Forensic Implications

The current findings are far from conclusive on the matter of incentives, whichare substantially more complex than any simple dichotomization. The next step willbe the investigation of combined incentives. In forensic practice, defendants arefaced concurrently with both positive and negative incentives. These incentives dif-fer substantially by the nature of the case. In many criminal cases, the goal of ma-lingering requires extreme presentations: For example, only grossly psychoticfabrications have any appreciable chance of being successful in achieving a verdict

282 Rogers and Cruise

of "not guilty by reason of insanity." In most cases, a failed attempt at malingeringis tantamount to a conviction, because the defendant must admit to the commissionof the criminal act in asserting an insanity defense. Even so, the magnitude of thenegative incentives may be highly variable. When the evidence of guilt is over-whelming (e.g., a videotaped confession), defendants may believe they have verylittle to lose by feigning.4 In civil cases, a full spectrum of feigning may be observed,from minor exaggerations of a genuine disorder to blatant and improbable fabri-cations. As a general rule, the greater the undeserved reward (e.g., fabrication ofgross impairment) is coupled with increased risk (i.e., the more extreme the feign-ing, the greater the likelihood of detection). Especially in forensic cases, incentivesmust be examined within a specific legal context in order to understand the rele-vance of the results,

The present study contrasted an applicable scenario (academic failure) witha less applicable forensic scenario.5 Pending further studies, the current data suggestthat the lack of an applicable scenario and sole reliance on positive incentives mayproduce less careful attempts at feigning. However, this result was found only onthe LI scale and does not appear to be generalizable to the SIMS as a whole.Likewise, relevance appears to have a circumscribed effect when participants areprovided with an applicable role. When participants are challenged by adversarialcircumstances, the applicability of scenarios may affect their performance (i.e., rele-vance x applicability). Such interactions suggest that even small changes in incen-tives, context, and relevance may have appreciable effects on subsequent attemptsat feigning.

For measures (e.g., the MMPI-2) that rely almost exclusively on simulationdesigns, forensic experts must carefully evaluate threats to external validity. Thefollowing parameters must be considered:

• Are incentives offered differentially for "successful" and "unsuccessful"feigning?

• Are negative incentives provided?« Are scenarios (contexts) offered that are (a) relevant to the participants

and (b) germane to the psycholegal issue?® Are the comparison groups relevant to forensic evaluations with respect

to (a) representative range of mental disorders and (b) criminalbackgrounds of persons evaluated?

The implications of these parameters to forensic practice are far-reaching. Atpresent, psychologists are observed testifying about malingering on the MMPI-2based solely on standard texts (e.g., Butcher & Williams, 1992; Graham, 1993).These distillations amalgamate only a minority of current research studies and donot take into account the effects of incentive or forensic context on malingeredpresentations. Instead, forensic psychologists will need to be conversant with specific

4We suspect that this conclusion is not entirely accurate and that some malingerers in the criminaljustice system may receive harsher sentences.5We did not empirically test this assumption. Because participants were not anonymous (i.e., the designrequired notification of positive and negative incentives), we were concerned with their forthrightnessin admitting to criminal charges.

Malingering with Simulation Designs 2S3

malingering studies and whether these studies address pertinent psycholegal issues,are methodologically sophisticated, and compare their results to relevant forensicpopulations.

REFERENCES

American Psychological Association (1985). Standards for educational and psychological testing.Washington, DC: Author.

Bagby, R. M., Gillis, J. R., & Dickens, S. E. (1990). Detection of dissimulation on a new generation ofobjective personality measures. Behavioral Sciences and the Law, 8, 93-102.

Bagby, R. M., Rogers, R., Buis, T., & Kalemba, V. (1994). Malingered and defensive response styleson the MMPI-2: An examination of validity scales. Assessment, 1, 31-38.

Beaber, R. J., Marston, A., Michelli, J., & Mills, M. J. (1985). A brief test for measuring malingeringin schizophrenic individuals. American Journal of Psychiatry, 142, 1478-1481.

Berry, D. T. R., Adams, J. J., Clark, C. D., Thacker, S. R., Burger, T. L., Wetter, M. W., Baer, R. A.,& Borden, J. W. (1996). Detection of a cry for help on the MMPI- 2: An analog investigation.Journal of Personality Assessment, 67, 26-36.

Berry, D. T. R., Baer, R. A., & Harris, M. J. (1991). Detection of malingering on the MMPI: Ameta-analytic review. Clinical Psychology Review, 11, 585-598.

Berry, D. T. R., Wetter, M. W., & Baer, R. A. (1995). Assessment of malingering. In J. N, Butcher(Ed.), Clinical personality assessment: Practical approaches (pp. 236-248). Oxford: Oxford UniversityPress.

Butcher, J. N., & Williams, C. L. (1992). Essentials of MMPI-2 and MMPI-A interpretation. Minneapolis,MN: University of Minnesota Press.

Castaneda, R., Lifshutz, H., Galanter, M., & Franco, H. (1994). Empirical assessment of theself-medication hypothesis among dually diagnosed inpatients. Comprehensive Psychiatry, 35,180-184.

Derogatis, L. R. (1977). The SCL-90-R version manual: Scoring administration and procedures for theSCL-90. Baltimore: John Hopkins University, School of Medicine.

Derogatis, L. R., Lipman, R. S., Uhlenhuth, E. H., & Covi, L. (1974). The Hopkins Symptom Checklist(HSCL): A self-report symptom inventory. Behavioral Science, 19, 1-15.

Franzen, M. D., Iverson, G. L., McCracken, L. M. (1990). The detection of malingering inneuropsychological assessment. Neuropsychology Review, 1, 247-279.

