The Rey Auditory Verbal Learning Test Effort Equation: A Cross-validation Study.

摘要

The purpose of this study was to cross-validate the utility of the RAVLT effort equation as an embedded measure to detect malingering. The original study (Boone et al., 2005) found that the RAVLT effort equation had a sensitivity of 73.8% and specificity of 90% when using a cutoff of ≤12. The current study examined credible (n=114) and noncredible (n=144) groups and found slightly less robust results, with sensitivity of 59% and specificity of 90% when using a cutoff of ≤10. When cut-off scores were selected to maintain 90% specificity in credible subjects, the most sensitive scores (>50% sensitivity) were recognition correct (70.1% sensitivity), the effort equation (59.0% sensitivity), and total score across learning trials (52.8% sensitivity). Thus, while the effort equation was found to be more effective in the original study, in the current investigation, the pure recognition score was more sensitive, suggesting that it may be the preferable score when using the RAVLT as an SVT. Women scored higher than men on some RAVLT scores, and some cut-offs could be made more stringent in women, while five scores had to be made more lenient for men. Interestingly, educational level appeared to moderate underperformance in noncredible subjects; specifically, less educated noncredible subjects appeared to fake more blatantly on the RAVLT; when cut-offs were adjusted to insure 90% specificity was maintained in credible subjects with <12 years of education, sensitivity rates in noncredible less educated subjects averaged 55.4%, while sensitivity rates in better educated noncredible subjects (≥12 years education) averaged 32.5%.