Ph.D. Candidate
Sustainable Development
604 West 115th Street Apt. 1B
New York City, NY 10025

Phone: (617) 335-1232
[email protected]

Research Statement

Systems involving humans – such as political or military coalitions, banking systems and universities – depend not only on their parts and on some fixed relationships between them, but also on human beliefs about both parts and relations. Beliefs alter behavior, which in turn has the potential to alter both a system's parts and their relationships to each other, making the systems either adaptive or maladaptive. My research interests concern human beliefs about systems, the diversity of these beliefs – cognitive diversity – and how both emerge from the complexity of the systems that are their content, as well as from social processes and cultural factors. I then seek to link changes in cognitive diversity – through mechanisms such as conformity – to phase transitions in social movements and systems such as markets and democracies.

During my PhD, I have built an open-source experimental platform (bettingisbelieving.com) in order to study belief dynamics in the laboratory and on the Internet. The innovation of this platform is that it allows participants to record their beliefs, in addition to interacting in a game. In my dissertation, I use this platform to test a theory of financial bubbles by Scheinkman and Xiong (2003). On the platform, participants view data of a financial system, and construct causal models of the system. They then bet on the value of one focal variable in a prediction market, and can adjust their model over time in light of their performance and new data. Thus, the data I collect are not restricted to participants' bets, as is common, but they feature mental models – structures, formalized as Bayes Nets, that capture how participants process and interpret signals – in line with a cognitive theory of human learning (Griffiths, Kemp, and Tenenbaum, 2008). This formalism allows me to extend Scheinkman's theory of bubbles to study how the probability of bubble formation is affected by the complexity of the financial system. Because the platform can simulate any stochastic causal system, at any level of complexity, it allows me and others to study system-dependent social learning experimentally.

Next, I will shed light on how institutions, social networks, and cultural diversity influence belief formation, group-think and the polarization of opinions. In the near term, I have plans for three projects. The first is to study how culture constrains minds, the second turns to the influence of political institutions on the polarization of beliefs, and lastly, my third project will study how culture affects politicians' supply of emotional and logical arguments. In all three projects, I bring together scientific models of human cognition with political economic and sociological theories to derive hypotheses about the social conditions that affect collective cognition. Simple modifications of my platform will then allow me to test these hypotheses experimentally.

My modeling framework allows me to investigate how social and political influences are mediated by the complexity of the environments in which people act. Intuitively, if circumstances are easy to understand, beliefs should not differ much, no matter what the social and political forces are, while when circumstances are complex and hard to comprehend, political opportunists and social forces can separately and interactively bias and homogenize opinions.

In my next project, I will address the interaction between culture and cognitive diversity. How do some cultural conditions jointly constrain belief systems, leading to more or less cognitively diverse communities? Is it the case that in highly hierarchical societies, individuals at the top of the hierarchy influence everyone else, as Bénabou (2012) suggests? Does the influence of culture depend on how complex the systems that people engage in are? My experimental platform allows me to control the complexity of the systems that participants are mentally engaged with and which they are deriving benefits from. In addition, my platform collects Facebook profile information from participants. These data include measurements of social networks and a wealth of social, cultural and geographical information which – unlike survey data – have been posted by the participants themselves over the course of several years, presumably in part with the intent to communicate cultural identity. Using these data and some techniques detailed by Gross and Rayner (1985), I can derive two well motivated measures of culture pertaining to communities (Douglas, 1978; Thompson, Ellis, and Wildavsky, 1990), called "group" and "grid" and experimentally study their effects on collective cognition.

The significance of this work lies in its ability to bridge rationalist and constructivist approaches to human collective computation. In search of universal explanations, rationalists, as are most economists and political scientists, give short shrift to the variety of cognitive constructs, which to social constructivists constitute exactly the substance that social scientists should be concerned with. Social constructivists on the other hand, as are most anthropologists and sociologists, by and large resist generalization and quantitative models because they see beliefs as too contingent to find it possible to generalize at all. In my work, I recognize the contingency of beliefs, but seek the cultural and systemic conditions that explain their diversity. In turn, I seek to understand how this cognitive diversity leads to better or worse collective computations.

In a second project, I will adapt a theory of deliberation dynamics (Hafer and Landa, 2007) to develop a theory of political persuasion in more or less complex policy environments. In a system in which there are a small number of politicians who are competing to supply logical arguments, how do the institutional arrangements between speaking politicians and listening voters affect polarization of beliefs? Do more competitive settings lead to more polarization or to more subtle differentiation? Does this depend on the complexity of the circumstances in which politicians compete for votes? In order to answer these questions I will add political institutions to my platform, so that a few experimental participants compete for the votes of large numbers of others by supplying reasons for and against certain actions to be taken within the system.

This project has the potential to shed light on how political institutions affect a society's ability to understand and manage complex problems. Building on this project, I plan to also evaluate the conditions under which speakers who compete for political support find it more beneficial to shape listeners' beliefs via emotional cues that are derived from shared cultural references, rather than via logical arguments. This last objective builds on Glaeser (2004) as a theoretical foundation for how politicians might benefit from supplying emotional frames and on Cushman and Greene (2012) for a "dual-process" theory of cognition whereby people are predicted to become conflicted when an emotional concern and a logical concern are incongruous.

In all, the subject of my work is the human mind in its social, economic, political and cultural embeddings. There are forces that bring minds closer together and some that polarize them, while yet others operate on minds so as to set the tone of the debate in either emotional or logical terms. I study various forces operating on minds, using modern interdisciplinary methods and tools. The principle aim of my work is to make significant strides toward understanding how collectives can better manage beliefs to increase understanding between groups and facilitate the discovery of solutions to complex problems.


Roland Bénabou. Groupthink: Collective delusions in organizations and markets. Review of Economic Studies, 2012.

F.A. Cushman and J.D. Greene. Finding faults: How moral dilemmas illuminate cognitive structure. Social Neuroscience, 2012.

M. Douglas. Cultural Bias. London: Royal Anthropological Institute, 1978.

Edward L. Glaeser. Psychology and the market. NBER Working Paper 10203, 2004.

T. L. Griffiths, C. Kemp, and J.B. Tenenbaum. Bayesian models of cognition., chapter in Cambridge Handbook of Computational Cognitive Modeling. Cambridge University Press., 2008.

Jonathan L. Gross and Steve Rayner. Measuring Culture. Columbia University Press, 1985.

Catherine Hafer and Dimitri Landa. Deliberation as self discovery and institutions for political speech. Journal of Theoretical Politics, 2007.

José A. Scheinkman and Wei Xiong. Overconfidence and speculative bubbles. Journal of Political Economy, 2003.

M. Thompson, R. Ellis, and A. Wildavsky. Cultural Theory. Westview Press, Colorado, CO, 1990.