Studied 338 patients, including 74 normal Ss, 83 psychiatric patients, and 181 neurological patients. Odd–even splits were used to determine the split-half reliabilities, which ranged from .89 (Memory) to .95 (Reading). Item–scale consistency was determined by correlating each item with each scale score and assigning the item to the scale with which it correlated highest. The scale that the item was on according to the test and the scale to which it was empirically assigned agreed for 250 of the 269 items. (5 ref)