Abstract
Objective
Performance validity tests (PVTs) are widely used in neuropsychological evaluations but are often lengthy and time-consuming. Therefore, development of more efficient PVTs is critical. The current study evaluated the utility of the first 10 items of the Warrington Recognition Memory Test (RMT) as an indicator of performance validity.
Methods
This study utilized archival data from 134 veterans presenting for a clinical neuropsychological evaluation at a southeastern Veterans Affairs hospital. Patients were deemed “credible” if they failed less than two unrelated PVT measures, had a FSIQ > 70, and did not carry a diagnosis of dementia. This resulted in 103 patients in the credible group [Mage = 49.76(11.67); Meducation = 14.09(2.88); 77.67% Male]. The 31 patients in the non-credible group met Slick et al. (1999) criteria[Mage = 45.13(11.18); Meducation = 12.77(2.53); 87.10% Male]. Frequency counts were run to determine sensitivity and specificity values at established cut-offs and to create new cut-offs for the first 10 items.
Results
Using established cut-off scores, the RMT total score had adequate specificity (91%) with a corresponding sensitivity of 84%, and the RMT time score had adequate specificity (93%) with a corresponding sensitivity of 34%. When examining the first 10 items on the RMT, a score of < 8 resulted in 96% specificity and 61% sensitivity.
Conclusions
The established cutoffs for the RMT performed well in this patient sample. Additionally, using a cutoff score of ≤ 8 on the first 10 items of the RMT may be a useful indicator of probable failure on the RMT, significantly reducing test administration time.