CaltechTHESIS
  A Caltech Library Service

Neurometrically Informed Mechanism Design and the Role of Visual Fixations in Simple Choice

Citation

Krajbich, Ian Michael (2011) Neurometrically Informed Mechanism Design and the Role of Visual Fixations in Simple Choice. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/8Y3X-6W36. https://resolver.caltech.edu/CaltechTHESIS:10262010-130106361

Abstract

The young field of neuroeconomics has already produced many important insights into the neurobiological underpinnings of decision making. However, at this early stage it is still unclear how much influence the field will have on mainstream economics. Here, I show how a neuroeconomics approach can shed light on two classic economic problems. First, I show that it is possible to predict individuals’ values for public goods, using functional magnetic resonance imaging (fMRI)-based pattern classification. With such predictions in hand, I demonstrate that it is possible to solve the free-rider problem, by taxing individuals based both on the values that they themselves report and on the predicted values (using fMRI). I go on to more generally prove that by using any informative signal of value, it is possible to overcome classic impossibility results in mechanism design. This allows us to construct mechanisms that simultaneously satisfy dominant strategy incentive compatibility, voluntary participation, budget-balance and social efficiency. Such mechanisms were previously thought to be impossible. I demonstrate how to construct such mechanisms, and test them in three different public goods experiments. Second, I show that individuals’ looking patterns are critical to the decision making process. When people make choices between options, they tend to look back and forth between them. One might think that these “fixations” are an unimportant by-product of the choice process, but I demonstrate that they are in fact intimately tied to the comparison process. By using a variant of the drift-diffusion models from the perceptual decision making literature, I find that fixations seem to bias the accumulation of evidence towards the item that is being looked at. Therefore, if one spends more time looking at one item over the other, then one is more likely to choose that item. Critically, I am able to show that this effect is not due to subjects looking longer at preferred items. The model has deep implications for how looking patterns (treated as exogenous) should bias choices, and I confirm these predictions using eye-tracking data from subjects choosing between snack foods.

Item Type:Thesis (Dissertation (Ph.D.))
Subject Keywords:neuroeconomics
Degree Grantor:California Institute of Technology
Division:Humanities and Social Sciences
Major Option:Behavioral and Social Neuroscience
Thesis Availability:Public (worldwide access)
Research Advisor(s):
  • Rangel, Antonio
Thesis Committee:
  • Rangel, Antonio (chair)
  • Camerer, Colin F.
  • Adolphs, Ralph
  • Ledyard, John O.
Defense Date:9 July 2010
Additional Information:Chapter 1 was published in Science, Volume 326, no. 5952, October 2009, pp. 596-599 : http:dx.doi.org/10.1126/science.1177302 Chapter 3 was published in Nature Neuroscience, Volume 13, no. 10, June 2010, pp. 1292-1298 : http:dx.doi.org/10.1038/nn.2635
Funders:
Funding AgencyGrant Number
NSF IGERTUNSPECIFIED
Record Number:CaltechTHESIS:10262010-130106361
Persistent URL:https://resolver.caltech.edu/CaltechTHESIS:10262010-130106361
DOI:10.7907/8Y3X-6W36
Related URLs:
URLURL TypeDescription
https://doi.org/10.1126/science.1177302DOIArticle adapted for ch. 1
https://doi.org/10.1038/nn.2635DOIArticle adapted for ch. 3
https://doi.org/10.1038/nn0911-1217bErrataErratum published for ch. 3 article
Default Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:6163
Collection:CaltechTHESIS
Deposited By: Ian Krajbich
Deposited On:07 Dec 2010 16:50
Last Modified:09 Oct 2019 17:06

Thesis Files

[img]
Preview
PDF - Final Version
See Usage Policy.

7MB

Repository Staff Only: item control page