Computational Biases in Decision-Making

Citation

Janowski, Vanessa (2012) Computational Biases in Decision-Making. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/J1AP-S293. https://resolver.caltech.edu/CaltechTHESIS:05172012-130309762

Abstract

Neuroeconomics has produced a number of insights into economics, psychology, and neuroscience in its relatively short existence. Here, I show how neuroeconomics can inform these fields through three studies in social decision making and decision making under risk. Specifically, I focus on computational biases inherent in our daily decisions.

First, using functional magnetic resonance imaging (fMRI), I examine how we make decisions for others compared to ourselves. I find that overlapping areas of the ventromedial prefrontal cortex (vmPFC) are involved in both types of decisions, though decisions for others are modulated by areas involved in social cognition. Specifically, activity in the inferior parietal lobule (IPL) encodes a variable measuring the distance between others’ and our own preferences, suggesting that we may anchor our choices for others on our own preferences and attempt to modulate these preferences with what we know about others.

Second, I investigate how visual looking patterns can critically influence the computation and comparison of values. In a first study using eye-tracking, I investigate the relationship between loss aversion and attention and find a correlation between how loss averse subjects are and how long they look at losses vs. gains when evaluating mixed gambles. Importantly, I show that this effect is not due to subjects simply looking longer at items of higher value. In a second study using Mouselab, I show how attention influences multi-attribute choice. I find that the display of different attributes has a significant effect on search among those attributes and, ultimately, choice.

Thesis Files