A Study of the SIRS With Severely Traumatized Patients

Abstract

Clinical research has revealed that traumatized patients often elevate feigning indicators on psychological measures, which raises the possibility that traumatization and concomitant dissociation may lead to misclassifications of malingering. Within the domain of feigned mental disorders, the Structured Interview of Reported Symptoms (SIRS; Rogers, Bagby, & Dickens, 1992) is a well established measure with excellent reliability and validity across clinical and forensic settings. Although recent studies have demonstrated its effectiveness with outpatient posttraumatic stress disorder (PTSD) samples, the SIRS usefulness with severely traumatized patients remains to be investigated. In this study, we recruited traumatized patients for a within-subjects simulation design in which we asked feigners to convincingly portray themselves as examinees claiming total disabilities. When compared to standard instructions, feigned presentations produced substantial effect sizes. Although the standard SIRS classifications produced moderately high sensitivities (M = .82), the false-positive rates were problematic. To minimize false-positives, we constructed a Trauma Index (TI) from 3 primary SIRS scales that appeared unaffected by severe trauma. Implementation of the TI substantially reduced false-positive rates (M = .09).

Acknowledgments

We gratefully acknowledge the support of the Research Opportunities Program at the University of North Texas. As a disclosure of financial interests, Dr. Rogers receives royalties for the SIRS and the SIRS professional manual.

Notes

¹According to Rogers (2008), Cohen's original benchmarks for interpreting effect sizes do not provide sufficient differentiation between criterion groups for the accurate assessment of response style. Instead, the following guidelines are recommended for d: = .75 for moderate, = 1.25 for large, and = 1.50 for very large effect sizes.

²We classified all simulators who acknowledged even minimal effort as “successful”; we wanted to avoid penalizing participants whose ability to simulate may have been negatively affected by their impairment.

^a Based on the 37 patients who received two standard administrations.

^b Due to the scoring method of counting items on SEL and SEV, measures of internal consistency are inappropriate.

^c Two of the four cells were empty, making kappa and Yule's Q impossible to calculate.

∗p ≤ .005,

∗∗ p ≤ .001.

∗p < .05.

∗∗ < .01.

∗∗∗ .001.

∗∗ p < .01.

∗∗∗ p < .001.

^a This compared the honest and feigning conditions for the same 38 participants.

^b Because the total score is only intended to be used in marginal cases, it was only applied when one or two scales were in the probable feigning range.

^c This compared the 37 honest participants in the first administration of the standard-standard condition to the 38 feigners.

^d This compared the 37 honest participants in the second administration of the standard-standard condition to the 38 feigners.

³For RO, a score of “2” represents a sudden worsening of the reported behavior.

⁴Manipulation checks have typically been used with experimental but not control conditions. Although this practice has usually been followed (Rogers, 1997) and may not be generally problematic, we recommend that all participants are given manipulation checks irrespective of their assigned conditions.