David T. R. Berry

Detection of malingering on the MMPI: A meta-analysis

Abstract We meta-analytically reviewed 28 studies that used the MMPI to detect malingering. Forty-four comparisons contrasting a malingering and honest group on the raw and T-scaled F score, F — K scale, the original and revised Dissimulation indices, and Subtle/Obvious scales were derived. The overall mean effect size of d = 2.07 suggests that these indices are good at detecting malingering, with the best scales being the T-scaled and raw F (ds = 2.34 and 2.22), the original Dissimulation scale (d = 2.17), and the F — K index (d = 1.89). Effect sizes were much smaller for the revised Dissimulation index and the Subtle/Obvious scales, suggesting caution in their use. Cutting scores for predicting malingering in the individual case were reviewed, with the wide range of cutting scores employed suggesting the need to investigate the appropriate cutting scores for separate use in subpopulations, such as prisoners, disability applicants, military personnel, and psychiatric patients.

Detection of Inadequate Effort on Neuropsychological Testing: A Meta-Analytic Review of Selected Procedures

Thirty-two studies of commonly researched neuropsychological malingering tests were meta-analytically reviewed to evaluate their effectiveness in discriminating between honest responders and dissimulators. Overall, studies using the Digit Memory Test (DMT), Portland Digit Recognition Test (PDRT), 15-Item Test, 21-Item Test, and the Dot Counting Test had average effect sizes indicating that dissimulators obtain scores that are approximately 1.1 standard deviations below those of honest responders. The DMT separated the means of groups of honest and dissimulating responders by approximately 2 standard deviations, whereas the 21-Item Test and the PDRT separated the groups by nearly 1.5 and 1.25 standard deviations, respectively. The 15-Item Test and the Dot Counting Test were less effective, separating group means by approximately 3/4 of a standard deviation. Although the DMT, PDRT, 15-, and 21-Item Tests all demonstrated very high specificity rates, at the level of individual classification, the DMT had the highest sensitivity and overall hit-rate parameters. The PDRT and 15-Item Test demonstrated moderate sensitivity, whereas the 21-Item Test demonstrated poor sensitivity. The less than perfect sensitivities of all the measures included in this review argue against their use in isolation as malingering screening devices.

A review of diagnostic techniques in the differential diagnosis of epileptic and nonepileptic seizures.

The diagnosis of psychogenic nonepileptic seizures (PNES) is complex. Long-term electroencephalogram monitoring with video recording (video EEG) is the most common method of differential diagnosis of epilepsy and PNES. However, video EEG is complex, costly, and unavailable in some areas. Thus, alternative diagnostic techniques have been studied in the search for a diagnostic method that is as accurate as video EEG, but more cost effective, convenient, and readily available. This paper reviews the literature regarding possible diagnostic alternatives and organizes findings into 7 areas of study: demographic and medical history variables, seizure semiology, provocative testing, prolactin levels, single photon emission computed tomography, psychological testing, and neuropsychological testing. For each area, the literature is summarized, and conclusions about the accuracy of the technique as a diagnostic tool are drawn. Overall, it appears unlikely that any of the reviewed alternative techniques will replace video EEG monitoring; rather they may be more successful as complementary diagnostic tools. An important focus for further investigations involves combinations of diagnostic techniques for the differential diagnosis of epilepsy and PNES.

A Comparative Meta-Analysis of Rorschach and MMPI Validity

Two previous meta-analyses concluded that average validity coefficients for the Rorschach and the MMPI have similar magnitudes (L. Atkinson, 1986; K. C. H. Parker, R. K. Hanson, & J. Hunsley, 1988), but methodological problems in both meta-analyses may have impeded acceptance of these results (H. N. Garb, C. M. Florio, & W. M. Grove, 1998). We conducted a new meta-analysis comparing criterionrelated validity evidence for the Rorschach and the MMPI. The unweighted mean validity coefficients (rs) were .30 for MMPI and .29 for Rorschach, and they were not reliably different (p = .76 under fixed-effects model, p = .89 under random-effects model). The MMPI had larger validity coefficients than the Rorschach for studies using psychiatric diagnoses and self-report measures as criterion variables, whereas the Rorschach had larger validity coefficients than the MMPI for studies using objective criterion variables.

Detection of feigned ADHD in college students.

Significant motivations and incentives exist for young-adult students to seek a diagnosis of attention-deficit/hyperactivity disorder (ADHD). With ADHD information readily accessible on the Internet, todays students are likely to be symptom educated prior to evaluation. This may result in false-positive diagnoses, particularly when students are motivated to convey symptoms. We evaluated the utility of ADHD symptom checklists, neurocognitive tests, and measures initially developed to detect feigned neurocognitive or psychiatric dysfunction (symptom validity tests [SVTs]). The performance of 31 undergraduates financially motivated and coached about ADHD via Internet-derived information was compared to that of 29 ADHD undergraduates following medication washout and 14 students not endorsing symptomatology. Results indicated malingerers readily produced ADHD-consistent profiles. Symptom checklists, including the ADHD Rating Scale and Connerss Adult ADHD Rating Scale-Self-Rating Form: Long, were particularly susceptible to faking. Connerss Continuous Performance Test-II findings appeared more related to motivation than condition. Promising results were seen with all cognitive SVTs (Test of Memory Malingering [TOMM], Digit Memory Test, Letter Memory Test, and Nonverbal-Medical Symptom Validity Test), particularly TOMM Trial 1 when scored using Trial 2 criteria. All SVTs demonstrated very high specificity for the ADHD condition and moderate sensitivity to faking, which translated into high positive predictive values at rising base rates of feigning. Combining 2 or more failures resulted in only modest declines in sensitivity but robust specificity. Results point to the need for a thorough evaluation of history, cognitive and emotional functioning, and the consideration of exaggerated symptomatology in the diagnosis of ADHD.

