Abstract

The main goal of a severe impairment profile (SIP) on a performance validity test (PVT) is to help reduce the false-positive rate when identifying non-credible effort in people who are truly impaired. A secondary goal is to help with clinical judgment about impairment itself. Although there is adequate specificity for the SIP in severely impaired individuals, a large proportion of simulators can produce an SIP. Given that Social Security Disability (SSD) claimants are typically low functioning and also seeking compensation, it was of interest to know whether the SIP can be used to exclude truly low-functioning claimants, or whether SSD claimants identified as malingering also produce the SIP, as the simulators in a recent study of this profile. By comparing the SSD claimants to a group of low-functioning Child Protection (CP) claimants who were motivated to do well in order to get their children returned from State custody, the findings clearly show that the SIP is easily produced in criterion-malingerers, but not in those low-functioning CP claimants motivated to do well.

Introduction

The development of special profiles in Green's performance validity tests (PVTs) (seeLarrabee, 2012 for usage change from Symptom Validity Tests to PVTs) to delineate true impairment from deliberate failure (Green, 2003, 2004, 2008; Howe & Loring, 2009) may be helpful in policy initiatives with the Social Security Administration (SSA; Chafetz, 2010, 2012), where there is concern about the failure of PVTs as due to impairment. Initially called a “Dementia Profile” (Green, 2004; Howe & Loring, 2009), this special configuration of scores on the Medical Symptom Validity Test (MSVT; Green, 2004) has more recently been termed a “severe impairment profile” (SIP; Carone, 2009) to recognize the broader types of impairment that can produce a particular pattern of scores.

The SSA concern that low-functioning claimants might fail PVTs due to impairment (e.g., lack of understanding of task demands, lack of ability) has been appreciated in Chafetz (2010, 2011b), who suggested that the use of the SIP might help delineate truly impaired claimants from those who are feigning impairment for disability compensation, thereby reducing the false-positive (FP) identification of feigning. As Howe and Loring (2009) have stated, the primary purpose of the SIP when the MSVT is used as a PVT is to reduce FPs, but a secondary purpose is to provide clinically relevant information about an examinee's functioning.

Previous research has shown adequate specificity of 90.5% for the SIP in severely impaired (i.e., demented) versus normal-functioning individuals (Howe & Loring, 2009). In this context, specificity refers to the percentage of people with normal cognitive functioning not obtaining the SIP. Thus, in dementia work-ups, the MSVT would be a useful diagnostic tool, helping decide when the memory impairment is so severe that recognition memory becomes involved, and cued and free recall are severely impaired.

In low-functioning Social Security Disability (SSD) claimants, however, the goal is to meet very strict disability requirements (Chafetz, 2011b). The SSD claimant might produce the SIP while attempting to appear very impaired. Indeed, Singhal, Green, Ashaye, Shankar, and Gill (2009) also included a simulator group, in which 40% of simulators obtained the SIP while attempting to simulate memory impairment on the MSVT. Thus, it is reasonable to suspect that a large proportion of SSD claimants who are motivated to appear impaired may actually produce the SIP.

From another point of view, Axelrod and Schutte (2010) have shown in a Veterans Administration Medical Center setting that the algorithm that creates an SIP does not differentiate the SIP subjects from “poor effort” subjects (who merely fail the effort subtests of the MSVT) in terms of overall neuropsychological findings. However, the primary limitation of their study was in not identifying which of their subjects were seeking compensation.

The present study is concerned with whether individuals motivated to appear intellectually disabled can produce an SIP on the MSVT. By comparing the SSD group with a group of Child Protection (CP) claimants with an opposite motivation–to perform well in order to have their children returned from State custody—we hoped to gain an understanding of how motivation affects impaired individuals' performance on the SIP. The purpose if this study was to determine whether the SIP was practically useful in a forensic context in separating truly impaired claimants from those who are feigning impairment.

Methods

Archived Records

Archived records from SSD referrals by the local Disability Determinations Services (DDS) office were used. Low cognitive functioning is a frequent feature of these referrals (Chafetz, Abrahams, & Kohlmaier, 2007; Chafetz, Prentkowski, & Rao, 2011), and only those subjects whose IQ <76 were used in the present study. These previously described subjects were used for the development of the Symptom Validity Scale (SVS) for low-functioning individuals (Chafetz et al., 2007) and for base-rate research (Chafetz, 2008). The adult sample of this study includes 50 subjects who obtained scores on the Wechsler Adult Intelligence Scale-III, MSVT, A-Test (see below), and the SVS for low-functioning individuals (see below) during the course of their examinations. These 50 subjects all had IQ < 76, with 7 of the 50 participants (14%) in this group performing at significantly below-chance levels on at least one of the MSVT recognition subtests.

