Identifying Overreporting in a Correctional Setting

Abstract

The current study examined the utility of the Minnesota Multiphasic Personality Inventory–2 Restructured Form (MMPI-2-RF) validity scales in a sample of incarcerated offenders. Utilizing an analogue simulation design, we compared a group of 36 inmates instructed to feign symptoms of mental illness with 56 inmates who had been referred for psychiatric treatment and completed the test under standard instructions. We also had a group of 63 inmates with no history of mental illness who completed the test under standard instructions. Our results indicated large effect sizes between the feigning and comparison groups for all the validity scales. Infrequent Responses (F-r) exhibited utility in discriminating between feigning and control inmates, whereas Infrequent Psychopathology Responses (Fp-r) and Infrequent Somatic Responses (Fs) exhibited the largest effect size in discriminating between the feigning and psychiatric inmate groups. Classification accuracy generally revealed greater specificity than sensitivity for the MMPI-2-RF validity scales.

Keywords

MMPI-2-RF malingering correctional psychology

Symptom exaggeration comes in many forms and occurs for many reasons in clinical assessment. When it represents a deliberate and intentional attempt at distorting symptoms that is motivated by an external secondary gain, it is classified as malingering. Malingering is a V code (other conditions and problems that may be a focus of clinical attention) in the Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR; American Psychiatric Association [APA], 2000) and was maintained in the DSM-5 (5th ed.; APA, 2013). In correctional settings, malingering can occur for a variety of reasons. Resnick and Knoll (2008) reported that inmates might malinger psychosis or other symptoms of a serious mental illness to serve “easy time” in an inpatient correctional setting. Gallagher, Ben-Porath, and Briggs (1997) suggested that inmates might want to appear more psychologically disturbed to acquire a less restrictive cell placement or to be exempted from work assignments. Some inmates (particularly in county detention centers and jails) may maintain the appearance of mental illness while awaiting trial to avoid charges or criminal responsibility for their behavior. Walters (2006) noted that inmates may malinger serious mental health problems to avoid predation from other inmates, to gain attention, or to simply amuse themselves through manipulation of the system. Although the base rate of malingering and symptom exaggeration has been examined in numerous settings (e.g., forensic, medical settings; Mittenberg, Patton, Canyock, & Condit, 2002), it is not entirely clear how often malingering occurs in correctional settings. Nevertheless, malingering remains an important consideration in correctional mental health assessment.

The Minnesota Multiphasic Personality Inventory–2 (MMPI-2; Butcher et al., 2001) is the most widely studied psychological measure for feigned mental disorders and the infrequency scales (i.e., F, Fp) are particularly effective in identifying overreported psychopathology (Rogers, Sewell, Martin, & Vitacco, 2003).

The MMPI-2 Restructured Form (MMPI-2-RF; Ben-Porath & Tellegen, 2008/2011) represents the newest revision in the MMPI family of instruments. The MMPI-2-RF is a restructured version that comprises 338 of the 567 items of the MMPI-2. The test is organized in hierarchical format, with 3 higher-order scales, 9 restructured clinical (RC) scales, 23 specific problem scales, 2 interest scales, and 5 revised Personality Psychopathology Five scales (see Ben-Porath, 2012, for a general description of the instrument and interpretation). It also contains nine validity scales that detect various forms of response bias (i.e., overreporting and underreporting of symptoms) as well as noncontent-based invalid responding, such as random responding.

The current study is focused on the MMPI-2-RF validity scales associated with symptom overreporting. The Infrequent Responses (F-r) scale was designed to capture general overreporting by including items rarely endorsed in the normative sample (Ben-Porath, 2012). The Infrequent Psychopathology Responses (Fp-r) scale serves as an indicator of overreported symptoms of severe psychopathology and includes a slight revision to the MMPI-2 version of Fp (Arbisi & Ben-Porath, 1995). The Infrequent Somatic Responses (Fs) scale serves as an indicator of overreported somatic complaints and includes items reflecting somatic content that are rarely endorsed among medical and chronic pain patients (Ben-Porath, 2012). The Symptom Validity (FBS-r) scale assesses noncredible somatic and neurocognitive complaints in civil forensic and medicolegal settings and works similarly to its MMPI-2 variant (Lees-Haley, English, & Glenn, 1991). Finally, the Response Bias Scale (RBS; Gervais, Ben-Porath, Wygant, & Green, 2007) assesses symptoms associated with overreported cognitive problems.

Much of the research on these scales was performed in civil forensic settings or for neuropsychological evaluations. Wygant et al. (2009) found that Infrequent Responses was effective in detecting general overreporting of psychological symptoms, whereas both Infrequent Somatic Responses and Symptom Validity were found to be effective in the detection of exaggerated somatic and neurocognitive complaints in a civil forensic setting. Youngjohn, Wershba, Stevenson, Sturgeon, and Thomas (2011) found that Symptom Validity was effective at distinguishing between traumatic brain injury (TBI) litigants who pass or fail formal effort tests. Wygant et al. (2011) found that Infrequent Responses and the RBS performed the best in predicting malingering in a sample of personal injury and disability claimants who were classified with specific malingering criteria.

