A Variety of Effects: How Variety Alters Experience by Influencing Both Choice and Hedonic Value

This paper makes a conceptual distinction between the effects of variety on choice and its effects on the hedonic value of chosen alternatives, to help explain consumers’ persistent and seemingly perverse tendency to choose less-preferred items from a varied set. In particular, we show that although high-variety sets have an unfavorable effect on choice (in the sense that worse items are chosen), an item chosen from a large set is preferred to that same item chosen from a smaller set. Revisiting data previously analyzed in Ratner, Kahn, and Kahneman (1999, Experiment 5), using a model of the Heckman type (Heckman 1979), we find substantially stronger effects of variety on hedonic value than in the earlier analysis. Our re-analysis does not contradict RKK’s earlier findings of a retrospective bias towards varied sequences, but suggests a specific mechanism that may underlie them: memory may favor varied sequences if it is driven by actual consumption experience, rather than choice context. A VARIETY OF EFFECTS: HOW VARIETY ALTERS EXPERIENCE BY INFLUENCING BOTH CHOICE AND HEDONIC VALUE Decision makers seem to persist in choosing less-preferred items from a set despite the apparently diminished enjoyment that results. This phenomenon, often described as varietyseeking, has been the focus of much research (Kahn and Lehmann 1991, Kahn, Ratner, and Kahneman 1997; Menon and Kahn 1995; Ratner, Kahn, and Kahneman 1999; Ratner, Novemsky, and Kahneman 1998; Read and Loewenstein 1995; Simonson 1990). Consumers often act as though they believed varied sequences are superior to repetitive ones. For example, Simonson (1990) found that subjects choosing three candy bars at once for consumption over three future occasions chose a wider variety than subjects choosing on three separate occasions for immediate consumption. Ratner, Kahn, and Kahneman (1999) investigated whether decision makers’ occasional tendency to choose less-preferred items might be explained as a reaction to satiation (i.e., diminished hedonic value) on favorites due to overly-frequent consumption. Inconsistent with that explanation, they found that “more varied sequences result in diminished enjoyment during consumption,” even though “participants did not become satiated with top-ranked songs” (p. 11). Despite this, consumers’ memories were positively biased for varied sequences. Similarly, Ratner, Novemsky, and Kahneman (1998) showed a prospective bias in favor of varied sequences. This paper makes a simple but telling conceptual distinction between the effects of variety on choice and its effects on judgment (i.e., assessment of the hedonic value of a choice), to help explain this paradoxical preference of consumers for variety. We demonstrate that 1 To be consistent with earlier researchers (specifically, Ratner, Kahn, and Kahneman 1999), we are concerned here with a particular form of variety, that created by a set in which each alternative appears only once. In this