Gothard, S., Viglione, D. J., Meloy, J. R., & Sherman, M. (1995). Detection of malingering incompetency to stand trial evaluations. Law and Human Behavior, 19, 493-505.

Graham, J. R. (1993). MMPI-2: Assessing personality and psychopathology (2nd ed.). Oxford: OxfordUniversity Press.

Graham, J. R., Watts, D., & Timbrook, R. E. (1991). Detecting fake-good and fake-bad MMPI-2 profiles.Journal of Personality Assessment, 57, 264-277.

Green, D. E., Walkey, F. H., Taylor, A. J., & McCormick, I. A. (1989). Replication of the factor structureof the Hopkins Symptom Checklist with New Zealand and United States respondents. New ZealandJournal of Psychology, 18, 60-64.

Greene, R. L. (1997). Assessment of malingering and defensiveness by multiscale inventories. In R.Rogers (Ed.), Clinical assessment of malingering and deception (2nd ed.) (pp. 169-207). New York:Guilford.

Iverson, G. I., Franzen, M. D., & Hammond, J. A. (1995). Examination of inmates' ability to malingeron the MMPI-2. Psychological Assessment, 7, 118-121.

Kazdin, A. E. (1992). Research design in clinical psychology (2nd ed.). Boston: Allyn and Bacon.Kraemer, H. C., & Thiemann, S. (19S7). How many subjects? Statistical power analysis. Newbury Park,

CA: Sage.Lees-Haley, P. R. (1991). MMPI-2 F and F-K scores of personal injury malingerers in vocational

neuropsychological and emotional distress claims. American Journal of Forensic Psychology, 9, 5-14.Lees-Haley, P. R., English, L. T., & Glenn, W. J. (1991), A fake bad scale on the MMPI-2 for

personal-injury claimants. Psychological Reports, 68, 203-210.Matt, G. E, Dean, A., Wang, B., & Wood, P. (1992). Identifying clinical syndromes in a community

sample of elderly persons. Psychological Assessment, 4, 174-184.Morey, L. C. (1991). Personality Assessment Inventory professional manual. Odessa, FL: Psychological

Assessment Resources.

2S4 Rogers and Cruise

Mossman, D., & Hart. K. J. (1996). Presenting evidence of malingering to courts: Insights from decisiontheory. Behavioral Sciences and the Law, 14, 271-291.

Patrick, C. J., & lacono, W. G. (1986). The validity of lie detection with criminal psychopaths.Psychophysiology, 23, 452-453.

Rogers, R. (1986). Conducting insanity evaluations. New York: Van Nostrand Reinhold.Rogers, R. (1997). Researching dissimulation. In R. Rogers (Ed.), Clinical assessment of malingering and

deception (2nd ed.) (pp. 398-426). New York: Guilford.Rogers, R., Bagby, R. M., & Dickens, S. E. (1992). Structured Interview of Reported Symptoms (SIRS)

and professional manual. Odessa, FL: Psychological Assessment Resources.Rogers, R, & Cavanaugh, Jr., J. L. (1983). "Nothing but the truth" ... A reexamination of malingering.

Journal of Law and Psychiatry, 11, 443-460.Rogers, R., Harrell, E. H., & Liff. C. D. (1993). Feigning neuropsychological impairment: A critical

review of methodological and clinical considerations. Clinical Psychology Review, 13, 255-274.Rogers, R., Hinds, J, D., & Sewell, K. W. (1996). Feigning psychopathology among adolescent offenders:

Validation of the SIRS, MMPI-A, and SIMS. Journal of Personality Assessment, 67, 244-257.Rogers, R., & Mitchell, C. N. (1991). Mental health experts and the criminal courts: A handbook for

lawyers and clinicians. Toronto: Carswell.Rogers, R., & Salekin, R. T. (1996). Beguiled by Bayes: A reanalysis of Mossman and Hart's estimates of

malingering. Manuscript submitted for publication.Rogers, R., Salekin, R. T., Sewell, K. W., & Goldstein, A. (in press). Prototypical analysis of malingering:

A comparison of forensic and nonforensic cases. Law and Human Behavior.Rogers, R., Sewell, K. W., & Goldstein, A. (1994). Explanatory models of malingering: A prototypical

analysis. Law and Human Behavior, 18, 543-552.Rogers, R., Sewell, K. W., & Salekin, R. (1994). A meta-analysis of malingering on the MMPI-2.

Assessment, 1, 227-237.Rogers, R., Ustad, K. L., & Salekin, R. T. (1998). Convergent validity of the Personality Assessment

Inventory: A study of emergency referrals in a correctional setting. Assessment, 5, 3-12.Schretlen, D. J. (1988). The use of psychological tests to identify malingered symptoms of mental

disorder. Clinical Psychology Review, 8, 457-476.Schretlen, D., & Arkowitz, H. (1990). A psychological test battery to detect prison inmates who fake

insanity or mental retardation. Behavioral Sciences and the Law, 8, 75-84.Smith, G. P. (1992). Detection of malingering: A validation study of the SLAM Test. Unpublished doctoral

dissertation, University of Missouri-St. Louis, St. Louis, MO.Smith, G, P. (1997). Assessment of malingering with self-report instruments. In R. Rogers (Ed.), Clinical

assessment of malingering and deception (2nd ed.) (pp. 351-372). New York: Guilford.Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics (3rd ed.). New York: Harper Coffins.Timbrook, R. E., Graham, J. R., Keiller, S. W., & Watts, D. (1993). Comparison of the Wiener-Harmon

subtle-obvious scales and the standard validity scales in detecting valid and invalid MMPI-2 profiles.Psychological Assessment, 5, 53-61.

Malingering with Simulation Designs 285