The Confirmation Bias : Does it Result from Mental Accounting?

Abstract

An experiment was conducted to explore whether the confirmation bias can be interpreted in terms of prospect theory: Useful butdisconfirming information is assessed more negatively than information confirming own assumptions. Consequently it was predicted thatpeople value utility of relevant information less than satisfaction about irrelevant information which is compatible with their preconceptions.This hypothesis was tested within the Wason selection task paradigm. After selecting the cards to be turned, subjects were told whether thecard was consistent with the given statement. Next they had to assess their satisfaction about this information and its utility. Satisfactionratings on cards consistent with the statement were higher than utility ratings on relevant results. Utility scores on relevant and irrelevantcards as well as for consistent and falsifying results were assessed rationally. However, all subjects failed to select only relevant cards.Explanations for these inconsistent findings are discussed.