A highly motivated CP group was also analyzed, consisting of parents sent from the Department of Children and Family Services (DCFS) seeking to have their children returned from State custody (Chafetz et al., 2011). This group consisted of 33 claimants who completed all the tests (all with IQ < 76) who were recently analyzed in another study (Chafetz & Biondolillo, 2012) for other purposes. Table 1 shows the demographics of the sample subjects for both groups. The IQ differences between the SSD group and the CP group were statistically significant (t = 4.37, p < .001, d = 0.97), as recognized previously (Chafetz et al., 2011), but no significant differences were observed when controlling for effort in the SSD group by comparing an SSD subgroup with no evidence for malingering (NE) with the CP group (t = 0.7, d = −0.24). The section on criterion groups below will further explain these subgroups.

Table 1.

Sample characteristics for SSD and CP groups

 SSD CP t (pCM SE NE 
Sample n 50 33 — 18 14 18 
Gender 29-M, 21-F 7-M, 26-F — 10-M, 8-F 8-M, 6-F 11-M, 7-F 
Ethnicity 43-A, 7-C, 0-O 14-A, 14-C, 5-O — 15-A, 3-C 14-A, 0-C 14-A, 4-C 
Age 30.52 (12.22) 32.88 (9.60) 0.93 (.35) 36.1 (12.6) 28.4 (11.7) 26.7 (10.8) 
Education 9.26 (2.02) 9.45 (1.84) 0.45 (.66) 8.3 (2.7) 9.9 (1.2) 9.7 (1.5) 
FSIQ 61.12 (7.43) 67.79 (5.70) 4.37*** 55.3 (7.6) 61.6 (4.6) 66.6 (4.1) 
t-test — — — 7.5* 3.4* 0.7 
Cohen's d — — — −1.9 −1.2 −0.2 
 SSD CP t (pCM SE NE 
Sample n 50 33 — 18 14 18 
Gender 29-M, 21-F 7-M, 26-F — 10-M, 8-F 8-M, 6-F 11-M, 7-F 
Ethnicity 43-A, 7-C, 0-O 14-A, 14-C, 5-O — 15-A, 3-C 14-A, 0-C 14-A, 4-C 
Age 30.52 (12.22) 32.88 (9.60) 0.93 (.35) 36.1 (12.6) 28.4 (11.7) 26.7 (10.8) 
Education 9.26 (2.02) 9.45 (1.84) 0.45 (.66) 8.3 (2.7) 9.9 (1.2) 9.7 (1.5) 
FSIQ 61.12 (7.43) 67.79 (5.70) 4.37*** 55.3 (7.6) 61.6 (4.6) 66.6 (4.1) 
t-test — — — 7.5* 3.4* 0.7 
Cohen's d — — — −1.9 −1.2 −0.2 

Notes: SSD = Social Security Disability; CP = Child Protection; CM = criterion malingering group from SSD; SE = some evidence for malingering (failing SVS or A-Test) from SSD; NE = no evidence for malingering from SSD; FSIQ = Full-Scale IQ; M = male; F = female; A = African American; C = Caucasian; O = Other; For age, education, and FSIQ, the mean (SD) are reported.

t-test for IQ comparisons for each SSD criterion group with the CP group; Cohen's d for IQ comparisons for each SSD criterion group with the CP group.

*p < .05.

**p < .01.

***p < .001.

Permissions

At the consultation, when the claimants signed the HIPAA notification, they were asked for permission to use scores from their examination for research. They were assured that their identity would be protected. Permission was granted if the claimant initialed next to the research notification. Only one claimant declined and her scores were not used. The author presented preliminary findings of this research at a statewide Louisiana DDS meeting (28 April 2003), and the local DDS Medical Liaison Officer provided written acknowledgment of this research. Moreover, DCFS workers who have presented with the claimants in CP cases have all initialed their consent to the research.

