Searching for the highest number

Abstract

When viewing a collection of products how does a consumer decide which one to buy? To do this task, the consumer not only needs to evaluate the desirability of the products, taking into account factors such as quality and price, but also needs to search through the products to find the most desirable one. We studied the search process using an abstraction of a common consumer choice task. In our task, observers searched an array of numbers for the highest. Crucially, the observers did not know in advance what this number would be, which made it difficult to know when the search should be terminated. In this way, our search task mimicked a problem often faced by consumers in a supermarket setting where they also may not know in advance what the most desirable product will be. We compared several computational models. We found that our data was best described by a process that assumes that observers terminate their search when they find a number that exceeds an internal threshold. Depending on the observer and the circumstances, this threshold appeared either to be fixed or to decrease over the course of the trial. This threshold can explain why in some situations the observers terminate the search without inspecting all the numbers in the display, whereas in other situations observers act in a seemingly irrational manner, continuing the search even after inspecting all the numbers.