A re-evaluation of an automated tailored test of concept learning with elderly psychiatric patients

Aspects of a revised version of the Picture Matching Test were evaluated in a sample of 33 elderly psychiatric patients. The new version of the test shared many defects of the older form, e.g. an uneven increase in levels of difficulty and a low ceiling. The derived speed of response measure was more closely related to the way in which the system responded to a subject's errors than to any of the other more direct measures of a subject's performance. Furthermore, there was no evidence to suggest that the method of controlling for practice effects, as suggested by the test's distributors, would be effective.