Structured Interview of Reported Symptoms (SIRS)

Description and Development

The initial development of the SIRS was based on an exhaustive review of potential detection strategies for feigned mental disorders. SIRS scales were developed based on the likely effectiveness of the underlying detection strategy and the adaptability of each strategy to interview-based assessments. Final item selection was based on independent judgments by eight experts in malingering and was subsequently refined to improve scale homogeneity. The SIRS is composed of 173 items that are organized by eight primary and five supplementary scales.

The original validation of the SIRS combined samples from multiple studies to form four groups: 100 inpatients and outpatients, 97 controls from community, correctional, and college settings, 170 simulators including coached and uncoached participants, and 36 likely malingerers from forensic settings. Subsequent validation research has included clinical and correctional samples with an additional 255 participants.

Reliability

Internal consistencies (alpha coefficients) for SIRS primary scales were excellent: They ranged from .77 to .92, with a mean alpha of .86. The reliability of individual scores was examined via standard errors of measurement (SEM). The SEMs were low for both clinical and control samples, indicating high reliability for individual scores. A central issue for the SIRS was its interrater reliability. These estimates were impressive, ranging from .89 to 1.00. The median reliability was .99, which represents almost perfect agreement.

Validity

SIRS validation relied on a combination of simulation designs and known-group comparisons. The simulation design capitalizes on internal validity in its use of analog research with clinical comparison samples. In contrast, known-group comparisons provide unmatched external validity in their use of individuals from actual clinical-forensic settings who were independently evaluated as malingering by established experts. For the assessment of malingering, convergent results from simulation and known-group studies provided the strongest evidence of SIRS validity.

A major focal point of the SIRS is its discriminant validity. The critical issue is whether each of the primary scales systematically differentiates between genuine and feigned psychopathology. Combining across studies, effect sizes can be computed for the critical distinctions (a) simulators versus clinical honest and (b) suspected malingerers versus clinical honest. For simulators, Cohen's d's were very large: They ranged from 1.40 (SU) to 2.31 (RS) with an average d of 1.74. Cohen's d's were also very large for suspected malingerers but showed less variability: 1.20 (IA) to 1.98 (SEL). The average effect size for malingerers was identical to that of simulators (d = 1.74). These combined data provide very strong evidence of discriminant validity.

Convergent validity was evaluated by comparing the SIRS with other measures of feigned mental disorders. The SIRS evidenced robust correlations with MMPI–2 validity scales. For example, the SIRS primary scales are strongly correlated with Scale F (r's from .71 to .80). As also expected, they are negatively correlated with Scale K (Mr =−.35), a measure of defensiveness.

...