skip to main content

Ulric B. and Evelyn L. Bray Social Sciences Seminar

Monday, April 19, 2021
12:00pm to 1:00pm
Add to Cal
Online Event
Predictable Effects of Visual Salience in Experimental Decisions and Games
Xiaomin Li, Postdoctoral Scholar in Behavioral Economics and Neuroeconomics, Caltech,

Abstract: Since attention is scarce, economics can analyze it. This paper is about a specific kind of attention{ bottom-up visual salience, which depends only on features of a visual stimulus. This kind of salience is largely automatic, effortless, and exogeneous to a person's \top-down" perceptual goals. We use algorithms that were carefully trained to predict salience values for each pixel in any image. The economic question is whether these salience values help explain economic decisions. Our first experimental analysis shows that when people pick between sets of fruits that have artificially induced value, predicted salience (which is uncorrelated with value by design) leads to mistakes. Our second analysis uses evidence from games in which choices are locations in images. When players are trying to cooperatively match locations, predicted salience is highly correlated with the success of matching (r=.57). In competitive hider-seeker location games, players choose salient locations more often than predicted by the unique Nash equilibrium. This fact creates a disequilibrium "seeker's advantage" (seekers win more often than predicted in equilibrium). The result can be explained by level-k models in which predicted bottom-up salience influences level-0 choices and also influences top-down perceptions and choices by higher-level players. The third analysis shows effect of visual salience in matrix games is small and statistically weak. Potential applications to behavioral IO, price and tax salience, nudges, and design, and visually-influenced beliefs are suggested.

Written with Colin F. Camerer.

For more information, or if you are interested in attending this online seminar, please contact Letty Diaz by email at [email protected].