Impact of Number of Alternatives and Spatial Arrangement on Decoy Effects

The way in which people make decisions is largely guided by the context of the choice set. Choice sets that contain decoy alternatives can result in context effects that violate rational principles of decision making. Most studies on contextual choice manipulate preference among two alternatives in the context of a decoy using three option choice sets. However, many real-world decisions are made among choice sets with many more than three alternatives, such as in online shopping. We tested for attraction and compromise decoy effects in choice sets with varying numbers of alternatives by using a within-subjects preferential grocery shopping task. Experiment 1 (n = 50) examined the effects of attraction and compromise decoys in choice sets with three and nine alternatives. Experiment 2 (n = 68) tested for the effect of presentation order on attraction and compromise decoys in nine alternative choice sets by presenting alternatives in one of three organizational formats: best to worst price, best to worst quality, and randomized. Experiment 3 (n = 40) used eye tracking while presenting attraction decoy choice sets with three, nine, and 15 alternatives and manipulating the presentation order in two conditions, best to worst price and randomized. Significant preference reversals were replicated for both attraction and compromise decoys in three alternative choice sets. Context effects were found to be significantly reduced as the number of alternatives increased to nine and reduced further when increased to 15 alternatives. Overall, preference reversals caused by the attraction decoy were significant across all levels of alternative number. The compromise decoy in nine alternative choice sets did not produce the hypothesized preference reversals; however, analysis of individual alternative choice frequencies revealed a local preference for alternatives nearest to the contextual average. The present study produced conflicting results on the effect of presentation order on decoy effects, with Experiment 2 revealing a significant increase in attraction effects caused by ordered compared to randomized choice sets and Experiment 3 producing no effect of ordering. Analysis of eye tracking data showed that participants engaged in more dimension-wise comparisons as the number of alternatives increased, but, contrary to previous research, the proportion of alternative-wise to dimension-wise transitions was not predictive of decoy effects. In congruence with previous research, the proportion of the total information attended to in a choice set decreased as the number of alternatives increased, and those who attended to less information were more likely to select decoy alternatives. Participants were also more likely to engage in a lexicographic decision-making strategy as the number of alternatives increased. Although the effects of spatial arrangement are inconclusive, the present study provides insight into how the number of alternatives in a choice set influences how people make decisions