Procedures

All participants underwent a standard interview, mental status examination, and testing procedures for the SSD or for the CP examination. Validated indicators from the SVS (Chafetz et al., 2007) were administered during the mental status examination and gathered on the Wechsler scale. The data were analyzed to determine those individuals identified as failing two PVTs.

A-Test

The A-Test was incorporated into the mental status examination of Strub and Black (1993) as a bedside auditory continuous performance task administered by the examiner, not a computer. A full description of the A-Test and its use as a PVT is provided in Chafetz (2012). In this study, the combined total of A-Test errors (omission plus commission) of >2 was seen to have sufficient sensitivity (72%) and specificity (92%).

Symptom Validity Scale

The SVS (originally named the Malingering Rating Scale) has 11 items rated from the mental status examination and Wechsler scale testing during a Psychological Consultative Examination for SSD. The items were developed from experience and observations of these low-functioning claimants. For example, Chafetz and colleagues (2007) showed how it was necessary to eliminate a validity item (the Vocabulary–Digit Span difference) that had been validated in normal-functioning individuals (Mittenberg, Theroux, Zielinski, & Heilbronner, 1995) because this variable was ineffective in assessing validity in low-functioning claimants. The remaining items included observations associated with poor quality of effort, for example, Ganser-like answers (Chafetz et al., 2007; Drob & Meehan, 2000), which belie the truth by being consistently close to the right answer (e.g., 2 + 3 = 6; 3 + 4 = 8); and coding errors which often showed both right-left reversals and top-bottom reversals, where both kinds of errors together are atypical. The development of the SVS items is detailed in Chafetz and colleagues (2007). Musso, Barker, Jones, Roid, and Gouvier (2011) and Shandera and colleagues (2010) cited Chafetz and colleagues (2007) as the only study (prior to the Musso study) to date providing the development of an SVS specifically relevant to the question of motivated feigning of low IQ in an already low-functioning group. An SVS total score of >6 was seen to have sufficient sensitivity (78%) and specificity (91%; Chafetz, 2010).

Criterion-malingering groups

The Slick, Sherman, and Iverson (1999) guidelines establish four criteria for the diagnosis of malingered neurocognitive dysfunction: (A) The presence of a substantial external incentive; (B) evidence from neuropsychological testing; (C) evidence from self-report; and (D) evidence from testing and self-report are not fully accounted for by psychiatric, neurological, or developmental factors. According to the guidelines, malingering is considered probable when criteria A and D are met and there are two or more types of criterion B present (or one of B and one of C). The Larrabee, Greiffenstein, Greve, and Bianchini (2007) modification of these criteria suggests that failure on two PVTs is satisfactory for the identification of probable malingering and thus the satisfaction of these conditions represent the criterion-malingering (CM) group. In this case, failure of both the SVS and the A-test satisfied the criteria for the CM group. Failure on one of these tests alone was considered to provide some evidence (SE) for malingering. The remaining group constituted the NE group.

Medical Symptom Validity Test and the SIP

The MSVT (Green, 2004) is a forced-choice test consisting of 10 pairs of words presented twice to each claimant on a computer screen. Claimants are then asked to choose the correct target words from pairs consisting of a target and a foil. Due to low reading ability reported by many claimants, the examiner read all the words and choices on the screen to each participant, effecting a combined computer–oral administration. MSVT measures include Immediate Recognition (IR), Delayed Recognition (DR), Consistency (Cn), Paired Associate Recall (PA), and Free Recall (FR).

The requirements for the SIP, called the Dementia Profile by Howe and Loring (2009), are: (1) Failure (<90%) on a symptom validity indicator (IR, DR, or Cn); (2) no effort scores significantly below chance (<30%); (3) the mean of the easy tests (IR, DR, and Cn) at least 20 points more than the mean of the hard tests (PA and FR); (4) IR and DR both greater than FR; and (5) PA > FR. The latter two requirements—4 and 5—are considered an analysis of order violations. Individuals who produce low scores on a symptom validity indicator of the MSVT (<90%) but fail to meet the other four criteria are considered to exhibit the poor quality of effort. Using all these criteria represents the most conservative estimate of the SIP and will be denoted SIP5. Using the (at least) 20-point difference between the means of easy and hard subtests—in addition to failure on IR, DR, or Cn, with no scores significantly below chance—will represent the core of the criteria (SIP3), which has been found to be more specific in impaired individuals (Singhal et al., 2009). The SIP3 removes only the order violation requirements imposed by Criteria 4 and 5 of the SIP5.

