Multivariate Base Rates of Standard- and Skyline-Cutoff Elevations on the Personality Assessment Inventory: Do They Distinguish Simulated from Genuine PTSD?

Abstract

Multivariate base rates (MBR) of elevations are an emerging psychometric paradigm for enhanced interpretation of multiscale self-report data. The aims of this study were to calculate and compare MBR of scale/subscale elevations on the Personality Assessment Inventory (PAI) and determine the ability of MBR to differentiate between mood disorders (n = 524, k = 3), military-based posttraumatic stress disorder (PTSD; n = 252, k = 2), and coached PTSD-simulator (n = 160, k = 1) groups. Overall, having at least one standard (T ≥ 70) and skyline elevation on clinical scales and clinical subscales was common across the groups. However, differential abnormal elevation thresholds emerged for each group. For instance, it was unusual (i.e., MBR < 10%) for the mood disorders group to have ≥ 1 (9.7%) and for the genuine PTSD group to have ≥ 3 (9.1%) skyline-elevated clinical scales. For subscales, it was unusual for the mood and PTSD groups to have ≥ 3 (7.6%) and ≥ 7 (8.3%) skyline-elevated clinical subscales, respectively. Conversely, PTSD simulators commonly yielded profiles with standard- and skyline elevations on nearly all clinical scales and subscales. MBR cutoffs identified from receiver-operating characteristic curve analyses yielded robust sensitivity (.650-.806) and specificity (.833-.984) in differentiating genuine PTSD and mood disorder groups from PTSD simulators. MBR are useful in differentiating genuine from simulated psychopathology, consistent with broader scale-based infrequency approaches.