A tale of two biases: choosing versus rejecting and the decision-by-sampling model

Abstract

ABSTRACT When subjects are exposed to irrelevant background numbers, the decision-by-sampling model maintains that the subjective value of a particular target attribute is determined by its relative rank position within the range of background numbers. The subjective distance will be telescoped (shrink) when the irrelevant numbers lie between (above and/or below) two target attribute values. Prior research indicates that telescoping the two lottery prizes increases preference for the riskier, higher value (enriched) lottery prize versus the safer, low-value (impoverished) prize, whereas shrinking the subjective distance increases preference for the low risk, smaller value (impoverished) prize. According to the compatibility hypothesis, relative preference for an impoverished option over the enriched option should also increase when subjects are invited to reject, rather than select, an option. The compatibility hypothesis posits that a reject instruction makes negative attributes more salient to the decision-maker. The current study tests whether the relative preference for the impoverished prize predicted by decision-by-sampling model (in a decision environment that fosters the compression of the subjective distance) may be exacerbated when the decision is framed with a reject instruction. Our results fail to support both the predicted moderating effect of task instructions (choose versus reject) on decisions predicted by the decision-by-sampling model and the predictions of the compatibility hypothesis and decision-by-sampling model. The discussion considers the absolute, rather than relative, size of the impoverished option and the quantity of choose/reject decisions as possible explanations for the inconsistency of our results with those reported in previous studies.