Explaining Variation in the Description-Experience Gap: A Small-Samples Account

ABSTRACT Decisions based on descriptions of options often diverge from decisions based on experience with the same options, yielding a description-experience gap. Early research suggested a uniform contrast: descriptions engendered relative overweighting of rare outcomes, and experience engendered relative underweighting. However, recent findings indicate that the direction of the gap can vary. For instance, the classic gap robustly arises in decisions between a sure outcome and a risky alternative but can be eliminated or reversed when both options are risky. Thus, experience-based choices can appear consistent with relative underweighting, overweighting, or neither. To explain such variation, we propose that under experience, preferences reflect reliance on small samples: that is, they are governed by recall of a few encountered outcomes. Psychologically, whereas expected utility, prospect theory, and other traditional models of preferences under description ask “what will each option be worth to me on average,” small sample accounts ask “which sample beats the other?” But rare outcomes are often absent from limited draws from a wider distribution, which implies that experience-based preferences will overemphasize common outcomes. Given the particulars of the available options, this can yield apparent underweighting of rare outcomes, apparent overweighting, or neither. We tested our small samples account in two preregistered experiments (N = 674) and a reanalysis of data from 29 studies. In all cases, the predictions of the small-samples account consistently tracked variation in rare-event sensitivity and explained the direction of the description–experience gap.