WHEN TRUE MEMORIES SUPPRESS FALSE MEMORIES:EFFECTS OF AGEING Elizabeth A. Kensinger and Daniel L. Schacter Harvard University, Cambridge, USA After studying a list of words that are all associated to a nonpresented target word, people often falsely recall or recognise the nonpresented target. Previous studies have shown that such false memories are greatly reduced when study lists are presented and tested several times compared to a single study/test trial. We report that older adults, who are sometimes more susceptible to memory distortions than are young adults, failed to exhibit any reduction in false recall or false recognition after five study/test trials compared to a single trial. By contrast, younger adults showed robust suppression of false memories after five study/test trials compared to a single trial. These results are consistent with the idea that older adults rely on memory for the general features or gist of studied materials, but tend not to encode or to retrieve specific details of individual items. INTRODUCTION Numerous studies of ageing memory have docu- mented that older adults exhibit lower levels of veridical recall and recognition of recently studied information than do younger adults (for reviews, see Craik, Anderson, Kerr, & Li, 1995; Light, 1991). But a growing body of evidence converges on the conclusion that, compared to younger adults, older adults sometimes show equal or greater levels of false recall and false recognition of items not previously studied. Early studies by Smith (1975) and Rankin and Kausler (1979) investigated false recognition using a procedure in which sub- jects make old/new decisions about previously stud- ied words, new words that are related to a previously studied associate (related lures), and new words that are not related to previously studied words (unre- lated lures). Both Smith and Rankin and Kausler reported higher levels of false recognition in older adults than in younger adults, although the overall magnitude of the false recognition effect was small (see also Isingrini, Fontaine, Taconnat, & Duportal, 1995). More recent research has shown that larger false recognition effects can be obtained when subjects study numerous items that are conceptually or per- ceptually similar to a novel test item (Hintzman, 1988;Shiffrin, Huber, & Marinelli, 1995).Particu- larly striking demonstrations of robust false recog- nition and false recall have been reported by Roediger and McDermott (1995). They revived and modified a procedure, described initially by Deese (1959), in which subjects are initially exposed to lists of semantic associates (e.g., candy, sour, sugar, bitter, good, taste, and so forth) that all COGNITIVE NEUROPSYCHOLOGY, 1999, 16 (3/4/5), 399–415 Ó 1999 Psychology Press Ltd 399 Requests for reprints should be addressed to Daniel L. Schacter, Department of Psychology, Harvard University, 33 Kirkland St., Cambridge, MA 02138, USA. Supported by grant AGO8441 from the National Institute on Aging and a grant from the Human Frontiers Science Program. These experiments were conducted by the first author under the supervision of the second author as part of a senior honours thesis at Harvard College. We thank Carrie Racine for assistance with the research, and thank Marc Hauser, Larry Jacoby, Wilma Koutstaal, Michelle Leichtman, Kathleen McDermott, and Anthony Wagner for helpful comments and discussion.