Using Text as Data Methods to Discover, Measure, and Explain
Text as data methods are increasingly used in the social sciences to explore large scale collections of text. This talk draws on my recent papers to show the distinct social science tasks that text as data methods can accomplish and provides a framework for evaluating those methods. Using an example from the study of Congressional communication I show how text as data methods can help us to understand the connection between elected officials and constituents. And an example survey experiment shows how text can be used to understand constituents' decisions in a democracy.
Justin Grimmer is an Associate Professor of Political Science at Stanford University and will be joining the University of Chicago's Department of Political Science this fall. His research examines how representation occurs in American politics using new statistical methods and develops new tools for the analysis of large text collections. He is the author of two books,Representational Style in Congress: What Legislators Say and Why It Matters (Cambridge University Press, 2013) which was awarded the Richard J. Fenno Prize from the legislative studies section and The Impression of Influence: How Legislator Communication and Government Spending Cultivate a Personal Vote (Princeton University Press, with Sean J. Westwood and Solomon Messing). His work has appeared in the American Political Science Review, American Journal of Political Science,Journal of Politics, Political Analysis, Proceedings of the National Academy of Sciences, Regulation and Governance, Proceedings of the Association for Computational Linguistics, and other outlets. He is also a Research Consultant on Facebook's Data Science and Marketing Science teams.