Special Session Summary Constructing Preferences: the Influence of When and Where

Elizabeth Gelfand Miller, University of Pennsylvania
[ to cite ]:
Elizabeth Gelfand Miller (2003) ,"Special Session Summary Constructing Preferences: the Influence of When and Where", in NA - Advances in Consumer Research Volume 30, eds. Punam Anand Keller and Dennis W. Rook, Valdosta, GA : Association for Consumer Research, Pages: 11-13.

Advances in Consumer Research Volume 30, 2003     Pages 11-13



Elizabeth Gelfand Miller, University of Pennsylvania

The objective of this session was to explore how tastes and preferences develop over time. In particular, the papers in this session examined the effect of when and how items are experienced or presented on the development of preferences. The first paper focused largely on the role of initial experiences in the development of tastes, but also considered the role of explicit knowledge and expertise on the development and expression of preferences. The second paper also examined the influence of expertise and order of presentation, but focused more on how items evaluated in isolation might experience a brand positivity bias. The third paper, while also considering the influence of single evaluations vs. more simultaneous evaluations on preference, focused more on how a specific characteristic of displayBthe naming of colors or flavorsBaffects preferences and choice.

The first paper (Hoeffler and West) examined how experience impacts changes in taste over time. In particular, the paper explored how initial experiences and explicit knowledge influence preferences and taste formation. Hoeffler and West demonstrated that tastes change with experience and that initial choices in a product category are biased by people’s initial exposure. In addition, they show that such choices can lead to further biased sampling within the product category, impacting people’s "real" preferences.

The second paper (Posavac et al.) examined the influence of singular evaluation (the evaluation or appraisal of singular products or brands as opposed to the consideration of multiple alternatives at one time) on brand evaluations and behavioral intentions. Posavac et al. showed that singular evaluation is often characterized by a brand positivity effectBthe tendency for favorably regarded products or brands to be evaluated more positively than warranted when judged in isolationBand that this bias can influence choice intentions. In addition, their findings suggest that selective processing contributes to the brand positivity effect, while the consideration of additional alternatives and increased expertise in the product category reduce this effect.

The third paper (Miller and Kahn) also explored the role of presentation in influencing preferences by focusing on how the naming of colors and flavors influences preferences for items and consumers’ likelihood of purchasing them. Miller and Kahn’s findings demonstrate that ambiguously-named items (e.g., Florida red or Frost) are often more likely to be purchased than typically named items (e.g., dark red, grape) and that such differences arise due to consumers’ increased likelihood of making positive attributions about the ambiguously-named products. Miller and Kahn develop theory for this process using Grice’s "conversational norms" ideas.



Steve Hoeffler, University of North Carolina

Pat West, Ohio State University

In this paper we investigate the impact of experience on changes in taste over time. We explore the role of the initial experience in taste development. We will demonstrate that initial experiences play a key role in the development of tastes and that initial experiences can anchor taste in two distinct ways. First of all, the components of the actual experience itself impact taste formation. In addition, we will show that the explicit knowledge about the initial experience has a separate influence on tastes. Lastly, we will show that initial experiential anchors may lead consumers astray from their optimal preferences.

The goal of this project is to investigate the impact of the specific starting point on the development of tastes. As it is an early account of these types of leanings, we will raise more questions than we answer. For example, we don’t believe that the type of experience and the explicit knowledge about the experience are the only factors that impact taste. In fact, there are most likely many actors that influence the type and intensity of taste change over time; such as developmental changes, aging, discrimination, learning, adaptation, consumption vocabulary (West, Brown, and Hoch 1996), commitment, and social changes.

Effects of Exposure on Liking (Acquired Taste for Chili Peppers)

Rozin and Schiller (1980) examine the development of taste for chili peppers. Their goal was to examine the development of affect in the context of acquiring taste for chili peppers and specifically the irritation associated with chili peppers dubbed the "chili burn." Rozin and Schiller studied how tastes were acquired for chili peppers in a rural village in the highlands of Mexico versus North American subjects. In general, they found that in the Mexico sample, there was a gradual increase in preferences for chili peppers over a period of two to eight years beginning at around the age of three to five.

Chili’s were not offered to infants. In fact, in some chili eating cultures, chili is placed on the mother’s breast to facilitate weaning (Jelliffe 1962; cited from Rozin and Schiller 1980). Starting at around the age of three, mothers would expose their children to chili peppers. In particular; "In the case of salsa, a small amount was put on a tortilla for children in the 3- 5- (but occasionally 2-) year range. If the child rejected it, the mother would prepare another tortilla, with less or no salsa (Rozin and Schiller 1980, italics in original)." In general, at about the age of 5, children exposed themselves to chili peppers that were available (in the form of salsa). While explicit rewards for eating chili’s were not offered, the more subtle rewards associated with copying behavior of older siblings and parents were in effect.

