Effects of coaching on detection of malingering on the California Verbal Learning Test

The diagnostic accuracy of California Verbal Learning Test (CVLT) indices to detect malingered head injury was examined using a simulation paradigm that included naive malingerers, malingerers provided with information about head injury, and normal controls (N = 90). The application of diagnostic cutoff scores for Recognition Discriminability and Recognition Hits derived from populations of individuals with bona fide head injury and individuals suspected of malingering (Millis et al., 1995) proved highly sensitive and modestly specific in detecting feigned head injury among both simulation groups. Results of analyses of variance and logistic regression support previous findings that malingerers overestimate memory impairment associated with mild head injury; however, they indicate that exposure to a simple instructional set may render insensitive many indices of malingering. In contrast, indices based on more subtle principles of learning theory hold promise in the detection of malingering, even in the presence of an instructional set.