The Surplus Identification Task and Limits to Multiattribute Consumer Choice

We introduce and demonstrate a novel experimental method for investigating the accuracy of consumer decision making. The Surplus Identification (S-ID) task exploits techniques from detection theory. Experimental control over surpluses is established by incentivizing participants to adopt a predetermined, objectively defined preference function. Surplus is then manipulated across multiple forced-choice trials in which participants decide whether a product does or does not confer a surplus at a given price. The S-ID task can be used to investigate how precision, bias, and learning vary with multiple properties of prices, attributes, and contexts. We demonstrate the task via a series of experiments that test the ability to apply a weighted adding decision strategy (with equal weights) as the number of product attributes increases. Imprecision increases sharply with additional attributes and larger trade-offs between them. Participants display persistent biases across the price range. These vary systematically with the number of attributes, implying a precision-bias trade-off. The findings have implications for models of multi-attribute choice and for consumer policy.