Attempting to avoid illusory memories: robust false recognition of associates persists under conditions of explicit warnings and immediate testing

Roediger and McDermott (1995) showed that presentation of associated words can induce false recognition of a related, but nonpresented, associate. In three experiments, we placed this related associate in half of the study lists (but not in the other half) in an attempt to determine whether (and under what conditions) subjects could distinguish between cases in which the critical associate was and was not presented. Results suggest that subjects are quite poor at performing this straightforward task, even when explicitly informed of the false recognition phenomenon, instructed to pay careful attention to whether or not the critical linking associate was presented in the list, and given a 1-item recognition test immediately following each 15-word list. Although subjects were not able to perform accurately under these conditions, the warning instruction did attenuate the false recognition effect (relative to an uninformed condition). This illusion of memory appears to be remarkably robust and little affected by the instructional manipulations. © 1998 Academic Press Roediger and McDermott (1995) reported ex-periments showing that presentation of words (e.g., door, glass, pane, shade, ledge) that are associates of a nonpresented word (e.g., win