Past Events

Lucas Stanczyk joined the MIT faculty in the spring of 2012 as Assistant Professor of Political Science, after finishing his Ph.D. in Harvard’s Government Department. A political theorist by training, his research ranges across a variety of topics in political philosophy, constitutional and legal theory, and the history of moral and political thought.

His doctoral dissertation investigates the implications of the value of social justice for the organization of economic production in contemporary societies. He is also at work on a history of the modern concept of distributive justice and ideas of redistribution.

Abstract: This paper studies the political effects of a particular type of information intervention: one that improves voter expectations of government capacity. I argue that if citizens systematically underestimate what their government is capable of, then voting on the basis of performance is of little consequence, and politicians in turn have little incentive to perform well. I report evidence from a randomized field experiment in Mali that tests whether improving voter information about the scope of government effects voting behavior. A civic education course was provided to 370 villages in 64 randomly assigned municipalities dispensing information on the responsibilities of local government and the basics of democratic accountability to all treated villages, with an additional component on relative government performance to half of treated villages. A survey was then conducted in the 64 treated and 31 control municipalities. Voting simulations show that people in treated villages are more likely to vote based on performance: a poor-performing candidate had to pay more to buy votes of citizens in treated communities, and the votes of citizens in control communities were more easily swayed by dimensions such as kinship or gift-giving. Suggestive evidence points to two possible mechanisms underlying this behavior: treatment raised expectations of local government and improved coordination among voters. A behavioral outcome measure – the likelihood that villagers challenge local leaders at a town hall meeting – confirms the positive treatment effects found in the survey measures.

Professor Keele is an Associate Professor of Political Science at  Penn State University.  He pursues research and teaching interests in American politics (public opinion and elections) and in methodology (causal inference, design-based inference, and nonparametric statistics).  He has published articles in the Journal of Politics, American Journal of Political Science, and American Political Science Review.  His research seeks to understand how political institutions can either increase or decrease electoral participation.  Current his research focuses on how research design can affect inferences about voter turnout.  Other projects seek to understand how variation in geography can be used to draw causal inferences.

Abstract: We explore how variation in geography can be exploited as a natural experiment. We review the assumptions that must hold for geographic natural experiments to yield valid causal estimates. In particular, we focus on cases where a geographic or administrative boundary splits units into treated and control areas. Here, one suspects that units either sort around the boundary with error or the boundary is drawn with some amount of error. We derive the identifi cation assumptions behind this geographic natural experiment, and show that it redefi nes the causal estimand in a counterintuitive fashion. We also suggest some testable implications for the key identifi cation assumption, and develop an estimation framework that is faithful to the inherently spatial qualities of the design. We illustrate best practice with geographic natural experiments through an application on whether ballot initiatives increase turnout in Wisconsin and Ohio. The applications illustrate the strengths and weaknesses of causal inferences based on geographic natural experiments.