2006 HSR&D National Meeting Abstract
3055 — Racial Differences in Validity of Psychometric Test Data in Veterans Being Evaluated for PTSD
Boggs CD (Durham VAMC VISN-6 MIRECC)
Braxton L (Durham VAMC, Duke University Medical Center)
Beckham JC (Durham VAMC, VISN-6 MIRECC, Duke University Medical Center)
Collie C (Durham VAMC)
Yeatts B (Durham VAMC)
Calhoun PS (Durham VAMC HSR&D, VISN-6 MIRECC, Duke University Medical Center)
Straits-Troster K (Durham VAMC VISN-6 MIRECC)
Butterfield MI (Durham VAMC HSR&D, VISN-6 MIRECC, Duke University Medical Center)
Racial disparities in the awarding of VA service-connected disability for PTSD have been documented, for example African American veterans are less likely to be service-connected for PTSD. The use of psychometric test data is part of the recommended “best practice” guidelines for conducting diagnostic evaluations for PTSD. Psychometric test data is often incorporated into PTSD evaluations to offer additional information regarding psychopathology and the possibility of exaggeration of symptoms or feigning. The current study explored racial differences in validity rates on the Personality Assessment Inventory (PAI) and the Minnesota Multiphasic Personality Inventory-2 (MMPI-2).
Veterans evaluated for PTSD at a VA PTSD clinic completed the PAI, MMPI-2 along with measures of PTSD, combat exposure, and demographic variables. The validity of profiles of African American veterans (n=194) was compared to Caucasian veterans (n=121). Test validity was based upon widely accepted criteria.
A greater percentage of African American veterans produced invalid profiles on the PAI (51.5% vs. 48.5%, p=.05). Racial differences in validity profiles were more pronounced on the MMPI-2 (67% vs. 33%, p=.01). There were no differences on scales of inconsistent responding on either measure. More African American veterans had profiles suggestive of over-reporting on Negative Impression Management (NIM) and Infrequency-Psychopathology (F(p)), scales designed to distinguish over-reporting of psychological symptoms from elevated profiles due to psychological distress. On the PAI, in regression analyses controlling for age, employment, SES, marital status, combat exposure, and PTSD severity, African American race was significantly associated with invalidity due to over-reporting (OR= 1.784, 95%CI 1.012-3.03). When restricting analyses to African American veterans, the PAI appeared to generate fewer invalid profiles than the MMPI-2 (67% vs. 51.5%, p=.000).
In an examination of validity profiles of the PAI and MMPI-2 in veterans being evaluated for PTSD, African American veterans generated more invalid profiles due to elevations on scales suggestive of over-reporting.
More research is needed to examine potential bias in validity rates of the PAI and MMPI-2. The presence of racial bias in validity rates of psychometric tests widely used in VA might contribute to the disproportionate denial of PTSD disability claims in African Americans.