A modified way of administering the process dissociation procedure to the false fame paradigm is described. Multidimensional signal detection theory (SDT) is used to correct for recollection as well as familiarity false alarms, and two experiments are reported that compare this method of false alarm correction with the hybrid procedure preferred by Jacoby et al. (1993). In experiment 1, it is shown that recollection and familiarity are lost at the same rate in normal subjects over a delay of 1 d when an SDT analysis is used. Analysis with the hybrid procedure fails to find any forgetting over the 1-d delay. In experiment 2, amnesics are shown to have preserved familiarity in the face of impaired recollection for names when the results are analyzed by either method. An additional analysis showed that the amnesics' familiarity was normal even for relatively novel surnames. The SDT analysis also revealed that the amnesics, relative to controls, showed a conservative recollection and a liberal familiarity response bias. The results indicate that it is important to correct for recollection as well as familiarity false alarms. (JINS, 1995, I, 469–482.)
