Coordinating Author Information: Megan Keen | Texas Tech University
Session Abstract: The Minnesota Multiphasic Personality Inventory-3 (MMPI-3; Ben-Porath & Tellegen, 2020a,b) was recently released and includes several important revisions, including item-content addition and wording revision to Anxiety [AXY], the premiere measures of PTSD symptoms on the MMPI-3’s predecessor (Sellbom et al., 2021; Wolf et al., 2008). Along with the re-norming of scale content, the revised Anxiety Related Experiences (ARX) scale includes 15 (rather than 5) items. Although the revisions to ARX appear promising for its improved clinical utility based on evidence available in the MMPI-3 Technical Manual, independent investigation is warranted. The current study evaluates the psychometric validity ARX in a sample of college students, examining its concurrent validity using correlation and classification analyses (e.g., sensitivity and specificity) as well as incremental validity through hierarchical regression.
Potential participants were college students (n = 419) recruited from a university participant pool as part of a study on college health. After removing individuals who elevated any MMPI-3 validity scale, individuals in the final sample (n = 347) were mostly white (67.4%), female (66.9%), and averaged 19.97 years of age. Participants were administered the MMPI-2-RF-EX, from which both the MMPI-2-RF and MMPI-3 may be scored (Hall et al., 2021), as well as a Qualtrics survey evaluating elements of college student health and internalizing psychopathology, including a focus on trauma.
We evaluated ARX’s relationship to criterion measures of PTSD and evaluated the impact of changes made to the ARX scale in the MMPI-3 revision. Specifically, we (1) correlated ARX with the PTSD symptom Checklist (PCL-5) total and cluster scores as well as total scores on the Severity of Acute Stress Symptoms (SASSS) and Life Events Checklist (LEC), (2) calculated classification accuracy using area under the curve analyses, and (3) used hierarchical logistic and linear regressions to evaluate incremental utility of ARX over the MMPI-2-RF’s AXY scale for PCL-5 total and cluster scores.
In general, results support use of ARX (α = .86) in predicting PTSD symptoms. Correlations with trauma-related symptoms and experiences were small to moderate (r =.26 to .57), lower than those observed in the technical manual. Similarly, item-level relationships to PCL-5 clusters were negligible to moderate (r = .03 to .47). At T65 (20.7% of sample), ARX had moderate sensitivity and high specificity (.37 and .94) using standard PCL-5 screening scores. When scoring the PCL-5 for symptom criteria qualification, sensitivity at T65 was substantially higher (.75) while specificity was slightly lower (.88). Both had large Area Under the Curve effects (AUC = .82 and .81, respectively). ARX was incremental in predicting PCL-5 total score, as well as each symptom cluster (R2Δ = .02 to .05). In conclusion, researchers and clinicians should feel confident using ARX as a screening measure of PTSD and this study suggests that changes made to AXY/ARX in the MMPI-3 were meaningful and incremental. Replication is needed with PTSD diagnoses verified by clinical interview, and further work is warranted on relationships between ARX and cluster symptoms clusters (e.g., is it predictive of specific or global PTSD symptomology).
Megan A. Keen | Texas Tech University
Bryce A. Robinson | Texas Tech University
Nicole Morris | Texas Tech University
Paul B. Ingram, PhD | Texas Tech University