Profile picture of Geoffrey Castillo

Geoffrey Castillo

Senior research fellow, Centre for Behavioural Sciences, Nottingham Trent University

geoffrey.castillo@ntu.ac.uk

CV

Publications

Working papers

  • Similarity and social discounting, with Benjamin Beranek.
    Social discounting refers to the idea that decision-makers discount payoffs as a function of social distance. We introduce a method to measure social distance using interpersonal similarity; that is, how similar or different others are to the decision-maker. We use data from our own preregistered experiments as well from an existing, independently conducted, lab-in-the-field experiment to estimate the structural parameters of social discounting and find evidence for it. Our experiments control for competing explanations to isolate the effect of similarity and thus show that people have a preference for more similar others. Our estimates imply that in order for a decision-maker to willingly forgo $1 and have it instead benefit a dissimilar other, then it would need to increase to at least $1.25. We also find evidence for quasi-hyperbolic social discounting.
  • The Virtual Bingo Blower: An open-source tool to generate ambiguity and risk in experiments, with Ola Andersson and Erik Wengström.
    We propose the Virtual Bingo Blower (VBB) as a way to generate credible risk and ambiguity in computerized experiments. Using a physics engine—a computer simulation of a physical system—the VBB simulates a conventional bingo blower. Different aspects of the VBB, such as the number of balls, their color, and their speed, can be easily modified. In an online experiment, we measure ambiguity attitudes and vary the source of ambiguity, using either the VBB or natural events. We find that the VBB and natural events result in a similar degree of ambiguity aversion. Further, we find that, by manipulating the number of balls, the VBB can be used to manipulate the perceived level of ambiguity.
  • How does the way we represent lotteries affect risk preferences?, with Chris Starmer.
    In a risk preference elicitation experiment, we manipulate the way we represent lotteries. We represent probabilities, payoffs, or both at the same time. We find that the representation has no effect on the raw, elicited certainty equivalents. We find, however, a significant effect on the structural parameters estimated via maximum likelihood.
  • Do different people report the same social norms?, with Lawrence Choo and Veronika Grimm.
    If the Krupka-Weber (2013) norm-elicitation task captures pre-existing social norms, then the elicited norms should be independent of one’s role in a game or one’s social preferences. We test this idea in a complex game that features rich interactions. We find that different people, even when they have conflicting incentives, report the same social norms. Our results further validate the use of the Krupka-Weber task to measure social norms.

In progress

Tools

  • bingo-blower.js, a virtual bingo-blower for ambiguity experiments
  • ios.js, a continuous implementation of the Inclusion of Other in the Self scale

Posts