Issues in Consumer Choice With Uncertain Product Outcomes

David E. Hansen, Virginia Polytechnic Institute and State University
[ to cite ]:
David E. Hansen (1992) ,"Issues in Consumer Choice With Uncertain Product Outcomes", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 175-176.

Advances in Consumer Research Volume 19, 1992      Pages 175-176


David E. Hansen, Virginia Polytechnic Institute and State University

In consumer choice, two fundamental kinds of uncertainty have been identified: uncertainty over the importance of attributes (e.g., is performance in a car more important to me than luxury) and uncertainty over product outcomes or the level of attributes (e.g., is gas milage going to be as good as the EPA sticker claims). This paper focuses on uncertainty over product outcomes or risky choice, a topic that has been studied in other fields but that has received little attention in the consumer literature. In risky choice there are two emerging directions of research. One is normative, in that it focuses on how people should make decisions under ideal risky choice conditions in order to maximize their expected utility (Hauser and Urban 1978). The other is more descriptive, in that it focuses on how people actually do make decisions under ideal conditions, such as the work by Kahneman and Tversky (1979). Another descriptive area involves uncertain choice under less than ideal risky choice conditions such as ambiguous probabilities (Kahn and Sarin 1989; Hansen 1991) and is the area of concern in the present paper.

Under ideal conditions in uncertain choice, each choice alternative (e.g., a car) is described as having different possible outcomes or levels for the same attribute (e.g., time between auto repairs), each outcome with its own explicit probability. These ideal conditions describe what is known as a gamble (e.g., a 20% chance that the repair interval is only one month and an 80% chance that it is six months). Theories of risky choice maintain that individuals evaluate alternatives as if they are forming statistical expectations based on the probability-weighted outcomes of gambles which implies the use of a compensatory choice process. Gambles have become the decision theorist's "fruit fly", since gambles are assumed to be representative of the way people set up risky decisions and because they are easy to work with. However, there is a problem in using gambles to describe consumer choice, since in most consumer choice situations explicit well-defined probabilities rarely exist and are otherwise difficult to formulate (Kahneman, Slovic, and Tversky 1984). Furthermore, people may not even recogize that the variation in the attribute level is due to randomness, although they can identify variation and thus uncertainty in attribute levels (Hogarth 1987). Under the less than ideal conditions found in consumer settings, decisions based on expections or probabilities may be quite difficult, inviting the use of decision shortcuts or heuristics (Payne 1982).

The main issue here concerns how consumers actually make decisions under uncertainty when probability is not explicitly given, cannot easily be formulated, and thus may be ignored or supressed. Much of the descriptive research on how people make decisions under uncertainty has found that people are highly concerned with negative information or losses (Einhorn and Hogarth 1981). It has also been found that outcomes are segregated into losses and gains relative to some neutral reference point and that losses loom larger than gains (Kahneman and Tversky 1979), that probability is often ignored (Einhorn and Hogarth 1981), and that most decisions are made using heuristics (Payne 1982). Because of this, it may be that in decisions without explicit probability consumers with little ability to formulate subjective probabilities use a choice heuristic that focuses on potential losses. By focusing on the worst possible loss of each alternative and minimizing losses, the alternative chosen should have the smallest potential loss of all alternatives. This choice strategy is non-compensatory in that it focuses only on losses as opposed to the compensatory nature of expectations which consider both losses and gains. Statistically naive consumers may actually use such a strategy to suppress uncertainty in the decision, if they do not know how to deal with it otherwise.

This leads us to ask whether individuals who are more experienced in uncertain choice deal with the lack of explicit probabilities differently than the naive consumers. Despite the ambiguity of the less than ideal conditions for consumers, it may be that individuals with statistical training look for ways to deal with the uncertainty rather than trying to suppress it. Hansen (1991) found that when sample statistics were used as choice data, the salience of the displayed data guided the use of a compensatory choice strategy by buyers and managers. Consumers who are not statistically naive may also use a more compensatory strategy in which potential gains and losses are considered, which can produce different decisions than those of the statistically naive consumers. Such a strategy does not make the implicit assumption of the naive consumer that only a loss will occur, since weight is given to both losses and gains. The issue this raises is what difference between naive and experienced consumers could explain the use of such different strategies. One reason may simply be that the trained consumers are able to recognize the existence of randomness and formulate subjective probabilities to help make the decision.

The implication of these proposals, if empirically supported, is that when statistically naive consumers make decisions under uncertainty, they may not be using the same decision formulation or choice strategy assumed by the decision theorist. That is, they may not formulate the decision problem as a gamble, let alone solve it via expectations. They may be using simple choice strategies which focus only on losses, unlike less naive decision makers who may try to use more of the outcome distribution (i.e., losses and gains). This coincides with the notion that the naive consumers may not be using concepts related to probability, either due to lack of recognition of the need to use them or lack of ability to use them. In effect their choice strategy may be a way to suppress uncertainty. It may also be construed as "certaintizing" the decision and is similar to Kahneman and Tversky's observation that very high or low probabilities are changed to certainties--the certainty effect (1979). However, whether such a choice strategy is used merely to simplify the decision or because naive consumers are unable to recognize randomness or formulate subjective likelihoods is another issue in need of empirical research.


Einhorn, H. J. and R. M. Hogarth (1981), "Behavioral Decision Theory: Processes of Judgment and Choice, " Annual Review of Psychology, 32, 52-88.

Hansen, D. E. (1991), "Salience Bias and Format Effects With Descriptive Statistics in Buyer Choice Under Uncertainty," working paper, Virginia Polytechnic Institute and State University.

Hogarth, R. M. (1987), Judgment and Choice, Second Edition, New York: Wiley.

Kahn, B. E. and R. K. Sarin (1989), "Modeling Ambiguity in Decisions Under Uncertainty," Journal of Consumer Research, 15, 265-272.

Kahneman, D., P. Slovic and A. Tversky (1984), Judgment Under Uncertainty: Heuristics and Biases, Cambridge: Cambridge Press.

Kahneman, D. and A. Tversky (1979), "Prospect theory: An Analysis of Decision Under RIsk," Econometrica, 47, 263-291.

Payne, J. (1982), "Contingent Decision Behavior," Psychological Bulletin, 92, 382-402.