Nocturnal hypoxia and Neuropsychological variables

Hypoxia is a well known cause of brain dysfunction. Neuropsychological impairments have been observed in normal subjects experiencing hypoxia iatrogenically as well as in patients with chronic lung disease. Recent investigations have demonstrated significant nocturnal hypoxia in subjects with sleep-disordered breathing. In the present study, heavy-snoring males, a group known to experience frequent episodes of sleep-disordered breathing received neuropsychological testing and a night of continuous monitoring of respiratory parameters. Partial correlations, controlling for age, weight, and education, indicated reliable relationships between nocturnal hypoxia and measures of general intelligence, verbal and nonverbal memory, and expressive verbal fluency. It is proposed that heavy-snoring males may potentially serve as a population in which to model the neurobehavioral effects of hypoxia. Further research in subjects with sleep-disordered breathing may help clarify the extent of the possible cognitive deficits as well as point out possible ameliorative treatments.

Sensitivity of MMPI-2 validity scales to random responding and malingering.

This study explored the effects of both random responding and malingering on the validity scales of the MMPI-2 in an analog design using 4 groups of college students. One group completed the entire answer sheet randomly, a second group was instructed to malinger a moderate psychological disturbance, a third group was asked to simulate a severe disturbance, and a fourth group was given standard test instructions

Detection of underreporting of psychopathology on the MMPI: A meta-analysis

Abstract We reviewed the empirical literature on detection of underreporting of psychopathalogy on the MMPI. Meta-analytic techniques were applied to 25 studies in which subjects responding honestly were compared to subjects underreporting psychopathology. Several indices of underreporting were examined, including the L and K scales , F- K, L + K, social desirability scales, subtle/obvious scales, and the Positive Malingering scale. An overall mean effect size of 1.05 was obtained, suggesting that subjects underreporting psychopathology differ from subjects responding honestly by approximately 1 standard deviation, on the average, on these scales. The Positive Malingering scale and one of the social desirability scales showed promising effect sizes of approximately 1. 5 standard deviations. For the L and K scales, effect sizes of just under 1 standard deviation were noted. Cutting scores for detection of underreporting on these scales also were examined. A wide variety of cutting scores were employed in different studies, suggesting that consistently effective criteria have yet to be established. Applications of these findings to the recently published MMPI-2 are discussed .

Rey-Osterrieth Complex Figure : Psychometric characteristics in a geriatric sample

Abstract We studied the psychometric characteristics of the Rey-Osterrieth Complex Figure (ROCF) in an aged sample. One hundred and seven healthy white males and females, 50 to 79 years of age, received the ROCF and Taylor Complex Figure (TCF) as well as a screening battery of neuropsychological tests. Interrater, alternate form, test-retest, and internal consistency reliability were adequate to good for recall trials of the ROCF, with the copy trial generally lower in reliability. Validity procedures confirmed the sensitivity of the ROCF to Alzheimers Disease, while factor analysis suggested that the ROCF loaded on a visuospatial perceptual/memory dimension. ROCF scores were significantly correlated with age and years of education, although not with sex. These results are generally supportive of the psychometric quality of the ROCF in an elderly sample. Once adequate normative data become available, the ROCF should prove a useful tool for clinicians in assessing elderly patients.

Response validity in forensic neuropsychology: Exploratory factor analytic evidence of distinct cognitive and psychological constructs

Forensic neuropsychology studies usually address either cognitive effort or psychological response validity. Whether these are distinct constructs is unclear. In 122 participants evaluated in a compensation-seeking context, the present Exploratory Factor Analysis examined whether forced-choice cognitive effort measures (Victoria Symptom Validity Test, Test of Memory Malingering, Letter Memory Test) and Minnesota Multiphasic Personality Inventory, Second Edition (MMPI-2) validity scales (L, F, K, FBS, Fp, RBS, Md, Dsr2, S) load on independent factors. Regardless of factor rotation strategy (orthogonal or oblique), four response validity factors emerged by means of both Principal Components Analysis (82.7% total variance) and Principal-Axis Factor Analysis (74.1% total variance). The four factors were designated as follows: Factor I, with large loadings from L, K, and S--underreporting of psychological symptoms; Factor II, with large loadings from FBS, RBS, and Md-overreporting of neurotic symptoms; Factor III, with large loadings from VSVT, TOMM, and LMT--insufficient cognitive effort; and Factor IV, with the largest loadings from F, Fp, and Dsr2--overreporting of psychotic/rarely endorsed symptoms. Results reflect the heterogeneity of response validity in forensic samples referred for neuropsychological evaluation. Administration of both cognitive effort measures and psychological validity scales is imperative to accurate forensic neuropsychological assessment.