Rozin and Schiller (1980) examine (and mostly discounted) a host of reasons for the acquired taste of chili peppers (e.g., receptor desensitization, associative learning, opponent process, and benign masochism). Finally, they mention both mere exposure and social factors as the most likely mechanisms leading to an acquired taste for chili peppers. Note that the social factor was not in the form of social pressure and was not thought to be the most important factor. Instead, they claim that exposure to gradually increasing amounts seemed to be the major factor leading to the change in taste.

Role of Initial Experience in Acquiring Taste

One account of how preferences are developed in a novel environment is found in the recent award winning work of Heilman, Bowman, and Wright (2000). The authors look at the evolution of preferences for parents who are new to the diaper market. In this work, initial choices in an environment are thought to be driven by two competing objectives; the desire to obtain information about the product category space, and the desire to avoid risky alternatives. The authors posit the existence of different stages of buying where information is collected and then consolidated into a stable preference for the product that provides the most utility for each consumer.

Hoeffler and Ariely (1999) examined the process of preference consolidation in a novel environment (aversive sounds). They examined the role of experience, effort, and the choice process in preference stabilization. A key factor that led to preference stabilization in the Hoeffler and Ariely studies was exposure to tradeoffs in the environment. In these studies, the authors focused on showing that preferences stabilize after experience in the environment. Thus, they did not examine the impact of acquiring taste. It is interesting that in many of the food environments where people are thought to acquire taste over time, the presence of tradeoffs and the dimensions on which tradeoffs are made are not so obvious. For instance, in the wine category, are White Zinfandel drinkers really giving up the more complex tastes associated with Chardonnay if their preferences had not evolved to a stage where they can appreciate these complex tastes?

One phenomenon that implies a strong influence of an initial experience is called the anchoring and adjustment bias (Tversky and Kahneman 1974). Standard anchoring studies usually focus on judgment (Tversky and Kahneman 1974) and have been shown to hold over a variety of stimulus materials and contexts (for a recent review see Epley and Gilovich 2001). Ariely, Loewenstein, and Prelec (2002) have extended the basic research on anchoring, by showing how the initial starting point of experience has a large role on valuation of alternatives. Ariely, Loewenstein, and Prelec showed how arbitrary initial anchors in novel environments (e.g., aversive sounds) have a large impact on how subjects value these experiences. They dubbed the first choice in an environment, the foundational choice, because this choice serve as the foundation from which following choices were measured. Ariely, Loewenstein, and Prele point to the role of awareness by noting that people are sensitive to decision making variable of which they are aware, and insensitive to ones that they are not directly aware of (because the don’t experience variation among unseen variables).


Heilman, CM, Bowman, D, and Wright, GP (2000). The evolution of brand preferences and choice behaviors of consumers new to a market. Journal of Marketing Research, 37, 139-155.

Hoeffler, S and Ariely, D (1999). Constructing stable preferences: A look into dimensions of experience and their impact on preference stability. Journal of Consumer Psychology, 8 (2), 113-139.

Rozin, P and Schiller, D (1980). The nature and acquisition of a preference for chili pepper by humans. Motivation and Emotions, 4, 77-101.

Tversky, A and Kahneman, D (1974). Judgment under uncertainty: Heuristics and biases. Science, 185, 1124-1131.

West, PM, Brown, CL, and Hoch, SJ (1996). Consumption vocabulary and preference formation. Journal of Consumer Research, 23, 120-135.



Steven S. Posavac, University of Rochester

David M. Sanbonmatsu, University of Utah

Frank R. Kardes, University of Cincinnati

Gavan J. Fitzsimons, Wharton

Taken together, six experiments provide compelling evidence that when consumers form brand preferences by evaluating brands in isolation, substantially different judgmental outcomes are achieved than if the brands had been evaluated in a comparative setting. Singular judgments of favorable consumer products are often characterized by a brand positivity biasCevaluations of focal brands belonging to several different categories were consistently more positive than warranted. Moreover, this judgmental tendency had implications for choice intentions in which the stated probability of purchase of the focal brand was higher than was the probability of choice of the average member of the category. Several experiments contribute some understanding of why the brand positivity effect occurs. In one experiment, when the judgmental context facilitated consideration of alternatives to a focal option, the likelihood of the brand positivity effect decreased. In a second study, as the consideration of alternatives to a focal brand increased, ratings and purchase likelihood of the focal brand decreased. These experiments, then, suggest that selective processing contributes to the brand positivity effect.

In addition to demonstrating this basic effect, our research also delineated important moderators of the brand positivity effect. As noted above, one experiment demonstrated that the brand positivity effect is unlikely to occur in judgmental contexts that facilitate the consideration of alternatives to a focal brand. This is the case because information about multiple alternatives may be gathered, thus reducing the likelihood of informational disparities arising between the focal and non-focal alternatives, and preventing the overly rosy perception of focal objects typically engendered by selective confirmatory hypothesis testing. Another experiment showed that being familiar with a focal brand increased the likelihood that judgments of it will be overly positive. Here, an individual trying to make a fair assessment of a familiar brand may be betrayed by the richness of information he or she has regarding it. Specifically, singular focus on a well-known brand may readily result in a large disparity between information considered regarding the focal versus non-focal brands. Moreover, evidence for the worth of a familiar brand will be readily obtained, and selective hypothesis testing may result in the focal brand appearing particularly attractive.

