Predicted performance experiments attempt to quantify an impaired individual's awareness of his or her deficit. These experiments measure perceived ability by the individual's prediction of his or her performance on a specific cognitive task and actual ability by his or her subsequent performance on that task. To date, the most comprehensive predicted performance experiment is the one proposed and implemented by McGlynn and Kaszniak (1991b). This experiment is potentially capable of removing a number of influences that may be confounded with deficit unawareness; however, it is not obvious what method of quantitative analysis best exploits this capability. In the present report, several possibilities are discussed. The limitations of McGlynn and Kaszniak's method are identified, and a more satisfying measure of deficit unawareness is proposed.