Moderation of preference reversals in the long run.

Abstract

The evaluation of a monetary gamble requires consideration of two basic types of information: (a) the probabilities associated with winning and losing, and (b) the amounts to be won and lost. The well-documented phenomenon of preference reversals provides strong evidence that the processing of this information depends on how the evaluation is expressed (Goldstein & Einhorn, 1987; Grether & Plott, 1979; Lichtenstein & Slovic, 1971, 1973; Lindman, 1971; Slovic & Lichtenstein, 1983). When presented with two gambles close in expected value, many subjects choose a bet with a high probability of winning a modest amount of money (P-bet) over a bet with a low probability of winning a large amount ($-bet), even though they place a higher minimum selling price on the $-bet. When confronted with the apparent inconsistency in their behavior and given the chance to change either of their responses, the majority of these subjects stand by their decisions (Slovic, 1986). In the typical preference reversal task, response modes may vary in two ways: (a) subjects either make a choice between two lotteries or make a judgment of each separately, and (b) the focus is either on the monetary value of the gamble or on its attractiveness. Goldstein and Einhorn (1987) manipulated these two factors independently and demonstrated preference reversals associated with each. Although recent research (Tversky, Sattath, & Slovic, 1988) has demonstrated preference reversals for nonprobabilistic stimuli, the vast majority of preference reversal studies have used gambles as stimuli. In this article, we explore the possibility that these preference reversals are largely due to the way in which people evaluate probabilities for unique, one-shot events. Two experiments test this hypothesis by varying the number of times a gamble