Textual data, data collection, and (computational) data analysis

Developing better measures of the concepts we hope to study is a recurring challenge of social science research. By offering access to new information sources, computational methods present new opportunities for testing theories and posing research questions. Our goal in this chapter is to not only provide an overview of text as data methods for social scientists, but also to promote informed use of these methods. We begin with a motivating example that illustrates the critical but too often overlooked importance of carefully assessing the validity and reliability of results produced from any text corpus. We then consider why text can be such a useful data source, starting with what information can be found in text and proceeding to consider four stages of a typical project. Throughout, our goal is to convey the many opportunities text as data methods affords for studying questions of interest to students of comparative politics.