Impact of simulating borderline personality disorder on the MMPI-2: a costs-benefits model employing base rates.

This investigation extends current work on the impact of role-specific response sets on dissimulated Minnesota Multiphasic Personality Inventory-2 (MMPI-2; Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989) profiles. Specifically, one group of subjects was given specific symptom information and instructed to simulate borderline personality disorder on the MMPI-2. This dissimulation performance was compared to a clinical comparison group of clients diagnosed with borderline personality disorder, a group of subjects instructed to simulate major depression, and a group of subjects who responded honestly to the MMPI-2. In both simulating conditions, subjects produced significantly higher elevations on a majority of clinical and validity scales than the clinical comparison group. A rigorous evaluation of the F scale was conducted to provide a range of predictive values for different F scale cutting scores for detecting malingering. Sensitivity, specificity, and positive and negative predictive power were computed for this range of F scale values. A similar analysis was conducted for the Borderline Personality Disorder Scale (BPDsc) of Morey, Waugh, and Blashfield (1985). These results are integrated with extant research. A costs-benefits model is proposed for employing F scale cutting scores in clinical situations with different base rates of malingering.