To our knowledge, only three studies have examined the MMPI-2-RF validity scales in criminal settings, and all of these involve pretrial forensic evaluations. Sellbom, Toomey, Wygant, and Kucharski (2010) examined the utility of the MMPI-2-RF validity scales to detect feigned psychopathology in a criminal forensic setting. Using a criterion-groups design that incorporated the Structured Interview of Reported Symptoms (SIRS; Rogers, Bagby, & Dickens, 1992) as a means to distinguish defendants who exhibited evidence of malingering from those without such evidence on the SIRS, their results demonstrated that Infrequent Responses and Infrequent Psychopathological Responses scales differentiated nonmalingering and malingering groups most effectively. Wygant et al. (2010) examined the utility of the RBS to detect cognitive response bias in a criminal forensic (as well as a civil forensic) sample and found that this scale added incrementally over other standard MMPI-2 and MMPI-2-RF overreporting scales in detecting symptom exaggeration associated with feigned cognitive problems. Interestingly, the RBS was correlated significantly higher with the Infrequency scales and the Aberrant Experiences (RC8) scale in the criminal sample than in the civil sample, suggesting that higher RBS scores are associated with overreporting of psychopathology symptoms to a greater extent among criminal defendants versus civil litigants. Finally, Marion et al. (2013) examined whether criminal defendants undergoing a forensic evaluation who had significant psychopathic traits (as assessed by the Psychopathy Checklist–Revised [PCL-R]; Hare, 2003) were better at avoiding detection of feigned psychopathology than defendants low in psychopathy. Their results suggested that psychopathy did not affect the MMPI-2-RF validity scales in detecting overreporting.

Several published studies have examined the MMPI-2-RF validity scales’ ability to assess feigned symptoms of severe mental illness. Purdon, Purser, and Goddard (2011) examined MMPI-2-RF overreporting scales in patients referred to a first-episode psychosis (FEP) clinic. Their results indicate that moderately elevated scores on Infrequent Psychopathological Responses may imply psychopathological symptom overreporting; however, they suggested caution in that clinician ratings of positive symptoms of psychosis were also associated with Infrequent Psychopathological Responses. Sellbom and Bagby (2010) found that Infrequent Psychopathological Responses best differentiated undergraduate students instructed to overreport psychopathology and genuine psychiatric patients. Marion, Sellbom, and Bagby (2011) also compared undergraduate students instructed to feign symptoms with genuine psychiatric patients (diagnosed with Major Depressive Disorder, Schizophrenia, Posttraumatic Stress Disorder). Their simulators were either naïve (no training in mental disorders) or had advanced training in psychopathology or personal experiences with a disorder. Again, Infrequent Psychopathological Responses best differentiated simulators from patients regardless of training or experience with mental illness. Although these studies did not utilize a correctional sample, evaluation of the validity scales among patients with severe mental illnesses is relevant in correctional settings, where inmates may attempt to feign symptoms of severe mental illness.

Goal of the Present Study

To date, there have been no studies that have examined the MMPI-2-RF validity scales using a simulation design with correctional inmates. The current study compared inmates with no psychiatric history who were asked to feign symptoms of mental illness with inmates who completed the test under standard instructions. Moreover, we utilized a comparison sample of inmates with known psychiatric history who completed testing as part of their intake process.

Given that Infrequent Responses was designed to capture responses not typically endorsed by the general population, it is hypothesized that this scale will successfully discriminate between the inmate feigning group and those inmates with no psychiatric history who completed the test under standard instructions. Because Infrequent Psychopathological Responses was designed to capture responses not typically endorsed by psychiatric populations, it is also hypothesized that the Infrequent Psychopathological Responses scale will outperform Infrequent Responses and the other overreporting scales in discriminating between the feigning group and the group of psychiatric inmates who completed the test as part of their intake process. Given that Symptom Validity and the RBS were developed for use in civil forensic settings (e.g., disability evaluations), it is hypothesized that these scales will exhibit less utility at discriminating between the groups in this study. Finally, although Infrequent Somatic Responses was specifically developed to assess somatic overreporting, this scale might still exhibit utility in the current study given that the items still reflect infrequent symptoms that may be endorsed by individuals feigning various symptoms of mental illness. Furthermore, we evaluated two malingering detection strategies for correctional settings developed for the MMPI-2 by Steffan, Morgan, Lee, and Sellbom (2010), termed the sequential model and the correctional model.¹

Method

Participants

Data from this sample were first used by Gallagher (1998) and are composed of two groups, including male inmates recruited for participation in a simulation study and a set of archival data from inmates identified as receiving psychiatric services from the Ohio Department of Rehabilitation and Corrections. The first group included 196 male inmates recruited from the general population of a correctional reception center, who were identified as possessing at least a sixth-grade reading level as assessed by the Tests of Adult Basic Education (TABE; CTB/McGraw-Hill, 1987) and had no history of referral for psychiatric services. The mean age of this group was 29.3 (SD = 11.0) and the group was predominantly African American (48%), with 39.8% Caucasian and 3.1% Latino. Ten percent of the sample did not have information pertaining to their ethnic background. Seventy-seven percent of this sample had at least a high school diploma.

The second sample included archival data on 73 male inmates who received psychiatric services during their incarceration. Their mean age was 29.8 (SD = 9.7) and they had a mean education of 9.8 years (SD = 5.8). The sample was predominantly Caucasian (56.2%), with 41.1% being African American. Most inmates presented with mood and anxiety disorders (70.2%), although 12.3% had psychotic disorders, 10.5% had adjustment disorders, and 7% had substance use disorders.

