Minding Your “ps and qs” When Using Forced-Choice Recognition Tests

Denney (1996), and Frederick, Carter, and Powel (1995) described a forced-choice recognition test (FCRT) to evaluate suspicious complaints of remote memory loss. Although the two-alternative forced-choice trials in symptom validity testing have equal prior probabilities of correct answers (p) and incorrect answers (q), ps and qs might vary from 0.5 on individual trials of FCRTs. FCRTs nonetheless remain conservative procedures for evaluating suspicious memory loss, as long as the overall p approximates 0.5. In computer simulations, distributions in which p q 0.5 resulted in more conservative decision making than distributions in which p = q = 0.5. The authors demonstrate the ease in constructing FCRTs with overall p = 0.5 and discuss the proper application of FCRT findings in a clinical evaluation.