Attributions of Fluency in Fame Judgments by Younger and Older Adults

Judgments about stimulus characteristics are affected by enhanced processing fluency that results from an earlier presentation of the stimulus. By monitoring for an episodic source of processing fluency, younger adults can more easily avoid this influence than can older adults. In Experiment 1, older adults discounted the effects of fluency when task demands encouraged the use of analytic judgments based on general knowledge, rather than an appeal to episodic source monitoring. Younger subjects were not reliably affected by these same task demands and their judgments continued to be affected by processing fluency. In Experiment 2, introduction of more stringent demands led younger adults also to discount the effects of fluency. We conclude that the influence of processing fluency on younger and older adults varies, depending on whether memory for source or general knowledge is put forward in place of fluency as a basis for judgments. The fluency with which a stimulus is perceived and identified can be strongly enhanced by a single prior presentation of that stimulus (Jacoby & Dallas, 1981). Moreover, the fluency generated by the previous exposure of an item can influence judgments that an observer makes regarding its properties. For example, it has been argued that recognition memory is based in part on an attribution regarding feelings of familiarity or fluency experienced when encoding a stimulus on a recognition test (Jacoby, Kelley, & Dywan, 1989). In this view, fluency is attributed to prior experience with the stimulus, prompting the conclusion that it has been encountered before. An important difficulty with using fluency as a basis for judging an item, a process that has been called the fluency heuristic (Jacoby & Brooks, 1984), is that many factors contribute to fluent processing, so an observer may be unable to make an accurate ascription regarding the source of fluency. In particular, perceptual fluency caused by a recent experience with a stimulus may not be ascribed to that occurrence. Instead, it might mistakenly be attributed to some characteristic that the stimulus does not actually possess, Canadian Journal of Experimental Psychology, 1995, 49:3, 287-310 288 Masson, Carroll, and Micco such as fame in the case of a name (Jacoby, Woloshyn, & Kelley, 1989), or to presentation conditions that do not actually prevail, such as longer viewing time or reduced background noise (Jacoby, Allan, Collins, & Larwill, 1988; Witherspoon & Allan, 1985). Conversely, fluency produced by stimulus presentation conditions may be erroneously attributed to the prior presentation of the stimulus, leading to false alarms on a recognition test (Jacoby & Whitehouse, 1989; Johnston, Hawley, & Elliott, 1991). The problem of determining the appropriate source of processing fluency is analogous to the problem of source amnesia, in which information is remembered but its episodic source cannot be recalled or the information is attributed to an inappropriate source (Schacter, Harbluk, & McLachlan, 1984). Difficulty with memory for source appears to be especially pronounced among older adults (Cohen & Faulkner, 1989; Hashtroudi, Johnson, & Chrosniak, 1989). Similarly, the elderly appear to be less able than younger adults to monitor the source of processing fluency when making stimulus judgments. Dywan and Jacoby (1990) showed that among older adults, classification of a nonfamous name as famous was more likely if it had been read earlier as part of a list of nonfamous names. Because subjects were informed before reading the original list that it contained only nonfamous names, successful monitoring of the source of fluency during the fame judgment task would have prevented incorrect classification of the previously read names. In fact, younger subjects showed just this effect and were more likely to judge new, rather than old, nonfamous names as famous. Older subjects, however, produced the opposite result, suggesting that they were less successful in establishing the source of fluency as prior occurrence on a list of nonfamous names. In an experiment that used a very similar procedure, Squire and McKee (1993) found that amnesic subjects were unable to recollect the prior exposure of nonfamous names, and consequently were quite susceptible to the influence of past exposure in making fame judgments. In addition, Bartlett, Strater, and Fulton (1991) have shown that older subjects are more influenced than younger subjects by prior presentation of a face when making false alarms on fame and recency judgment tasks. Susceptibility to the inappropriate influence of prior exposure to an item is not restricted to older adults and amnesic subjects. When operating under divided attention, normal younger adults have difficulty using memory for source as a basis for discounting the influence of prior occurrence (Jacoby, Woloshyn, & Kelley, 1989). Divided attention effects obtained with