Methoden der Vergleichenden Regierungslehre I
durchgeführt von Dominic Ponattu
Dienstags 17:15 - 18:45 Uhr in B 6, 23-25, Raum A 204
Startdatum: 6.9.2016; Enddatum: 6.12.2016
A solid research design is paramount to carefully testing theories. In particular, modern political science requires a compelling empirical strategy, capable of identifying causal effects (and thus go beyond mere correlation). This course starts off with a brief review of theory building and how to derive testable hypotheses. We will then focus on the key element of this course -- which is on how to apply empirical methods for causal inference. To this end, we will first recap OLS, how to interpret this estimator and understand its assumptions. We will then move on to study empirical strategies that have become increasingly important in modern political science. These include: fixed effects estimation, difference-in-differences, matching, instrumental variables, regression discontinuity designs and quantile regression. Our goal is to understand how to implement these methods and under what circumstances we would choose one over the other. To do so, we will discuss specific research questions and which methods of causal inference are most appropriate to examine this particular question. Moreover, we will implement these methods using STATA. Students wishing to enroll in this course are expected to have acquired some basic knowledge of both the OLS model and STATA.