The Effects of Prior Use on Preference

Abstract

The Effects of Prior Use on Preference Samuel B. Day ([email protected]) Department of Psychology, Northwestern University 2029 N. Sheridan Drive Evanston, IL 60208-2710 USA Daniel Bartels ([email protected]) Department of Psychology, Northwestern University 2029 N. Sheridan Drive Evanston, IL 60208-2710 USA Holyoak, 1983; Loewenstein, Thompson & Gentner, 1999). This suggests that the act of comparison may serve to highlight meaningful structural commonalities between the examples, making those structures more available for application to new situations. Similar results have been found in developmental research, where comparison has been shown to facilitate the acquisition and application of difficult abstract concepts (e.g., Gentner & Medina, 1998). These findings can be interpreted within the framework of structure mapping theory (Gentner, 1983), which proposes that comparisons involve the alignment of representational structures (e.g., Markman & Gentner, 1993). These structural alignments lead to a special role for “alignable differences,” or differences that are related to structural commonalities. For instance, the fact that cars have four wheels while motorcycles have two is an alignable difference, since it involves different values on the shared dimension of “number of wheels.” On the other hand, the fact that motorcycles have a kickstand is a unique feature for which a car has no corresponding attribute, making it a “nonalignable difference.” Prior research has repeatedly shown a privileged role for alignable differences, in areas such as perceived similarity (Markman & Gentner, 1996), recall (Markman & Gentner, 1997), and preference (e.g., Roehm & Sternthal, 2001). With respect to the current studies, these findings suggest that comparison of alternatives that vary along particular dimensions should serve to highlight those dimensions. Our primary question is whether this sort of differential emphasis could lead to generally increased salience, causing increased dimension weighting on subsequent choices. Some work exists examining the relationship between structural alignment and decision processes. In the domain of consumer research, Zhang and Markman (1998) showed that after initially learning facts about a particular brand of product, future alternative products were evaluated largely in terms of their alignable differences with this “first entrant.” Values on dimensions that were shared with this initial product were better remembered, and played a larger role in preference judgments. In another set of studies, Markman and Medin (1995) found that justifications for preferences were more likely to include alignable than nonalignable differences. (For a review of interesting parallels between similarity and choice processes, see Medin, Goldstone & Markman, 1995). Another way in which comparison and decisions seem to be linked is in the shared influence of the context created by Abstract There is little research examining the effects of decision processes on the individual, and the work that does exist has emphasized very high-level processes, associated with complex decisions made over extended periods of time, with substantial levels of engagement. We present two studies demonstrating very basic effects, in which use of a dimension in a decision leads to its increased salience, and subsequent greater weighting in later decisions. Effects were based on both recency of use and total amount of prior use. Introduction A large body of research has focused on the factors that influence people’s decisions. (For a recent review, see Markman & Medin, 2002). However, relatively little attention has been given to the effects that the act of making a decision may have upon the individual. Of course, there are the obvious tangible effects of the actions people undertake as a result of their choices. However, it is also reasonable to suggest that there could be an impact on the mental representations that are used in decision processes. In the present studies, we are investigating very specific, low-level effects, determining whether the particular dimensions that are used in making a decision will become generally more salient, and thus play a larger role in subsequent choices. We presented participants with alternative models of portable video players, in which certain dimensions (such as hard drive capacity or battery life) were diagnostic and thus useful for determining relative preference. We then examined the extent to which dimensions that were used in a decision would become more important on later choices between different models. Comparison and Choice One reason to expect choice processes to affect mental representations is the fact that decisions generally require comparison between alternatives, and there is evidence that comparison itself may lead to representational change. Some of the most compelling evidence in this regard comes from demonstrations that comparison may facilitate knowledge transfer. For example, people were found to be more successful at solving difficult problems or applying optimal strategies after having previously made comparisons between analogous examples, and this advantage was far superior to that of situations in which the prior analogues were simply encoded separately (Gick &