BEAT Seminar - Investigating preference reversals and response times in risky choice using different equivalence methods

Abstract: A robust finding in risky choice is that people’s preferences in choices between gambles are often inconsistent with the values they assign to those gambles. The classic pattern of these preference reversals is to give a higher Certainty Equivalent (CE) to a low probability large outcome bet ($-bet), than a high probability small outcome bet (P-bet), despite preferring the P-bet to the $-bet in direct choice. In recent years there have been attempts to develop dynamic models of both pricing and choice tasks, which produce these preference reversals using a common evidence accumulation process. However, CEs are just one method for eliciting the value of a bet, and such a common process should apply equally to other valuation methods. Across four studies we test for reversals using 8 methods, eliciting several types of risky amount equivalents (AEs) and probability equivalents (PEs). While for AEs existing theories can capture the patterns we observe, we identify several unexpected PE reversals. Additionally, we find weak evidence that response times for valuations are impacted by the similarity between the elicitation method and the bet under consideration. This is partially consistent with the anchoring processes assumed by current dynamic models.