Validity of the M Test: Simulation-Design and Natural-Group Approaches

Abstract

The utility of the M Test (Beaber, Marston, Michelli, & Mills, 1985) as a screening measure for malingering was assessed using simulation-design (subjects asked to fake mental illness) and natural-group (forensic assessment cases suspected of malingering) approaches. A total of 318 subjects from community, clinical, and correctional settings completed the M Test under instructions to respond honestly or to simulate a major mental disorder. A factor analysis of the M Test items supported the original rationally based scale assignment, and the scales were found to have adequate internal reliability. To examine the predictive validity of the M Test, results revealed that sensitivity using the cutting scores suggested by Beaber et al. (1985) was much higher for simulating subjects (79.8%) than for the suspected malingerers (40.0%). The findings emphasize the danger of exclusive reliance on simulation studies for validating measures of malingering. It was concluded that, in its present form, the M Test does not constitute a good screening measure for assessing malingering.