23–25 June June 2021
Linking Theory and Text as Data is a course that synthesizes skills learned by students in the previous three modules (game theory, statistics, and QTA). Focusing mainly (though not entirely) on papers using non-cooperative game theory, this module pairs theoretic modules of information with quantitative textual data. This class is especially useful for students trying to write papers on and complete research projects using the framework of EITM when their empirical applications use textual data.
Over the three days, we will review a number of theoretical models including signaling models with one (or more) experts, models of delegation, and bargaining models that are useful for motiving empirical applications with text. We will also review research on strategic selection, strategic ambiguity, and audience effects. In the framework of EITM, we will also discuss measurement construction from theoretically motivated model parameters, and cover topics related to measurement reliability and validity.
Students will also be exposed to a number of cutting-edge, substantive applications in the literature that use a Text as Data approach, from a variety of disciplines including political science, health, economics and finance, and computer science. Problems (and solutions) to using multilingual texts are briefly considered. Finally, students will also be exposed to new research linking textual analysis and causal inference using survey experiments.
Nicole Baerg is a lecturer at the Department of Government at the University of Essex.