Results

Comparing the Differently Motivated Groups

Table 2 shows that 3 of 33 (9%) CP claimants (attempting to present well) obtained the SIP5, whereas 11 of 50 (22%) of the SSD claimants (attempting to obtain disability benefits) obtained the SIP5. This difference is not significant (χ2 = 1.63, n.s.). Table 2 also shows that 18 of 50 (36%) of SSD claimants and 3 of 33 (9%) of CP claimants obtained the SIP3, a significant difference (χ2 = 6.26, p < .05). A noteworthy point is that the three CP claimants who obtained the SIP had IQs of 56, 56, and 59. None of the CP claimants with IQ in the range of 60–75 obtained the SIP at either level.

Table 2.

Comparison of differently motivated (CP and SSD) groups on frequencies of SIP profiles and of malingering identification

 SIP5 SIP3 CM SE At least NE 
CP 3/33 (9%) 3/33 (9%) 0/33 (0%) 3/33 (9%) 3/33 (9%) 30/33 (91%) 
SSD 11/50 (22%) 18/50 (36%) 18/50 (36%) 14/50 (28%) 32/50 (64%) 18/50 (36%) 
χ2 (Yates) 1.53 6.26* 13.12*** 3.28 22.4*** 22.4*** 
 SIP5 SIP3 CM SE At least NE 
CP 3/33 (9%) 3/33 (9%) 0/33 (0%) 3/33 (9%) 3/33 (9%) 30/33 (91%) 
SSD 11/50 (22%) 18/50 (36%) 18/50 (36%) 14/50 (28%) 32/50 (64%) 18/50 (36%) 
χ2 (Yates) 1.53 6.26* 13.12*** 3.28 22.4*** 22.4*** 

Notes: SSD = Social Security Disability; CP = Child Protection; SIP = severe impairment profile; CM = criterion malingering; SE = some evidence for malingering (fail one PVT); At least (some evidence for malingering) = CM + SE; NE = no evidence for malingering. The results clearly show the differences between the differently motivated groups in the frequency of malingering identification and the production of the SIP. Note that the At least and the NE analyses are the same but are shown for completeness. Also note that the p-value for the SE comparison is .07.

*p < .05.

**p < .005.

***p < .001.

Concerning the CM group, 18 of 50 (36%) of SSD claimants and 0 of 33 (0%) of CP claimants met criteria for identification of malingering (χ2 = 13.12, p < .001). The SE group for malingering was obtained when claimants failed one of the criterion PVTs (SSD, 14/50 = 28%; CP, 3/33 = 9%). The observed frequency difference is not significant (χ2 = 3.28, p = .07). If one compares the differently motivated groups concerning the frequencies for “at least SE” (SE + CM), the frequency difference between SSD (32/50 = 64%) and CP (3/33 = 9%) is significant (χ2 = 22.4, p < .001). The demographic characteristics of these additional groups are shown in Table 1.

Comparing Groups with Varying Evidence for Malingering on Obtaining SIP

Table 3 presents a validation study in the SSD and CP claimants comparing those having met criteria for malingering (CM), those with SE for malingering (failing one PVT), and those with NE in relation to their frequencies of obtaining the SIP5 and SIP3 profiles. The table shows that 27.8% of the CM group, 35.7% of the SE group, and only 5.6% of the NE group obtain the SIP5. When these frequencies are compared with that in the CP obtaining the SIP5, the frequency differences are not significant, though the comparison between SE and CP frequencies for SIP5 approach significance (p = .07), as does the comparison for at least SE versus CP (p = .05). As might be expected, the NE group (1/18 = 5.6%) is quite similar to the CP group (3/33 = 9%) in frequency of SIP5.

Table 3.