Measure

MMPI-2-RF

Although participants in this study were administered the full MMPI-2, the present study only focuses on the overreporting validity scales from the MMPI-2-RF (Ben-Porath & Tellegen, 2008/2011), which include the Infrequent Responses (F-r), Infrequent Psychopathological Responses (Fp-r), Infrequent Somatic Responses (Fs), Symptom Validity (FBS-r), and RBS. The RC scales are also presented to illustrate symptom endorsements among the three groups. Information regarding these scales can be found in the MMPI-2-RF Manual for Administration, Scoring, and Interpretation (Ben-Porath & Tellegen, 2008/2011).

Procedure

Participants from the first sample were randomly assigned to either a standard instruction group or the feigning group. Inmates in the standard instruction group completed the MMPI-2 after being read the standard administration instructions from the test’s manual. Inmates in the feigning group received the following instructions:

We are interested in how well people can pretend to be psychologically or emotionally disturbed on the MMPI-2. Please complete this test as if you were trying to create the impression that you have serious psychological problems. Imagine that you want the Psychology Department to think you are so disturbed that you cannot work and that you should get your own cell.

Participants in the correctional psychiatric services group completed the MMPI-2 under standard administration instructions.

One hundred fourteen participants were excluded from further analysis as they exhibited nonresponding (Cannot Say; CNS > 18), variable responding (Variable Response Inconsistency; VRIN-r > 80), and fixed responding (True Response Inconsistency; TRIN-r > 80). This resulted in final sample sizes of 63 (Standard Instructions), 36 (Feigning), and 56 (Psychiatric Inmates). Participants who produced valid versus invalid profiles were compared for age, education, marital status, and race. They differed in terms of race and marital status. Among participants who produced valid profiles, 60.6% were Caucasian, 34.2% were African American, and 3.8% reported another ethnic origin. Among participants who produced invalid profiles, 21.9% were Caucasian, 62.3% were African American, and 10.6% reported another ethnic origin. This difference was statistically significant, χ²(7) = 44.75, p < .001. Among participants who produced valid profiles, 24.6% were married, 14.9% were divorced, 4.5% were separated, 54.5% were single, and 0.6% were widowed. Among participants who produced invalid profiles, 19.5% were married, 8.0% were divorced, 2.7% were separated, 63.7% were single, and 1.8% were widowed. This difference was statistically significant, χ²(6) = 13.19, p = .04.

Results

Validity Scale Group Differences

A one-way MANOVA was conducted to determine whether there were overall differences on the MMPI-2-RF overreporting validity scales across the three conditions. The overall model was statistically significant, Hotelling’s T = .866, F(10, 294) = 12.73, p < .001. We followed up with univariate analyses of variance for each individual overreporting validity scale to test differences among the standard, psychiatric, and faking groups. These results are presented in Table 1. All five validity scale scores in the feigning group were significantly higher compared with those of the standard and psychiatric inmate groups. The RBS, Infrequent Responses, and Infrequent Psychopathological Responses scales did best in differentiating between the feigning and standard groups. Infrequent Psychopathological Responses and Infrequent Somatic Responses did best in differentiating between the feigning and psychiatric inmate groups.

Table 1:

ANOVA Between Groups

	Group
	Standard (n = 63)		Psychiatric Inmates (n = 56)		Feigning (n = 36)		ANOVA		Effect Size
	M	SD	M	SD	M	SD	F(2, 152)	p	η²	d ₁	d ₂
F-r	66.6_a	22.3	76.9_a	34.7	126.1_b	50.8	34.8	<.001	.46	1.68	1.18
Fp-r	62.2_a	27.0	63.9_a	26.3	134.4_b	58.7	52.7	<.001	.69	1.75	1.68
Fs	57.8_a	18.9	63.3_a	20.7	109.4_b	46.3	42.3	<.001	.56	1.63	1.39
FBS-r	48.5_a	12.9	57.7_b	16.3	72.6_c	19.4	26.5	<.001	.35	1.55	0.85
RBS	53.0_a	14.4	62.7_b	19.7	88.8_c	27.0	37.9	<.001	.50	1.80	1.14

Note. Means with different subtext are significantly different (Tukey’s HSD). d₁ = Cohen’s d between Standard and Feigning groups; d₂ = Cohen’s d between Psych. Services and Feigning groups; F-r = Infrequent Responses; Fp-r = Infrequent Psychopathology Responses; Fs = Infrequent Somatic Responses; FBS-r = Symptom Validity; RBS = Response Bias Scale.

RC Scale Group Differences

To examine the types of symptoms presented by the three groups, we present the mean profiles of the RC scales for each group in Table 2. Again, a one-way MANOVA was conducted to determine whether there were overall differences on the MMPI-2-RF RC scales between the three groups. The model was significant, Hotelling’s T = .655, F(18, 286) = 5.21, p < .001. Follow-up ANOVA shows that the groups had significantly different scores on all the RC scales except RC3 Cynicism and RC9 Hypomanic Activation.

