Detecting Response Bias with Performance Patterns on an Expanded Version of the Controlled Oral Word Association Test

Abstract

The present study investigated whether speeded word generation performance patterns seen in healthy subjects are also produced in genuine and feigned traumatic brain injury (TBI). An expanded version of the Controlled Oral Word Association Test was administered to healthy controls, TBI patients, simulated malingerers, and probable clinical malingerers. Four performance patterns were operationalized. Three of these patterns were replicated in the healthy control sample and found to be unaltered by genuine TBI. They were then combined into a logistic regression model that discriminated well between examinees who put forth adequate effort and those who evidenced response bias.

Keywords:

• Effort
• Malingering
• Phonemic fluency
• Response bias
• Traumatic brain injury
• Verbal fluency

ACKNOWLEDGMENTS

The authors would like to thank University of Windsor students Dawne Martens and Stephanie Solcz, as well as the neuropsychology trainees and staff of the Rehabilitation Institute of Michigan, for their assistance with data collection.

Notes

†Total score = sum of correct responses across all trials.

¹Clinicians may also want to correct for the pretest probability of response bias (i.e., base rate) that corresponds to their work setting. This can be accomplished by calculating likelihood ratios (see Millis & Volinsky, 2001).