Percentages of claimants in the CP, CM, SE, and NE groups with identified SIP

Claimant groups SIP5 Yates χ2 versus CP SIP3 Yates χ2 versus CP 
CP 3/33 (9%) — 3/33 (9%) — 
CM 5/18 (27.8%) 1.83 (p = .18) 9/18 (50%) 8.7 (p = .003)* 
SE 5/14 (35.7%) 3.2 (p = .07) 8/14 (57.1%) 10.1 (p = .002)* 
At least (CM + SE) 10/32 (31.3%) 3.7 (p = .05) 17/32 (53.1%) 12.8 (p = .000)* 
NE 1/18 (5.6%) 0.2 (p = .65) 1/18 (5.6%) 0.2 (p = .65) 
Claimant groups SIP5 Yates χ2 versus CP SIP3 Yates χ2 versus CP 
CP 3/33 (9%) — 3/33 (9%) — 
CM 5/18 (27.8%) 1.83 (p = .18) 9/18 (50%) 8.7 (p = .003)* 
SE 5/14 (35.7%) 3.2 (p = .07) 8/14 (57.1%) 10.1 (p = .002)* 
At least (CM + SE) 10/32 (31.3%) 3.7 (p = .05) 17/32 (53.1%) 12.8 (p = .000)* 
NE 1/18 (5.6%) 0.2 (p = .65) 1/18 (5.6%) 0.2 (p = .65) 

Notes: CP = Child Protection; CM = criterion malingering; SE = some evidence for malingering (fail 1 PVT); NE = no evidence for malingering; At least = at least some evidence for malingering (SE + CM groups).

*p < .05.

**p < .01.

***p < .001.

The table also shows that 50% of the CM group, 57.1% of the SE group, and 5.6% of the NE group obtain the SIP3. The at least SE group obtained the SIP3 at a 53.1% rate. Each of these frequency differences is significantly different from the CP group (none of whom was identified as malingering).

Contrasting IQ Levels for Those Obtaining SIP

Table 4 shows the mean IQ levels of the SSD criterion groups and the CP group for only those subjects obtaining the SIP5 or SIP3. Due to the IQ confound (correlation) with malingering (Chafetz et al., 2007), the CM groups cannot be statistically compared with the CP group. Moreover, the CM groups cannot be statistically compared with the NE group, as the latter only contains one subject obtaining the SIP. Nevertheless, one can see that in the CP group, the only subjects obtaining the SIP had IQ < 60. None of these met criteria for malingering. In the SSD criterion groups, there appears to be a linear relationship between malingering criteria and IQ, with the only subject with NE showing the highest IQ of 73. We contrast these results with those presented in Table 1, showing that the mean IQ levels in the CM and SE groups, but not the NE group, are significantly different from the that in the CP group. The IQ levels in the two groups with NE are equivalent in the mid-60's: CP, 67.79 (5.70) and NE, 66.6 (4.1).

Table 4.

IQ Scores (mean + SD) of claimants in the CP, CM, SE, and NE groups with identified SIP

Claimant groups SIP5 SIP3 
CP 57.0 (1.7) 57.0 (1.7) 
CM 60.8 (5.2) 56.1 (7.3) 
SE 62.4 (5.4) 60.4 (5.0) 
NE 73 73 
Claimant groups SIP5 SIP3 
CP 57.0 (1.7) 57.0 (1.7) 
CM 60.8 (5.2) 56.1 (7.3) 
SE 62.4 (5.4) 60.4 (5.0) 
NE 73 73 

Notes: CP = Child Protection; CM = Criterion malingering; SE = some evidence for malingering (fail 1 PVT); NE = no evidence for malingering.

*p < .05.

**p < .01.

***p < .001.

Discussion

The senior author has been concerned with providing SSA (through the DDSs) as a tool from scientific psychology to assist in separating out invalidity from true impairment (Chafetz, 2010, 2011a). Following Green (2004) and Howe and Loring (2009), the idea was to utilize an SIP to cull those claimants who were truly impaired from those who were providing non-credible responding. This current study sought to determine the feasibility of this profile method for delineating true impairment in claimants.

The principal finding in this study is that a large proportion of compensation-seeking SSD claimants produce the SIP with the more or less restrictive criteria (SIP5 or SIP3). Of the SSD claimants, 22% obtain the SIP5 and 36% the SIP3, while only 9% of the well-motivated CP claimants obtain the SIP5 or the SIP3. These frequency differences between the well-motivated CP group and the compensation-motivated SSD group are significant for the SIP3 but not the SIP5.

We note that the SSD group has significantly more criterion-malingerers (36%) than the CP group (0%) and significantly more claimants with at least SE for malingering (64% vs. 9%).

