In the past decade, quantitative text analysis has established itself as a frequently used method in political science to study political actors and processes. Typical research questions include the identification of concepts and topics, text classification, and the measurement of latent policy positions. While quantitative text analysis reduces the costs of analysing texts, and while various user friendly open-source libraries have been developed recently, several challenges remain. In this presentation, I will describe popular text-as-data methods from the perspective of implementation in R, one of the most commonly used statistical programming software in the social sciences. Afterward, I will turn to the challenges that political scientists are currently facing when analysing textual data. I will also discuss how computer scientists could assist social scientists in solving some of these problems. Finally, I outline potential future developments of quantitative text analysis.
23.05.2019 - 10:15