Table 2:

ANOVA Between Groups

	Group
	Standard (n = 63)		Psychiatric Inmates (n = 56)		Feigning (n = 36)		ANOVA		Effect Size
	M	SD	M	SD	M	SD	F(2, 152)	p	η²	d ₁	d ₂
RCd	55.2_a	10.1	59.7_a	13.4	68.1_b	13.9	12.3	<.001	.16	1.11	0.61
RC1	53.4_a	12.4	60.4_a	18.4	78.9_b	19.2	30.8	<.001	.41	1.68	0.91
RC2	50.4_a	9.5	57.4_b	15.2	70.1_c	16.8	25.5	<.001	.34	1.56	0.80
RC3	59.8_a	11.3	58.1_a	12.7	66.6_b	12.3	5.4	.005	.07	0.58	0.46
RC4	68.3_a	10.2	67.2_a	12.5	76.8_b	13.6	8.0	<.001	.11	0.74	0.53
RC6	64.6_a	15.9	64.8_a	20.0	87.1_b	20.5	22.2	<.001	.29	1.27	0.85
RC7	53.2_a	11.9	55.3_a	13.5	68.9_b	14.6	17.8	<.001	.23	1.21	0.98
RC8	56.7_a	12.1	59.2_a	16.5	79.5_b	21.6	24.9	<.001	.32	1.41	1.09
RC9	58.2_ab	11.5	54.2_ab	14.0	63.6_bc	12.7	5.9	.003	.08	0.45	0.70

Note. Means with different subtext are significantly different (Tukey’s HSD). d₁ = Cohen’s d between Standard and Feigning groups; d₂ = Cohen’s d between Psych. Services and Feigning groups; RCd = Demoralization; RC1 = Somatic Complaints; RC2 = Low Positive Emotions; RC3 = Cynicism; RC4 = Antisocial Behavior; RC6 = Ideas of Persecution; RC7 = Dysfunctional Negative Emotions; RC8 = Aberrant Experiences; RC9 = Hypomanic Activation.

Incremental Validity

Next, we conducted hierarchical logistic regression analyses to determine whether Infrequent Somatic Responses, RBS, and Symptom Validity added to Infrequent Responses or Infrequent Psychopathological Responses in differentiating between the feigning and standard instruction groups and the feigning and psychiatric inmate groups. We chose these scales because they exhibited large effect sizes in differentiating between the groups in the current study. Moreover, RC8 (Aberrant Experiences), which captures psychotic symptoms exhibited the largest effect between the feigning and psychiatric inmate groups. Infrequent Responses and Infrequent Psychopathological Responses were specifically designed to detect symptom exaggeration of this nature. We conducted six separate planned regression analyses. First, we examined whether Infrequent Psychopathological Responses, Infrequent Somatic Responses, RBS, and Symptom Validity would add incremental predictive utility to Infrequent Responses. In three regressions, Infrequent Responses was entered in the first step, and Infrequent Psychopathological Responses, Infrequent Somatic Responses, RBS, or Symptom Validity was entered in the second step. Next, three additional analyses were conducted in which Infrequent Psychopathological Responses was entered in the first step and Infrequent Responses, Infrequent Somatic Responses, RBS, and Symptom Validity were entered into their respective steps. Table 3 shows the results of the analyses, which indicate that only the RBS added incrementally to Infrequent Responses and Infrequent Psychopathological Responses in differentiating the feigning and standard instruction groups. Infrequent Psychopathological Responses and Infrequent Responses each added incrementally to each other in differentiating between the feigning and psychiatric inmate groups. Infrequent Somatic Responses also added incrementally to Infrequent Responses in this prediction as well.

Table 3:

Hierarchical Logistic Regression Analysis Results for Predicting Feigning Status

	Parameters
Step	Model Fit χ² (df)	w’	χ² Change (df)	w’ Change	R ²	ΔR²
Feigning group versus standard group
F-r entered first
1. F-r	43.11***	.66			.48
2a. Fp-r	46.34***	.68	3.23	.18	.51	.03
2b. Fs	43.57***	.66	0.46	.07	.49	.01
2c. RBS	47.77***	.69	4.66*	.22	.52	.04
2d. FBS-r	44.29***	.67	1.18	.01	.49	.01
Fp-r entered first
1. Fp-r	45.45***	.68			.50
2a. F-r	46.34***	.68	0.89	.09	.51	.01
2b. Fs	45.62***	.68	0.17	.04	.51	.01
2c. RBS	49.83***	.71	4.38*	.21	.54	.04
2d. FBS-r	46.84***	.69	1.40	.12	.52	.02
Feigning group versus psychiatric inmate group
F-r entered first
1. F-r	24.43***	.52			.32
2a. Fp-r	49.91***	.74	25.48***	.53	.57	.25
2b. Fs	31.97***	.59	7.54**	.29	.40	.08
2c. RBS	24.89***	.52	0.46	.07	.32	.00
2d. FBS-r	24.66***	.52	0.23	.05	.32	.00
Fp-r entered first
1. Fp-r	41.94***	.68			.50
2a. F-r	49.91***	.74	7.98**	.29	.57	.07
2b. Fs	41.94***	.68	0.00	.00	.50	.00
2c. RBS	44.98***	.70	3.05	.18	.52	.02
2d. FBS-r	44.40***	.69	2.46	.16	.52	.02

Note. Nagelkerke R² estimation was used for logistic regression. w’ = effect size for χ² statistic; w’ change = effect size for χ² change statistic; F-r = Infrequent Responses; Fp-r = Infrequent Psychopathology Responses; Fs = Infrequent Somatic Responses; RBS = Response Bias Scale; FBS-r = Symptom Validity.

p < .05. **p < .01. ***p < .001.