If we now just take the SSD claimants, separating them into groups of CM, SE for malingering, and NE, we see that the NE group has 5.6% claimants meeting criteria for the SIP5 and SIP3 profiles. The CM group has 27.8% claimants meeting SIP5 and 50% claimants meeting SIP3 criteria. If we combine the frequencies for the SE and SM groups (at least SE for malingering), then 31.3% of claimants meet the SIP5 criteria and 53.1% meet the criteria for SIP3. Considering the SIP3, the CM groups all show a significantly higher frequency of this impairment profile than the CP group, while the group with NE is similar to the CP group, which also does not include anyone meeting the criteria for malingering.

While we note that because the SIP5 has more criteria (order violations) than the SIP3 and thus eliminates more claimants (suggesting that they have exhibited evidence for malingering), it is clear that almost one third of the compensation-seeking claimants who have at least SE for malingering obtain the SIP5 and 53% the SIP3. Moreover, considering how low the sample sizes are, if we combine the groups with no malingerers (CP and NE) and then do the comparisons (not shown), we see that the SE group (χ2 = 5.0, p < .05) and the at least SE group (χ2 = 6.1, p < .05) show significantly more SIP5 profiles than the combined no-malingering groups. Thus, the results clearly show that criterion-malingerers (and those compensation seeking claimants with at least SE for malingering) produce an intolerably high level of impairment profiles. In criterion-malingerers, it is the appearance of impairment that is being feigned with the SIP.

But are those claimants really impaired? We see that 9% of the well-motivated CP group produces these impairment profiles, none of whom is malingering. In fact, none of the CP claimants with IQ levels between 60 and 75 even produce the SIP; the three CP claimants who obtained the SIP all had IQ < 60. Thus, it is not likely that the compensation-seeking SSD claimants who are producing this profile are truly impaired, and in fact, we see that only one of the NE group SSD claimants did so (5.6%), but this person had an IQ of 73! Given that this claimant was in the compensation-seeking SSD group, we suggest that this claimant was a false negative for the SVS and the A-Test and that the failure on the MSVT actually was a hit for malingering, rather than showing severe impairment.

How do we know this? The Chafetz and Biondolillo (2012) analysis shows that the well-motivated CP claimants simply do not fail PVTs (including SVS, A-Test, MSVT, and reliable digit span (RDS)). For the IQ range of 60–75, there was a 100% pass rate for these claimants on the MSVT, SVS, and A-Test, with a 94% pass rate on the RDS (two failures with cutoff < 6).

Our current study thus establishes that in a well-motivated low-IQ sample, “impairment” is not identified by the SIP unless IQ dips below 60. It is interesting to note that in the Blue Book listings used by the DDSs to delineate disability (US Department of Health and Human Services, 1994, revised 2006), the listing for disability for mental retardation (12.05) does not require other restrictions or limitations when the IQ is below 60. Thus, SSA already recognizes that an IQ below 60 represents a qualitatively different level of impairment, consistent with these findings.

Given how low the IQ must be before the SIP adequately identifies true impairment (in well-motivated subjects), we must now revisit the algorithm in the AI program (Green, 2011) to decide true impairment versus poor effort. If the criteria for the SIP3 are met (failing easy subtests, not below chance, and at least a 20-point difference between the means of the easy [IR, DR, and Cn] and hard [PA and FR] subtests), then the decision arrow points to the possible genuine memory impairment profile (called SIP here to distinguish it from the Genuine Memory Impairment Profile [GMIP] in which the recognition tasks are passed, but there is memory impairment on the memory subtests). The evaluator is warned by the AI program to be prepared to diagnose dementia or an equivalent condition, and if none of those conditions is possible, then poor effort needs to be considered.

This algorithm has been criticized by Axelrod and Schutte (2010) for applying circular reasoning. Indeed, illustrating the circularity, one often finds statements on list-serves to the effect: “The impairment profile is only valid if the claimant has a true impairment.” In the context of this criticism, we note that the D criterion from Slick and colleagues (1999) states that “Behaviors meeting necessary criteria from groups B or C are not fully accounted for by Psychiatric, Neurological, or Developmental Factors.” The Slick and colleagues (1999) algorithm operates in a similar manner, albeit in the other direction, mitigating the inference of malingering if there is true impairment from another source. The problem with the Green (2011) algorithm, however, is that the test profile itself is used to identify true impairment, which is then overturned only if the claimant does not have true impairment. That is the circularity.

