Algorithmic Text Analysis

Abstract

Abstract Algorithmic analysis of “text as data” has proliferated throughout academic research over the past decade, offering the opportunity to examine ever-larger and more diverse collections of documents, address more complex issues, and incorporate underrepresented voices into social science inquiry. Drawing from a new survey of twelve top political science journals, this chapter discusses how current methods address a multitude of questions—pointing to what their structural assumptions and fundamental characteristics imply about their insight into latent constructs and meanings. Particular consideration is paid to the mismatch between the objective orientation of computational methods and subjective nature of social scientists’ contextual, situated inquiry. Finally, the chapter looks into areas of potential expansion that would allow these algorithmic methods to enhance concept and discourse analysis, further advancing an interpretive and inclusive understanding of political communication.