2013 Graduate Workshop in Complexity and Computational Social Science

Student Projects

Each student began a research project during the two-week workshop. Below are brief descriptions of these various projects. These projects will form the basis for dissertation chapters and/or journal articles.

Jon Atwell, Sociology, University of Michigan (atwell@umich.edu).

Jon is intrigued by social construction, in particular, how mutually influential descriptions and knowledge of the world can lead to self (in)validating outcomes. He investigates this system using a model of environmental influence where, as agents attempt to improve their own payoffs, they alter the payoff landscapes of the other agents. The current model appears to lead to a rich set of behavioral regimes that can capture the critical cases of interest. 

Chenna Reddy Colta, Computational Social Science, George Mason University (ccotla@gmu.edu).

Chenna is using experimental results about the types of learners that arise in public-goods games to explore the impact of network structure on achieving cooperative outcomes. He models agents as adaptive learnings, using algorithms that are able to capture both reinforcement- and belief-based learning. He estimates the underlying parameters and their population distributions using maximum likelihood and Bayesian clustering techniques. Based on these estimates, he creates various network configurations populated by the calibrated agents, and derives estimates of overall cooperation. Enumerating all seven agents (nodes) connections, he finds that denser graphs encourage more contributions to the public good, a result that is at odds with current speculations. These predictions will be verified using laboratory experiments.

Martha G. Alatriste Contreras, Economics, Aix-Marseille School of Economics (martha.alatriste@ehess.fr)

Martha is studying how technological change influenced the French economy over the last fifty years. To perform this analysis, she takes existing input-output data (data that captures the various production demands across all economic sectors) and transforms them into an implied production network. This transformation provides a new window into how technological changes influence economic growth. Using this new framework, she is exploring the network signatures of technological changes and developing new hypotheses about their underlying mechanisms.

Marion Dumas, International and Public Affairs, Columbia University (mmd2172@columbia.edu).

Marion wants to understand the evolution of case law under precedent rules. To frame this problem, she assumes that cases differ across two dimensions, and that judges must develop schemas that classify cases into a binary decision (for or against). Judges seek patterns that respect the schemas in use (stare decisis), accommodate past decisions, and align with their own preferences. She has developed various methods to characterize the degree of path dependence in this system, and is using these to link the system's behavior to different degrees of stare decisis driving the judges. 

Ursula Kreitmair, Public and Environmental Affairs, Indiana University (ukreitma@indiana.edu).

Ursula is looking at the interaction between local information and cooperation in public-goods problems. She focuses on both the connections between agents and the type of information transmitted across those connections. Depending on the condition, agents can observe the action, the behavioral rule, or both, of their neighbors. Preliminary results suggest that systems with more links or transmitted information, lead to higher levels of cooperation. Ongoing work is exploring the impact of noisy signals on the system---under some conditions, such signals can result in higher levels of system-wide cooperation. 

Stephen Palley, Political Science, UCLA (spall@ucla.edu).

Steve is interested in exploring ethnic-based wars. To do so, he has developed an ethnic-warfare model that allows the exploration of biases against out groups, ethnic alliances, and power dynamics. Agents in the model have various types (modeled via binary strings) and are connected to one another in a fixed graph. Occasionally, agents decide if they want to go to war based on their likelihood of winning (using a Lanchester combat model), differences in ethnic identities, and potential ethnic alliances. Attacks and alliances influence future identities. Early results indicate the importance of feedback loops in the resulting dynamics and the potential for long-cycles of continuing conflict. 

Lynette Shaw, Sociology, University of Washington (shawl@u.washington.edu).

Lynette is interested in social construction: the general idea that agents construct the reality in which they take part. Specifically she is looking at the process of how mental representations interact with the environment to form such realities. To explore these ideas, she is using adaptive agents playing a coordination game, influenced by various environmental configurations. Even in cases of a random environment, she finds group formation, though the groups are more diverse and tend to have lower social conviction. In ongoing work, she is illuminating the conditions that cause developing cultures to mismatch their environment.

Milena Tsvetkova, Sociology, Cornell University (mvt9@cornell.edu).

Milena wants to understand the emergence of voluntary, contribution-based communities, ranging from social norms like opening doors to the help offered by anonymous participants in on-line forums. In prior experimental work she observed both generalized reciprocity and third-party imitation, the latter of which can result in cooperation easily dissolving over time. Here she combines models of contagion and thresholds of collective behavior assuming heterogeneity across the agents. She analyzes the impact of the "rivalness" of the good and can tie the likelihood of cooperative contagion to key parameters in the model, such as agent mobility and the ability to observe either recipients or contributors.

Alicia Uribe, Political Science, Washington University in St. Louis (aburibe@wustl.edu).

Alicia wants to further our understanding of the process of judicial appointments in the United States. She views this process as an assignment problem, whereby the President wants to allocate, a limited pool of judges, across the available openings based on the President's idealogical preferences tempered by the need to get the appointments approved by the Senate. Her results embrace a number of key features of this system, including predictions about the quality of judges, number of vacancies, and types of appointments, tied to the degree of ideological disagreement between the President and Senate. 

Jameson Watts, Management and Marketing, University of Arizona (jamesonw@email.arizona.edu).

Jameson is refining our understanding of industries characterized by rapid technological change. He models a technological space where firms compete with one another. Firms are able to search for new technologies adjacent to the ones that they already know. As knowledge bases begin to overlap, he explores various assumptions about how such competition is resolved in the market. Using this model, he answers questions about the importance of skill for firm success, how diversity impacts knowledge discovery, and the impact of variance across firm ability on market outcomes.

John H. Miller , miller@santafe.edu.