Reexamining performance validity cutoffs within the Complex Ideational Material and the Boston Naming Test-Short Form using an experimental malingering paradigm.

OBJECTIVE: This study was designed to cross-validate previously published performance validity cutoffs embedded within the Complex Ideational Material (CIM) and the Boston Naming Test-Short Form (BNT-15).

METHOD: Seventy healthy undergraduate students were randomly assigned to either a control condition (n = 40) and instructed to perform to the best of their ability or an experimental malingering (n = 30) condition and instructed to feign cognitive impairment while avoiding detection. All participants were administered the same battery of neuropsychological tests.

RESULTS: Previously published validity cutoffs within the CIM (raw score ≤9 or T-score ≤29) and BNT-15 (≤12) produced good classification accuracy using both experimental malingering and psychometrically defined invalid responding as criterion variable. However, a BNT-15 completion time ≥85 s produced a better signal detection profile than BNT-15 accuracy scores.

CONCLUSIONS: Results support the clinical utility of existing cutoffs. Given the relatively high base rate of failure even in the control group (5-15%), and the perfect specificity of CIM ≤9 and BNT-15 ≤ 11 to noncredible responding, relabeling this range of performance as "Abnormal" instead of "Impaired" would better capture the uncertainty in its clinical interpretation.