This current study offers a way out of the circularity. Once the easy subtests are failed (and not at a below-chance level), and the sufficient easy–hard difference is obtained (to make it an SIP profile), there should be arrows going to Evaluation Context boxes. If the evaluation context includes the compensation/avoidance of punishment (i.e., secondary gain), then another arrow should point to a probable poor effort inference. In other words, in a secondary gain context, the obtaining of the SIP profile would be considered further evidence adduced for malingering, much like simple failure on the MSVT. If the evaluation context is clinical or post-adjudication or any other non-compensation/avoidance of punishment context, the arrow should point to a neurological criteria box in order to evaluate whether IQ is <60, or a severe enough dementia is present. This is important, because the current study shows that one cannot trust the SIP to delineate impairment in a compensation context.

Axelrod and Schutte (2010) have shown in a Veterans Administration medical center setting that the algorithm creating an SIP does not differentiate the SIP subjects from the poor effort subjects in terms of the overall neuropsychological findings. They cite the Green (2004) database in which both Depression groups failing the MSVT have average scores that meet the SIP criteria. In this database, 33% (20 of 61) of one depression sample and 40% (19 of 47) of another depression sample failed the MSVT. It is not possible to tell how impaired these depression groups were, but given our findings we suspect that this is a matter of poor performance quality (invalidity) rather than true impairment.

Moreover, Axelrod and Schutte (2010) did not otherwise identify which of their subjects were seeking compensation, and so the relative frequencies in each of their groups are unknown. That is the main limitation of their study. The current study sought to improve on the Axelrod and Schutte (2010) study by tracing the findings in the criterion-malingerers, but we note that our findings are consistent with theirs.

The GMIP is a term often used to refer to an SIP (Green, 2011) or to failure of PA and FR in persons who do not fail the MSVT effort subtests (Howe & Loring, 2009). Following Howe and Loring (2009), we prefer to use the GMIP as referring to only those individuals who fail the memory subtests without failing the effort (validity) subtests. Supposedly, this indicates actual memory impairment, without the slippage of the recognition memory tasks, a more severe condition. This will obviously require more study, but it is likely that the community would benefit from delineating the GMIP and the SIP.

According to Howe and Loring (2009), the two main purposes for using an SIP for Green's MSVT are (a) to reduce the FP rate in low-functioning individuals when the purpose is to assess validity and (b) to provide clinically relevant information about a person's functioning when the purpose is to assess impairment. Low IQ by itself does not cause PVT failure (Chafetz & Biondolillo, 2012; Chafetz et al., 2011), though very low IQ (<60) may cause an SIP to occur even in well-motivated claimants. When the SIP occurs in claimants who are motivated to fail, it is more likely due to the feigning of impairment rather than actual impairment. However, in a dementia or intellectual disability study in which the motivation of the subjects is known, we see that a combination of the SIP and the GMIP may aid in defining the impairment.

One limitation of this study is the small sample sizes, and we hope that a replication will be attempted with larger sample sizes. We would also encourage a replication of this study with other of Dr Green's tests.

Funding

No funding was obtained for this research.

Conflict of Interest

None declared.

Acknowledgements

We are especially indebted to a reviewer who clearly spent a lot of time with this paper and who offered many helpful suggestions.

References

Axelrod
B. N.
Schutte
C.
Analysis of the dementia profile on the Medical Symptom Validity Test
The Clinical Neuropsychologist
 , 
2010
, vol. 
24
 (pg. 
873
-
881
)
Carone
D.
Test review of the Medical Symptom Validity Test
Applied Neuropsychology
 , 
2009
, vol. 
16
 (pg. 
309
-
311
)
Chafetz
M. D.
Malingering on the Social Security disability consultative examination: Predictors and base rates
The Clinical Neuropsychologist
 , 
2008
, vol. 
22
 (pg. 
529
-
546
)
Chafetz
M. D.
Symptom validity issues in the psychological consultative examination for Social Security disability
The Clinical Neuropsychologist
 , 
2010
, vol. 
24
 (pg. 
1045
-
1063
)
Chafetz
M. D.
Reducing the probability of false positives in malingering detection of Social Security disability claimants
The Clinical Neuropsychologist
 , 
2011a
, vol. 
25
 (pg. 
1239
-
1252
)
Chafetz
M. D.
The psychological consultative examination for Social Security disability
Psychological Injury and Law
 , 
2011b
, vol. 
4
 (pg. 
235
-
244
)
Chafetz
M. D.
The A-Test: A symptom validity indicator embedded within a Mental Status Examination for social security disability
Applied Neuropsychology
 , 
2012
, vol. 
19
 (pg. 
121
-
126
)
Chafetz
M. D.
Abrahams
J. P.
Kohlmaier
J.
Malingering on the Social Security disability consultative examination: A new rating scale
Archives of Clinical Neuropsychology
 , 
2007
, vol. 
22
 (pg. 
1
-
14
)
Chafetz
M. D.
Biondolillo
A. M.
Validity issues in Atkins death cases
The Clinical Neuropsychologist
 , 
2012
 
