The detection of feigning using multiple PAI scale elevations: a new index.

Archival data were collected from 98 male inmates at a psychiatric inpatient unit to examine the utility of the Multiscale Feigning Index (MFI) as a proposed feigning index for the Personality Assessment Inventory (PAI). MFI was compared with existing PAI feigning indices, Malingering (MAL), Negative Impression Management (NIM), and Rogers Discriminant Function (RDF), using performance on the Structured Interview of Reported Symptoms (SIRS) as the feigning criterion. Regression analyses revealed that MFI was a stronger predictor of SIRS outcome than NIM, MAL, and RDF. In addition, NIM, MAL, and RDF did not add substantial incremental validity to MFI in predicting SIRS outcome. Receiver operating characteristic analyses revealed sensitivity of 68.89% and specificity of 94.34% at an MFI cutoff of more than 76, which compared favorably with the utility of NIM, MAL, and RDF.