Examining independent and combined accuracy of embedded performance validity tests in the California Verbal Learning Test-II and Brief Visuospatial Memory Test-Revised for detecting invalid performance

• Published in: Applied neuropsychology. Adult Vol. 29; no. 2; pp. 252 - 261
• Main Authors: Resch, Zachary J., Pham, Amber T., Abramson, Dayna A., White, Daniel J., DeDios-Stern, Samantha, Ovsiew, Gabriel P., Castillo, Liliam R., Soble, Jason R.
• Format: Journal Article
• Language: English
• Published: United States Routledge 04.03.2022
• Subjects: Adult; Assessment; Cross-Sectional Studies; forensic; Humans; memory; Memory and Learning Tests; Neuropsychological Tests; performance validity; psychometrics; Reproducibility of Results; Sensitivity and Specificity; Verbal Learning; psychometrics; performance validity; memory; forensic; Assessment
• ISSN: 2327-9095; 2327-9109; 2327-9109
• DOI: 10.1080/23279095.2020.1742718

The California Verbal Learning Test-Second Edition (CVLT-II) Forced Choice Recognition (FC) and Brief Visuospatial Memory Test-Revised (BVMT-R) Recognition Discrimination Index (RD) are embedded performance validity tests (PVTs) assessing material-specific neuropsychological processes (i.e., verbal and visual memory, respectively). Prior research demonstrated the utility of these PVTs independently; however, no study has compared their diagnostic accuracy for identifying invalid performance relative to each other and in combination within a single sample. This cross-sectional study included an adult neuropsychiatric sample who underwent neuropsychological evaluation. Validity groups were determined via independent criterion PVT performance, and consisted of 103 participants with valid and 25 with invalid neurocognitive performance. FC and RD were not significantly correlated (r = .154), yet both differed between validity groups (η p 2  = .14-.19). Previously established FC (≤14) and RD (≤4) cutoffs evidenced 32-40% sensitivity/90-98% specificity, though receiver operating characteristic (ROC) analyses indicated a more liberal FC cutoff (≤15) was optimal. Logistic regression models utilizing both embedded PVTs indicated that FC did not significantly improve classification accuracy above and beyond RD. Results support the clinical utility of existing cutoffs for FC and RD for independently identifying invalid performance, though the latter showed relatively better ability to detect invalid performance when both are used together.