The value heuristic in judgments of relative frequency.

Estimating the relative frequency of a class of objects or events is fundamental in subjective probability assessments and decision making (Estes, 1976), and research has long shown that people rely on heuristics for making these judgments (Gilovich, Griffin, & Kahneman, 2002). In this report, we identify a novel heuristic for making these judgments, the value heuristic: People judge the frequency of a class of objects on the basis of the subjective value of the objects. Why use object value as a cue to object frequency? A psychological and economic principle of valuation is that the scarcity of objects increases their value (e.g., Brock, 1968; Hirshleifer, Glazer, & Hirshleifer, 2006; McKenzie & Chase, in press). Two different theoretical perspectives led us to hypothesize that people use this relationship to heuristically—and inversely—infer the frequency of a class of objects from their value. Kahneman and Frederick (2002) proposed that heuristic judgments rely on attribute substitution: In assessing an attribute (the target attribute, e.g., the frequency of an object) that is less readily assessed than a related property (the heuristic attribute, e.g., the value of the object), people unwittingly substitute the simpler assessment of the heuristic attribute for the assessment of the target attribute. This can bias judgments when the heuristic attribute is not diagnostic of the target attribute. Yet our proposition is also consistent with Brunswik’s (1952) notion of probabilistic functionalism. Subjective estimates of a distal variable rely on judgmental cues that are probabilistically related to it. Brunswik’s concept of vicarious functioning denotes the ability to analyze several cues that are correlated with the distal variable (e.g., frequency) and to substitute one cue (e.g., value) for another (e.g., ease of recall), a process that yields valid alternative routes to the distal variable. If people use the value of a class of stimuli heuristically to infer the frequency of the stimuli, they will assess more valuable stimulus classes as being less frequent even when value is not diagnostic of frequency. To test this hypothesis, we manipulated the value of stimuli only after participants had encoded them; then, participants assessed the frequencies of the stimuli. A randomization procedure ensured that stimulus value could not be diagnostic of frequency.