Classification Accuracy

Classification accuracies of the Infrequent Responses, Infrequent Psychopathological Responses, Infrequent Somatic Responses, and the RBS in differentiating feigning and psychiatric inmate groups were examined. We also calculated classification accuracy for the sequential and correctional models of detection (see Steffan et al., 2010, for more details), in addition to when any of the overreporting scales were elevated. We provided estimates of predictive powers across the different base rates: .15, .30, and .50. Positive and negative predictive powers (PPP and NPP) are indices of classification efficiency in that they provide a probability whether the individual is engaging in feigning given a certain cutoff value. We selected a range of possible optimal cutoff scores with an emphasis on reducing false positive predictions (approximately 10% false positive rate). Table 4 shows these results. In general, the MMPI-2-RF overreporting validity scales display better specificity than sensitivity, particularly at higher cutoffs. This suggests that impressions formed from these scales about symptom feigning (vs. bona fide psychopathology) will be stronger at higher cutoffs. Both Infrequent Responses and Infrequent Psychopathological Responses exhibit good specificity when they reach 100, with smaller (11% false positive rates) occurring at 120 for Infrequent Responses and 110 for Infrequent Psychopathological Responses. Not surprising, with better specificity than sensitivity, these scales exhibit good NPP, but PPP does not become strong until higher base rates are estimated. Infrequent Somatic Responses displayed better sensitivity and specificity than Infrequent Responses and Infrequent Psychopathological Responses, particularly at a cutoff of 100, with three fourths of the feigning group scoring in that range and less than 10% of psychiatric inmates scoring above this cutoff. The RBS did not do as well as the other scales in the classification of overreporting. The sequential and correctional detection models also displayed better specificity than sensitivity. The sequential model had sensitivity of .64 and specificity of .89 in differentiating between the two groups. The correctional model had sensitivity of .67 and specificity of .79. Any scale elevation performed similarly to the correctional model.

Table 4:

Classification Accuracy Statistics in Differentiating Between Feigning and Psychiatric Inmate Groups

				BR = .15		BR = .30		BR = .50
Cutoff Score	SN	SP	OCC^a	PPP	NPP	PPP	NPP	PPP	NPP
F-r
T = 120	.61	.89	.78	.49	.93	.70	.84	.85	.70
T ≥ 115	.64	.84	.76	.41	.93	.63	.84	.80	.70
T ≥ 105	.67	.82	.76	.40	.93	.61	.85	.79	.71
T ≥ 100	.67	.77	.73	.34	.93	.56	.84	.74	.70
Fp-r
T ≥ 110	.64	.89	.79	.51	.93	.72	.85	.86	.71
T ≥ 100	.67	.86	.78	.46	.94	.67	.86	.83	.72
T ≥ 90	.69	.84	.78	.43	.94	.65	.86	.81	.73
T ≥ 80	.72	.80	.77	.39	.94	.61	.87	.78	.74
Fs
T ≥ 115	.69	.98	.87	.86	.95	.94	.88	.97	.76
T ≥ 100	.75	.91	.85	.60	.96	.78	.89	.89	.78
T ≥ 90	.81	.88	.85	.54	.96	.74	.92	.87	.82
T ≥ 80	.81	.77	.78	.38	.96	.60	.90	.78	.80
RBS
T ≥ 90	.58	.86	.75	.42	.92	.64	.83	.81	.67
T ≥ 80	.69	.80	.76	.38	.94	.60	.86	.78	.72
T ≥ 70	.72	.73	.73	.32	.94	.53	.86	.73	.72
T ≥ 60	.75	.61	.66	.25	.93	.45	.85	.66	.71
Model
Sequential	.64	.89	.79	.51	.93	.72	.85	.86	.71
Correctional	.67	.79	.74	.35	.93	.57	.85	.76	.70
Any elevated	.67	.79	.74	.35	.93	.57	.85	.76	.70

Note. Optimal cut score in bold. BR = Base Rate, SN = Sensitivity; SP = Specificity; OCC = Overall Correct Classification; PPP = Positive Predictive Power; NPP = Negative Predictive Power; F-r = Infrequent Responses; T = T-score; Fp-r = Infrequent Psychopathology Response; Fs = Infrequent Somatic Responses; RBS = Response Bias Scale.

OCC values are based on base rates in current sample (.39).