younger adults might be taken as a model for the process of aging, by which it is assumed that older adults are less able to meet the requirements of dual task situations because of limited processing resources (e.g., Craik & Byrd, 1982). This approach leads to the suggestion that source monitoring difficulty observed among older adults may reflect a general deficit in handling multiple task demands. Alternatively, the source monitoring problems among older Attribution of Fluency 289 adults reported by Dywan and Jacoby (1990) and by Bartlett et al. (1991) may be specific to the episodic memory requirement involved in monitoring for source, rather than to problems with dual task performance. The possibility that a memory deficit is responsible for source monitoring problems receives support from the fact that age-related deficits in dual task performance is not a consistent finding (Plude & Hoyer, 1986; Salthouse, Fristoe, Lineweaver, & Coon, 1995; Wickens, Braune, & Stokes, 1987). In the experiments described here, we explored the hypothesis that older adults may be just as effective as younger adults at discounting the effects of processing fluency when doing so does not rely on episodic memory. We arranged task demands so that the effects of processing fluency could be avoided by relying on general knowledge, a skill that is maintained or even enhanced among older adults (Burke & Harrold, 1988; Hultsch, Hertzog, & Dixon, 1990; Perlmutter, 1986), and at worst shows only a small age-related decline (Salthouse, 1988). If avoidance of the effects of fluency is not compromised by general age-related deficits in processing capacity, older adults should show evidence of discounting fluency when task demands invoke a knowledge-based alternative heuristic. Experiment 1 We examined the influence of task demands on the use of the fluency heuristic using a version of the fame judgment task developed by Jacoby and his colleagues (Dywan & Jacoby, 1990; Jacoby, Woloshyn, & Kelley, 1989). Subjects initially read a list of famous and nonfamous names, and later were asked to classify a set of names as famous or nonfamous. Some of the test names had been read in the first part of the experiment, and consequently were expected to be more fluently identified by all the subjects. We attempted to verify this expectation by testing a subset of the names in a visual identification task. Subsequent use of the fluency heuristic in the classification task was expected to be revealed by a higher probability of classifying the previously read names as famous. To evaluate the claim that use of the fluency heuristic would be influenced by prevailing task demands, half of the subjects were asked only to make a decision about the fame of each name. The remaining subjects were given the same judgment task, but also were instructed to explain why a name was famous when they classified it as such. This request for an analytic, knowledge-based assessment of evidence for a claim of fame was expected to provide subjects with an alternative to the fluency heuristic; the ability to generate some plausible reason for fame could be used as grounds for classifying a name as famous. Given that this alternative basis for judging fame depends on general knowledge, which older adults often can access as effectively as younger subjects, it was expected that both age groups would be able to reduce the extent to which they relied on the fluency heuristic to 290 Masson, Carroll, and Micco make fame judgments. Subjects were not, however, required to explain why a name was famous; they had the option of classifying a name as famous but reporting that they did not know the reason for fame. This approach made it possible to test for a spontaneous rather than a forced shift in the basis for making fame judgments. The procedure used here differed in an important respect from that of Dywan and Jacoby (1990) and Bartlett et al. (1991). In those experiments, episodic memory for prior occurrence was placed in opposition to a positive decision about fame by exposing subjects to a study list comprised of only nonfamous names. Younger subjects were better able to monitor for names or faces that had been presented earlier, and this information helped inoculate them against the influence of fluency in making their decisions regarding fame. In the present experiment, rather than initially exposing subjects to a pure list of nonfamous names, we used a mixed list of famous and nonfamous names. This procedure neutralized the role of explicit recognition by making it uninformative with respect to the fame judgment a previously presented name could be famous or nonfamous. The instruction to provide a reason for fame, however, served to place the lack of general knowledge in opposition to a positive fame judgment, and thereby created an alternative to the fluency heuristic. The critical question was whether subjects, particularly older adults, would make use of this alternative basis for fame judgments.