Preliminary examination of sequential dual-threshold classification of performance validity findings

Conventional interpretation of performance validity tests (PVTs) involves tallying failed PVTs across a test battery. Failure of a single PVT reflects a common, but diagnostically ambiguous, indeterminate outcome affecting both clinicians and researchers. This study explored whether dual-threshold classification (DTC), which integrates sensitive and specific cut-scores in a given PVT, would reduce classification uncertainty compared to the PVT failure tally method.

This study used a simulation sample (N = 216) of healthy controls (n = 73) and simulators (n = 143) and a mixed clinical/forensic sample (N = 443). Participants were classified as valid, invalid, or indeterminate across PVT pairs combined using the PVT failure tally method and a sequential, two-PVT implementation of DTC. Classification outcomes were compared at observed invalidity rates and across resampled base rates of invalidity (10%-50%).

Sequential DTC significantly reduced indeterminate findings compared to the PVT failure tally method in 9 of 10 PVT pairs in the simulation sample and in 26 of 28 pairs in the clinical/forensic sample. At observed invalidity base rates, indeterminate classifications decreased from 13.4-30.6% to 0-12.5% in the simulation sample and from 9.9-28.7% to 0-8.1% in the clinical/forensic sample. Indeterminate classifications remained lower using DTC across resampled invalidity base rates. Sequential DTC generally increased identification of criterion-invalid cases, and classification of criterion-valid cases varied across PVT pairs.

These findings provide preliminary support for integration of sensitive and specific PVT cutoffs to reduce diagnostic indeterminacy in performance validity assessment.