Discussion

The goal of the present study was to examine the validity and clinical utility of the MMPI-2-RF overreporting validity scales in a correctional setting. As expected, we found that the validity scales significantly discriminated between the three groups. Infrequent Responses had the largest effect size in comparing the standard and feigning groups, which is consistent with this scale’s composition of items that are rarely endorsed in the test’s nonclinical normative sample. Given that the standard group in this study was composed of inmates with a history of mental illness, they would likely approximate the normative group on most scales. Not surprisingly, the standard group’s highest scores were on RC4 (Antisocial Behavior) and RC6 (Ideas of Persecution), which is consistent with correctional data sets. Infrequent Responses had a lower effect size in discriminating between the feigning and psychiatric inmate groups, which is not surprising, given that bona fide symptoms of mental illness will inflate scores on this scale (Ben-Porath & Tellegen, 2008). Infrequent Psychopathological Responses had the largest effect size in differentiating between the feigning and psychiatric inmate samples, which again is not surprising given that this scale was designed to identify symptom exaggeration in samples with higher rates of psychopathology (Arbisi & Ben-Porath, 1995). Indeed, Infrequent Psychopathological Responses comprises items rarely endorsed among bona fide psychiatric patients. In terms of classification accuracy, Infrequent Somatic Responses exhibited the best combination of sensitivity and specificity among the overreporting scales. Indeed, this scale has detected between 11% and 14% more of the feigning inmates than Infrequent Psychopathological Responses and Infrequent Responses at their optimal cutoffs, respectively. This suggests that one strategy used by feigning inmates was to report infrequent somatic and physical health complaints in addition to symptoms of severe psychopathology. The RBS, which was developed specifically to identify symptoms associated with neurocognitive malingering (Gervais et al., 2007), exhibited large effect sizes in differentiating the feigning group from the standard and psychiatric inmate groups. However, the RBS fared worse in the classification analyses. At the most optimal cutoff to avoid false positives, the scale classified just over half of the feigning group.

The current findings show that the optimal cutoff score for Infrequent Responses is 120T to avoid false positives, as individuals with bona fide symptoms of mental illness often exhibit elevated scores on this scale. Thus, correctional psychologists should focus more on Infrequent Psychopathological Responses in detecting malingering as this scale was more specific to feigning and less confounded with genuine psychopathology. The optimal cutoff score for Infrequent Psychopathological Responses is 110T. Infrequent Somatic Responses may so be useful in distinguishing feigning from bona fide mental illnesses in a correctional setting. This scale exhibited incremental predictive utility and had the best overall classification accuracy, with an optimal cutoff of 100T. Furthermore, the sequential and correctional detection models do not appear to work any better than any single validity scale on the MMPI-2-RF. Although these approaches are clinically useful for the MMPI-2 (Steffan et al., 2010), they do not appear to work as well for the MMPI-2-RF.

The sensitivity and specificity of the MMPI-2-RF Infrequent Responses and Infrequent Psychopathological Responses scales are roughly equivalent (albeit slightly lower) to the MMPI-2 F and Fp scales in classifying malingering. Using a T-score of 100 or greater for the F and Fp scales, Gallagher (1998) found that F had sensitivity of .71 and specificity of .90 and Fp had sensitivity of .74 and specificity of .93 in distinguishing between the feigning and psychiatric inmate groups. The current study found that both Infrequent Responses and Infrequent Psychopathological Responses had slightly lower sensitivity and specificity than their respective MMPI-2 scales in distinguishing between the same groups. Interestingly, Infrequent Somatic Responses (which does not have a counterpart on the MMPI-2) exhibited very similar sensitivity and specificity with the MMPI-2 F and Fp scales.

Although the current study is the first to examine the MMPI-2-RF validity scales in a simulation design in correctional setting, it is not without limitations. First, the sample sizes for the three groups are small due to a large percentage of noncontent-based responding. The rate of invalidity (23%) among the psychiatric group was not uncommon for such groups (e.g., Purdon et al., 2011). Furthermore, the high percentage (49%) of the fake bad group suggests that these participants may have attempted to malinger by responding in a random manner. However, future studies should attempt to replicate these findings with larger samples. Another limitation is the use of an analogue simulation design. Despite this approach having good internal validity, the feigning inmates might not generalize well to real world situations in which the incentives for malingering are higher. However, comparing our sample of feigning inmates with the sample of criminal defendants classified as malingering on the SIRS in a previous study (Sellbom et al., 2010) revealed small effects (Cohen’s ds ranged from .23 to .38) for Infrequent Responses, Infrequent Psychopathological Responses, and Infrequent Somatic Responses between the two samples.² This suggests that our simulators performed similarly to a group of criminal defendants classified as malingering by a well-established external criterion. This allays concerns to some degree that simulators do not approximate malingers classified in criterion-groups designs, which are credited as having greater external validity. Rogers (2008) suggested that the use of clinical comparison groups can address to some extent the limitations of simulation designs. To address this limitation, the current study utilized a psychiatric inmate sample as a clinical comparison group to evaluate the extent to which the MMPI-2-RF validity scales can differentiate genuine from feigned psychopathology. Our psychiatric inmate sample appears similar to the sample of male psychiatric inpatients reported in the MMPI-2-RF Technical Manual (Tellegen & Ben-Porath, 2008). Indeed, with the exception of Cynicism (RC3: d = 75), Antisocial Behavior (RC4: d = .46), and Ideas of Persecution (RC6: d = .41), the average effect size between our psychiatric sample and the MMPI-2-RF Psychiatric Inpatient comparison group on the remaining RC scales was d = .02.³ It is not surprising that a sample of prison inmates would score higher on a measure of cynicism, antisocial behavior, and suspiciousness than a psychiatric sample; however, this basic comparison suggests that our clinical comparison group approximates other psychiatric samples with respects to mean RC profile. Although RC2 was the only scale to significantly differentiate the psychiatric and standard inmate groups, this is likely due to increased variability on the RC scales for the psychiatric group. Indeed, the average standard deviation of the RC scales for the standard group was 11.7 versus 15.1 in the psychiatric inmate group. Although RC2 was the only statistically significant difference between the groups, there were still moderate differences between the two groups on RC1 (d = .48) and RCd (d = .38).

