Psychology and Aging Copyright 1996 by the American Psychological Association, Inc. 1996. Vol. 1 I, No. 2,258-271 0882-7974/96/$3.00 Fact Retrieval in Younger and Older Adults: The Role of Mental Models Gabriel A. Radvansky Rose T. Zacks University of Notre Dame Michigan State University Lynn Hasher Duke University Using a fan effect paradigm, three experiments tested whether younger and older adults differ in the retrieval of integrated and nonintegrated facts, where integration refers to the development of a mental model. Earlier work by G. A. Radvansky and R. T. Zacks ( 1991 ) had found that as long as facts can be integrated into a single mental model, young adults show no increase in retrieval time or error rates as the size of the subset of related facts increases (i.e., no fan effect). The present studies show a similar pattern for older adults. By contrast, and in confirmation of our previous findings on age differences and the fan effect (L. D. Gerard, R. T. Zacks, L. Hasher, & G. A. Rad- vansky, 1991 ), older adults show an exaggerated fan effect, at least in their error rates, on subsets of related facts not easily integrated into a single mental model. The purpose of the present experiments is to assess how younger and older adults compare in the construction and use of mental models (e.g., Johnson-Laird, 1983, 1989). A mental model is a representation of a described situation rather than a representation of a text itself or the propositions conveyed by a text. The structure of a mental model corresponds to the func- tional relations among entities as they would exist in the world. As such, a mental model can be thought of as a simulation of events in the world, either real or imaginary. The idea of a mental model can be conveyed by considering a study by Garnham (1982). In that study, people who originally heard the sentence "The hostess bought a mink coat from the furrier" were likely to mistakenly claim that the sentence "The hostess bought a mink coat at the furrier's" was heard before. In contrast, people who originally heard the sentence "The hostess received a telegram from the furrier" were not likely to mistak- enly claim that the sentence "The hostess received a telegram at the furrier's" was heard before. The first two sentences poten- tially describe the same situation, and therefore correspond to a single mental model, whereas the second two sentences do not. There is some evidence that young and older adults create and use mental models similarly (Morrow, Leirer, & Altieri, 1992; Radvansky, Gerard, Zacks, & Hasher, 1990). In a study mod- Gabriel A. Radvansky, Department of Psychology, University of No- tre Dame; Rose T. Zacks, Department of Psychology, Michigan State University; Lynn Hasher, Department of Psychology, Duke University. Portions of the research reported here were presented at the 1990 meeting of the Society for Cognition and Aging in Atlanta, Georgia. This research was supported by National Institute on Aging Grant RO 1 AGO 4306. We would also like to thank Karin Butler, Beth Chittenden, Doug Davidson, Heather Oonk, and Dan Spieler for their assistance in collecting the data. Correspondence concerning this article should be addressed to Ga- briel A. Radvansky, Department of Psychology, University of Notre Dame, Notre Dame, Indiana 46556. Electronic mail may be sent via Internet to gabriel.a.radvansky. 1 @nd.edu. 258 eled after the Garnham (1982) experiment described pre- viously, Radvansky et al. found that young and older adults used mental model representations to identify previously heard statements on a later recognition test. The two groups showed the same pattern of confusion errors on the recognition test, suggesting that younger and older adults create similar types of mental models and then use these representations during recognition. One specific aspect of mental models on which young and older adults seem not to differ is foregrounding. Foregrounding refers to keeping certain story elements, such as the protagonist, more available than other elements (e.g., Garrod & Sanford, 1983; Glenberg, Meyer, & Lindem, 1987). In a study by Mor- row et al. (1992), people read a passage followed by a series of comprehension questions that referred back either to the pro- tagonist or to minor characters. Although young adults outper- formed older adults overall, both groups more accurately iden- tified the foregrounded story protagonists than the minor char- acters. In fact, performance on questions referring to the protagonist was identical in the two age groups. These findings suggest that both age groups form similar mental models. The current experiments expand on this earlier research by compar- ing young and older adults' use of mental models with sets of facts presented outside of a connected discourse. In previous studies with young adults, we (Radvansky, Spieler, & Zacks, 1993; Radvansky & Zacks, 1991 ) have dem- onstrated that the use of mental models to integrate potentially competing items of information results in the elimination of the negative impact of competitors at retrieval. This finding was observed in the context of a fan effect paradigm (Anderson, 1974). The fan effect is an increase in response time (RT) or error rate in recognizing learned facts with an increase in the number of other facts having a concept in common with the probed fact. Our research has shown that the fan effect is elimi- nated when facts having a concept in common can be readily integrated into a single mental model. In contrast, the standard fan effect is obtained when facts having a common concept can-