Although our results suggest that the brand positivity effect is ubiquitous, with both student and mature consumer samples demonstrating positivity effects in several product categories, our final experiment suggests that experts are much less likely to be influenced by context in this manner than non experts. One reason why experts may be less susceptible to the brand positivity effect is that they may have preexisting rankings or preferences that serve as the basis for judgment. Even if preferences are not previously delineated prior to a decision, there are several additional reasons to expect that experts will be likely to be immune to the brand positivity effect. First, experts may be likely to have a clearly defined set of criteria or standards with which to evaluate choice options. Accordingly, the context in which judgments are rendered is less likely to exert influence on the judgment because experts may bring criteria with them into the context, and be unlikely to use contextual cues to develop criteria. Second, experts are likely to possess substantial knowledge, including specific evaluations, of various options within the choice category. Consequently, disparities in information acquisition are less likely to arise in expert versus novice judgment. In a related vein, knowledge of alternatives and attribute information may reduce the likelihood of selective hypothesis testing.



Elizabeth Gelfand Miller, Wharton

Barbara E. Kahn, Wharton

A trip to a cosmetics counter these days reveals a remarkable thing: there is no red. There may be Panic Button or Enigmatic or Seductress, but there is no red (Schultz, 2001). In fact, such unusual, and often ambiguous, names are appearing in all sorts of product categories and with astounding success. What accounts for the phenomenal success of these naming strategies? Building on Grice’s (1975) theory of "conversational implicature," we propose that consumers react favorably to the unusual names because they are essentially assuming that the marketing messages convey some useful information. Specifically, since consumers cannot interpret the literal meaning of the ambiguous label, they focus on what they assume is the pragmatic or underlying meaning or reason for the communication effort. Since consumers believe that packaging or advertising would only provide positive information, they make positive attributions about the brand based on the ambiguous descriptions. The results of three experiments provide empirical support for our proposal and rule out some alternative explanations for the success of ambiguous naming strategies.

According to Grice’s (1975) theory of conversational implicature, conversations are guided by a set of tacit assumptions. These assumptions enable people to mean more than they say (i.e., convey nonliteral meanings) and to make sense of sentences that might literally be seen as non sequiturs. Specifically, Grice argues that listeners interpret speakers’ utterances based on the assumption that the speaker is being cooperative, unless they have reason to believe otherwise. Grice’s formulation of this cooperative principle embodies four conversational maximsBthose of quantity (make your contribution as informative as required, no more, no less), quality (make your contribution one that is true), relation (make your contribution relevant to the exchange), and manner (be perspicuous or easy to understand). These maxims suggest that one key assumption people make during conversations is that all information contributed by participants is relevant to the goal of the ongoing conversation.

Similarly, we hypothesize that consumers may assume that all information offered to them by the marketer, including the name of the product’s color or flavor, is meant to be relevant or informative to their purchase process and that they will consequently try to make sense of this information. Thus they will try to assign meaning to the ambiguous color or flavor names. If the color or flavor name is uninformative in the literal or semantic sense (as is often the case with atypical names), consumers will search for a pragmatic meaning or reason for the communication (Harris and Monaco 1978; Gruenfeld and Wyer 1992). In this context, focusing on a pragmatic meaning would result in positive attributions about the brand. That is, consumers may see these names, realize they are uninformative about the color/flavor, and then wonder why the manufacturer has provided this information. Their conclusion would then be that the manufacturer named the item in this manner to communicate some other information, perhaps something about the brand’s quality or stylishness. Such attributions are then likely to lead to an increased preference for these (ambiguously-named) items and a consequently, higher likelihood of purchase.

We test this theory in three studies. In the first study, we show that consumers do have preferences for the ambiguously named items and we provide evidence that the process behind this effect is cognitive in nature. In study 2, we provide further evidence for our specific process. In study 3, we rule out an alternative explanation for the results. Taken together, the findings support the notion that ambiguous names do affect preferences and that such names operate through a cognitive pathway, by which consumers make additional attributions about the ambiguously-named products, leading to increased preference and likelihood of purchase.


Grice, HP (1975). Logic and conversation, in Peter Cole and Jerry L. Morgan (Eds.). Syntax and Semantics III: Speech Acts, New York: Academic Press, 41-58.

Gruenfeld, DH and Wyer, RS, Jr. (1992). Semantics and pragmatics of social influence: How affirmations and denials affect beliefs in referent propositions. Journal of Personality and Social Psychology, 62, 38-49.

Harris, RJ and Monaco, GE (1978). Psychology of pragmatic implications: Information processing between the lines. Journal of Experimental Psychology: General, 107, 1-27/

Schultz, C (2001). Nailing down just the right name. The Trenton Times, August 7, B1.