Furthermore, another limitation of the current study is the possibility of a cohort effect. The testing of the participants for this study with the MMPI-2 was performed in the 1990s, before the proliferation of medical information on the Internet. Many individuals attempting to malinger in medicolegal settings are now obtaining information through Internet searches about various symptoms and diseases, as well as resources and written material that describe psychological tests (including descriptions of validity scales; Rogers, 2008). Data from the current sample might differ from any attempt made to malinger illness in the present day. More research needs to be conducted to determine how information on the Internet has changed how individuals malinger.

Finally, the use of the MMPI-2 (from which the MMPI-2-RF scales were scored) may introduce another potential confound in the current study. Although the two versions of the test assess similar constructs, the significantly longer length of the full MMPI-2 may introduce factors such as examinee fatigue that might not be present with the briefer MMPI-2-RF administration. Recent research by Tarescavage, Alosco, Ben-Porath, Wood, and Luna-Jones (2014) compared the internal structure and invalidity rates on the MMPI-2-RF between criminal defendants who completed the MMPI-2 (from which the MMPI-2-RF scales were rescored) and the MMPI-2-RF. They found no statistically significant differences in the invalidity rates between defendants who completed the MMPI-2-RF and full MMPI-2. Nevertheless, future studies need to examine whether the findings from this study, where the MMPI-2-RF scales were scored from full MMPI-2 administrations, would generalize to the detection of feigning on the MMPI-2-RF.

Conclusion

In summary, our results suggest that the MMPI-2-RF has the potential to identify symptom exaggeration among correctional inmates. Psychologists are increasingly called on to provide both assessment and treatment of incarcerated offenders (see Wormith et al., 2007 for a review). Effective and efficient psychological assessment can assist in these endeavors. With its shorter length than the MMPI-2 and improved psychometric qualities, the MMPI-2-RF is well-positioned to assist correctional psychologists evaluate the needs of incarcerated offenders. The current study suggests that as a first step, the MMPI-2-RF can reliably detect when incarcerated offenders overreport psychopathology.

Footnotes

Notes

Tina D. Wall, MS, is a doctoral student in the Applied Developmental Program at the University of New Orleans. She received her BS in psychology from the University of Alabama and MS in clinical psychology from Eastern Kentucky University. Her research interests include investigating the personality traits that covary with psychopathy, as well as approaches to assess psychopathy, especially with noncriminal or “successful” psychopaths. She is also more broadly interested in forensic research and developmental psychopathology.

Dustin B. Wygant, PhD, is Associate Professor of Psychology and the Director of Clinical Training at Eastern Kentucky University. He received his BA in psychology at Miami University and his MA and PhD in clinical psychology at Kent State University. His primary research interests include the conceptualization of the psychopathic personality and the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5) Section III model of personality disorders, detection of malingering, and the use of personality assessment (primarily with the MMPI-2-RF) in forensic and medical settings. He is a licensed psychologist and maintains an active private practice providing forensic psychological evaluations and expert testimony.

Robert W. Gallagher, PhD, is director of psychological consultation and part owner of the Guardian Defense Group. Prior to joining his current firm, he completed nearly 20 years of government service in direct support of military and civilian intelligence. He was a leading figure in the development of operational psychology and helped solidify disparate entities across government into a thriving scientific community of practice. His primary interest area remains the application of psychological assessment in support of national security objectives.

References

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: American Psychiatric Publishing.

Arbisi

P. A.

Ben-Porath

Y. S.

(1995). An MMPI-2 infrequent response scale for use with psychopathological populations: The infrequency–psychopathology scale, F(p). Psychological Assessment, 7, 424-431.

Ben-Porath

Y. S.

(2012). Interpreting the MMPI-2-RF. Minneapolis: University of Minnesota Press.

Ben-Porath

Y. S.

Tellegen

(2008/2011). MMPI-2-RF (Minnesota Multiphasic Personality Inventory–2 Restructured Form): Manual for administration, scoring, and interpretation. Minneapolis: University of Minnesota Press.

Butcher

J. N.

Graham

J. R.

Ben-Porath

Y. S.

Tellegen

Dahlstrom

W. G.

Kaemmer

(2001). Minnesota Multiphasic Personality Inventory–2: Manual for administration and scoring (2nd ed.). Minneapolis: University of Minnesota Press.

CTB/McGraw-Hill. (1987). Tests of Adult Basic Education Forms 5 and 6: Technical report. Monterey, CA: Author.

Gallagher

R. W.

(1998). Detection of malingering at the time of intake in a correctional institution with the MMPI-2 validity scales (Unpublished doctoral dissertation). Kent State University, Kent, OH.

Gallagher

R. W.

Ben-Porath

Y. S.

Briggs

(1997). Inmate views about the purpose and use of the MMPI-2 at the time of correctional intake. Criminal Justice and Behavior, 24, 360-369.

10.