doi:10.1080/13854046.2012.730674
Chafetz
M. D.
Prentkowski
E.
Rao
A.
To work or not to work: Motivation (not low IQ) determines PVT findings
Archives of Clinical Neuropsychology
 , 
2011
, vol. 
26
 (pg. 
306
-
313
)
Drob
S. L.
Meehan
K. B.
The diagnosis of Ganser syndrome in the practice of forensic psychology
American Journal of Forensic Psychology
 , 
2000
, vol. 
18
 
3
(pg. 
37
-
62
)
Green
P.
Manual for the Word Memory Test
 , 
2003
Edmonton
Green's Publishing
 
(Revised 2005)
Green
P.
Manual for the Medical Symptom Validity Test
 , 
2004
Edmonton
Green's Publishing
Green
P.
Test manual for the Nonverbal Medical Symptom Validity Test
 , 
2008
Edmonton
Green's Publishing
Green
P.
Advanced interpretation (AI) program
 , 
2011
Edmonton
Green's Publishing
Howe
L. L. S.
Loring
D. W.
Classification accuracy and predictive ability of the Medical Symptom Validity Test's Dementia Profile and General Memory Impairment Profile
The Clinical Neuropsychologist
 , 
2009
, vol. 
23
 (pg. 
329
-
342
)
Larrabee
G. J.
Performance validity and symptom validity in neuropsychological assessment
Journal of the International Neuropsychological Society
 , 
2012
, vol. 
18
 (pg. 
1
-
7
)
Larrabee
G. J.
Greiffenstein
M. F.
Greve
K. W.
Bianchini
K. J.
Larrabee
G. J.
Refining diagnostic criteria for malingering
Assessment of malingered neuropsychological deficits
 , 
2007
New York
Oxford University Press
Mittenberg
W.
Theroux
S.
Zielinski
R. E.
Heilbronner
R. L.
Identification of malingered head injury on the Wechsler adult intelligence scale-revised
Professional Psychology: Research and Practice
 , 
1995
, vol. 
26
 (pg. 
491
-
498
)
Musso
M. W.
Barker
A. A.
Jones
G. N.
Roid
G. H.
Gouvier
W. D.
Development and validation of the Stanford Binet-5 Rarely Missed Items-Nonverbal Index for the detection of malingered mental retardation
Archives of Clinical Neuropsychology
 , 
2011
, vol. 
26
 (pg. 
756
-
767
)
Shandera
A. L.
Berry
D. T. R.
Clark
J. A.
Schipper
L. J.
Graue
L. O.
Harp
J. P.
Detection of malingered mental retardation
Psychological Assessment
 , 
2010
, vol. 
22
 (pg. 
50
-
56
)
Singhal
A.
Green
P.
Ashaye
K.
Shankar
K.
Gill
D.
High specificity of the Medical Symptom Validity Test in patients with very severe memory impairment
Archives of Clinical Neuropsychology
 , 
2009
, vol. 
24
 (pg. 
721
-
728
)
Slick
D. J.
Sherman
E. M. S.
Iverson
G. L.
Diagnostic criteria for malingered neurocognitive dysfunction: Proposed standards for clinical practice and research
The Clinical Neuropsychologist
 , 
1999
, vol. 
13
 (pg. 
545
-
561
)
Strub
R. L.
Black
F. W.
The Mental Status Examination in Neurology
 , 
1993
3rd ed.
Philadelphia
F. A. Davis
US Department of Health and Human Services, Social Security Administration
1994, revised 2006
 
Disability Evaluation Under Social Security, SSA Publication no. 64-039