The MMPI-2 Fake Bad Scale: Concordance and Specificity of True and Estimated Scores

Abstract

A number of recent studies have supported the use of the MMPI-2 Fake Bad Scale (FBS) as a measure of negative response bias, the scale at times demonstrating greater sensitivity to negative response bias than other MMPI-2 validity scales. However, clinicians may not always have access to True FBS (T-FBS) scores, such as when True-False answer sheets are unavailable or published research studies do not report FBS raw scores. Under these conditions, Larrabee (2003a) suggests a linear regression formula that provides estimated FBS (E-FBS) scores derived from weighted validity and clinical T-Scores. The present study intended to validate this regression formula of MMPI-2 E-FBS scores and demonstrate its specificity in a sample of non-litigating, clinically referred, medically intractable epilepsy patients. We predicted that the E-FBS scores would correlate highly (>.70) with the T-FBS scores, that the E-FBS would show comparable correlations with MMPI-2 validity and clinical scales relative to the T-FBS, and that the E-FBS would show an adequate ability to match T-FBS scores using a variety of previously suggested T-FBS raw score cutoffs. Overall, E-FBS scores correlated very highly with T-FBS scores (r = .78, p < .0001), though correlations were especially high for women (r = .85, p < .0001) compared to men (r = .62, p < .001). Thirty-one of 32 (96.9%) comparisons made between E-FBS/T-FBS correlates with other MMPI-2 scales were nonsignificant. When matching to T-FBS “high” and “low” scores, the E-FBS scores demonstrated the highest hit rate (92.5%) through use of Lees-Haley's (1992) revised cutoffs for men and women. These same cutoffs resulted in excellent overall specificity for both the T-FBS scores (92.5%) and E-FBS scores (90.6%). The authors conclude that the E-FBS represents an adequate estimate of T-FBS scores in the current epilepsy sample. Use of E-FBS scores may be especially useful when clinicians conduct the MMPI-2 short form, which does not include all of the 43 FBS items but does include enough items to compute each of the validity and clinical T-Scores. Future studies should examine E-FBS sensitivity in compensation-seekers with incomplete effort.

Notes

^an = 24.

^bn = 29.

^an = 24.

^bn = 29.

*p < .01.

**p < .001.

Note. ^cThis calculation tested the hypothesis that T-FBS and E-FBS correlates (with MMPI-2 validity and clinical scales by gender) are equal. The z score transformation test (Fisher's r-to-z transformation) makes use of the sample size employed to obtain each coefficient. Z-scores were compared, in a 2-tailed fashion to the unit normal distribution (used formula 2.8.5 from Cohen and Cohen, 1983, p. 54). By convention, values greater than 1.96 are considered significant if a 2-tailed test is performed.

^aN = 53.

^bn = 24.

^cn = 29.