Associations Between Symptom Validity Test Failure and Scores on the MMPI–2–RF Validity and Substantive Scales

Abstract

This study examined the association between Symptom Validity Test (SVT) failure and the Minnesota Multiphasic Personality Inventory–2 Restructured Form (MMPI–2–RF; Ben-Porath & Tellegen, 2008), in the Forensic Disability Claimant samples described in the MMPI–2–RF Technical Manual (Tellegen & Ben-Porath, 2008a, 2008b). SVTs used included the Word Memory Test (Green, 2003), the Computerized Assessment of Response Bias (Allen, Conder, Green, & Cox, 1997), the Medical Symptom Validity Test (Green, 2004), and the Test of Memory Malingering (Tombaugh, 1996). SVT failure was associated with significant elevations throughout the MMPI–2–RF overreporting validity scales and substantive scales. Pairwise contrasts between groups failing 0 and 3 SVTs revealed predominantly large effect sizes for the overreporting validity scales (d = 0.78–1.11), and many of the substantive scales, including the Cognitive Complaints (COG) scale. Results of this study demonstrate an association between SVT performance and elevated scores on the MMPI–2–RF. These results suggest that exaggeration of cognitive symptoms as demonstrated by SVT failure is also associated with overreported emotional, somatic, and neurocognitive complaints on the MMPI–2–RF.

Acknowledgments

Yossef Ben-Porath is a paid consultant to the MMPI–2–RF publisher, the University of Minnesota Press, and distributor, Pearson. He receives royalties on sales of MMPI–2–RF materials.

Portions of this study were presented at the 44th Annual Symposium on Recent Research with the MMPI–2, MMPI–2–RF, & MMPI–A, May 2009, Minneapolis, MN. The authors thank Anthony Tarescavage for his assistance with the tables.

Notes

To address concerns about potential redundancy among the overreporting predictors we tested a Poisson regression model in which F-r, Fp-r, Fs, and FBS-r were entered into an equation predicting number of SVT tests failed. We used backward elimination to generate an equation in which only statistically significant predictors remained in the model (robust standard errors; p > .05 was used for removal). Overall models were significant for both men, Wald χ²(2) = 100.21, p < .001; pseudo R² = .08; and women, Wald χ²(2) = 53.75, p < .001; pseudo R² = .06. The same predictors emerged as statistically significant in both models—F-r (b = .038, SE = .012, z = 3.16, p = .002 [men]; b = .046, SE = .014, z = 3.24, p = .001 [women]) and FBS-r (b = .095, SE = .017, z = 5.77, p = <.001 [men]; b = .078, SE = .020, z = 3.79, p < .001 [women]).