* This research was supported by grant number AG05552 awarded to Thomas M. Hess by the National Institute on Aging. The authors would like to thank Katherine Follett, Rachel Garfield, and Stephen Woodburn for their assistance in this project. Address correspondence to: Thomas M. Hess, Department of Psychology, Box 7801, North Carolina State Uni- versity, Raleigh, NC 27695-7801, USA. E-mail: hess@poe.coe.ncsu.edu. Accepted for publication: December 16, 1997. Aging, Neuropsychology, and Cognition 1382-5585/98/0501-027$12.00 1998, Vol. 5, No. 1, pp. 27-42 © Swets & Zeitlinger Category-Based versus Attribute-Based Processing in Different-Aged Adults* Thomas M. Hess and Cheryl A. Bolstad North Carolina State University ABSTRACT Adaptive social functioning requires flexibility in processing operations in response to the characteristics of specific situations. Three experiments were conducted in which the ability of different-aged adults to appropriately use category-based versus attribute-based information in making affective judgments was examined. Contrary to expectations, we found that performance was minimally related to age or to working memory efficiency. Adults of all ages made affective judgments based on category information when such information resulted in appropriate inferences, whereas attribute-based information was used when cate- gory-based inferences were inaccurate. We interpret these results to be consistent with the hypothesis that the processing of evaluative information occurs at a preconscious level with little drain on processing resources. When we encounter information about another person, our natural tendency is to create some sort of representation or impression about the person. How such impressions are formed can vary, based upon the type of information that is processed about the person. Generally, it is thought that impression formation proceeds through a sequence of stages, beginning with categorization of the target in terms of some eas- ily accessible categorical information (Brewer, 1988; Fiske & Neuberg, 1990; Gilbert & Hixon, 1991; Hilton & Darley, 1991). Thus, for exam- ple, when we encounter an older adult, our ste- reotype about old people might be activated, and we can then use this stereotypical information to make both affective and behavioral inferences about the person. If this top-down approach is successful (i.e., our category-based inferences serve our interactional goals), little additional processing is performed. In contrast, if the cate- gorization is unsuccessful, a more bottom-up approach may be adopted, in which the per- ceiver attempts to integrate specific pieces of information about the person in order to form either a person-specific impression or to identify a category that can be appropriately used to characterize the person. Thus, for example, when considering an applicant to graduate school, we might try to piece together informa- tion about the person’s academic background, standardized-test performance, interests, under- graduate institution, and work experience in or- der to evaluate his or her potential. Such top-down and bottom-up processing have been referred to as category-based and at- tribute-based (or piecemeal) processing, respec- tively, in the social cognition literature (e.g., Asch, 1946; Fiske & Neuberg, 1990), where re- searchers have been interested in examining the conditions under which each of these is used by people. Consistent with the notion of humans as ‘‘motivated tacticians,’’ Fiske and Neuberg