Comparing the Psychometric Properties of Eight Embedded Performance Validity Tests in the Rey Auditory Verbal Learning Test, Wechsler Memory Scale Logical Memory, and Brief Visuospatial Memory Test–Revised Recognition Trials for Detecting Invalid Neuropsychological Test Performance. (December 2021)

• Record Type: Journal Article
• Authors: Pliskin, Joshua I.; DeDios Stern, Samantha; Resch, Zachary J.; Saladino, Kevin F.; Ovsiew, Gabriel P.; Carter, Dustin A.; Soble, Jason R.
• Is Part Of: Assessment. Volume 28:Number 8(2021)
• Page Start: 1871
• Page End: 1881
• Publication Date: 2021-12
• Subjects: performance validity -- neuropsychology -- assessment -- memory -- psychometrics; Psychological tests -- Periodicals; Psychodiagnostics -- Periodicals; Personality tests -- Periodicals; 150.287
• DOI: 10.1177/1073191120929093
• Languages: English
• ISSNs: 1073-1911

Abstract

This cross-sectional study evaluated eight embedded performance validity tests (PVTs) previously derived from the Rey Auditory Verbal Learning Test (RAVLT), Wechsler Memory Scale–IV–Logical Memory (LM), and Brief Visuospatial Memory Test–Revised (BVMT-R) recognition trials among a single mixed clinical sample of 108 neuropsychiatric patients (83 valid/25 invalid) with ( n = 54) and without ( n = 29) mild neurocognitive disorder. Among the overall sample, all eight recognition PVTs significantly differentiated valid from invalid performance (areas under the curve [AUCs] = .64-.81) with 26% to 44% sensitivity (≥89% specificity) at optimal cut-scores depending on the specific PVT. After subdividing the sample by cognitive impairment status, all eight PVTs continued to reliably identify invalid performance (AUC = .68-.91) with markedly increased sensitivities of 56% to 80% (≥89% specificity) in the unimpaired group. In contrast, among those with mild neurocognitive disorder, RAVLT False Positives and LM became nonsignificant, whereas the other six PVTs remained significant (AUC = .64-.77), albeit with reduced sensitivities of 32% to 44% (≥89% specificity) at optimal cut-scores. Taken together, results cross-validated BVMT-R and most RAVLT recognition indices as effective embedded PVTs for identifying invalid neuropsychological test performance with diverse populations including examinees with and without suspected mild neurocognitive disorder, whereas LM had more limited utility as an embedded PVT, particularly when mild neurocognitive disorder was present.