Gervais

R. O.

Ben-Porath

Y. S.

Wygant

D. B.

Green

(2007). Development and validation of a Response Bias Scale (RBS) for the MMPI-2. Assessment, 14, 196-208.

11.

Hare

R. D.

(2003). Hare Psychopathy Checklist–Revised manual (2nd ed.). Toronto, Ontario, Canada: Multi-Health Systems.

12.

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.

13.

Marion

B. E.

Sellbom

Bagby

R. M.

(2011). The detection of feigned psychiatric disorders using the MMPI-2-RF overreporting validity scales: An analog investigation. Psychological Injury and Law, 4, 1-12.

14.

Marion

B. E.

Sellbom

Salekin

R. T.

Toomey

J. A.

Kucharski

Duncan

(2013). An examination of the association between psychopathy and dissimulation using the MMPI-2-RF validity scales. Law and Human Behavior, 37, 219-230.

15.

Mittenberg

Patton

Canyock

E. M.

Condit

D. C.

(2002). Base rates of malingering and symptom exaggeration. Journal of Clinical and Experimental Neuropsychology, 24, 1094-1102.

16.

Purdon

S. E.

Purser

S. M.

Goddard

K. M.

(2011). MMPI-2 Restructured Form overreporting scales in first-episode psychosis. The Clinical Neuropsychologist, 25, 829-842.

17.

Resnick

P. J.

Knoll

J. L.

(2008). Malingered psychosis. In Rogers

(Ed.), Clinical assessment of malingering and deception (pp. 51-68). New York, NY: Guilford Press.

18.

Rogers

(2008). In Rogers

(Ed.), Clinical assessment of malingering and deception (3rd ed., pp. 411-434). New York, NY: Guilford Press.

19.

Rogers

Bagby

R. M.

Dickens

S. E.

(1992). Structured Interview of Reported Symptoms: Professional manual. Odessa, FL: Psychological Assessment Resources.

20.

Rogers

Sewell

K. W.

Martin

M. A.

Vitacco

M. J.

(2003). Detection of feigned mental disorders: A meta-analysis of the MMPI-2 and malingering. Assessment, 10, 160-177.

21.

Sellbom

Bagby

R. M.

(2010). Detection of overreported psychopathology with the MMPI-2-RF validity scales. Psychological Assessment, 22, 757-767.

22.

Sellbom

Toomey

J. A.

Wygant

D. B.

Kucharski

L. T.

(2010). Utility of the MMPI-2-RF (Restructured Form) validity scales in detecting malingering in a criminal forensic setting: A known-groups design. Psychological Assessment, 22, 22-31.

23.

Steffan

J. S.

Morgan

R. D.

Lee

Sellbom

(2010). A comparative analysis of MMPI-2 malingering detection models among inmates. Assessment, 17, 187-196.

24.

Tarescavage

A. M.

Alosco

M. L.

Ben-Porath

Y. S.

Wood

Luna-Jones

(2014). Minnesota Multiphasic Personality Inventory–2–Restructured Form (MMPI-2-RF) scores generated from the MMPI-2 and MMPI-2-RF test booklets: Internal structure comparability in a sample of criminal defendants. Assessment. Advance online publication. doi:1073191114537347. Retrieved from http://asm.sagepub.com/content/early/2014/06/13/1073191114537347

25.

Tellegen

Ben-Porath

Y. S.

(2008). Minnesota Multiphasic Personality Inventory–2 Restructured Form: Technical manual. Minneapolis: University of Minnesota Press.

26.

Walters

G. D.

(2006). Coping with malingering and exaggeration of psychiatric symptomology in offender populations. American Journal of Forensic Psychology, 24, 21-40.

27.

Wormith

J. S.

Althouse

Simpson

Reitzel

L. R.

Fagan

T. J.

Morgan

R. D.

(2007). The rehabilitation and reintegration of offenders: The current landscape and some future directions for correctional psychology. Criminal Justice and Behavior, 34, 879-892.

28.

Wygant

D. B.

Anderson

J. L.

Sellbom

Rapier

J. L.

Allgeier

L. M.

Granacher

R. P.

(2011). Association of the MMPI-2 Restructured Form (MMPI-2-RF) validity scales with structured malingering criteria. Psychological Injury and Law, 4, 13-23.

29.

Wygant

D. B.

Ben-Porath

Y. S.

Arbisi

P. A.

Berry

D. T. R.

Freeman

D. B.

Heilbronner

R. L.

(2009). Examination of the MMPI-2 Restructured Form (MMPI-2-RF) validity scales in civil forensic settings: Findings from simulation and known group samples. Archives of Clinical Neuropsychology, 24, 671-680.

30.

Wygant

D. B.

Sellbom

Gervais

R. O.

Ben-Porath

Y. S.

Stafford

K. P.

Freeman

D. B.

Heilbronner

R. L.

(2010). Further validation of the MMPI-2 and MMPI-2-RF Response Bias Scale: Findings from disability and criminal forensic settings. Psychological Assessment, 22, 745-756.

31.

Youngjohn

J. R.

Wershba

Stevenson

Sturgeon

Thomas

M. L.

(2011). Independent validation of the MMPI-2-RF somatic/cognitive and validity scales in TBI litigants tested for effort. The Clinical Neuropsychologist, 25, 463-476.