Is it possible to elicit the ideological positions of voters by having information about the structure of their social media networks? Can we make predictions about upcoming policy changes by analyzing the speeches of statesmen, although they could be considered cheap talk? Is there a way to analyze all court decisions of a jurisdiction in order to identify individual biases of judges, thereby suggesting a way how to make the legal system more impartial? Or can we develop a reliable index of organized crime and subversion in industrial areas, typical hotbeds of such crimes, taking into account a wide range of Internet, social media and administrative data sources? These questions are within the domain of Institutional and Organizational Economics (IOE). And all of them could not be seriously studied, let alone answered, by traditional empirical methods. Data science, a new toolkit combining statistics with computer science, is changing this.
Together with Patricia Prüfer, I have written a brief introductory essay that will be published in a handbook, “A Research Agenda for New Institutional Economics,” edited by Claude Menard and Mary Shirley. We describe the most prominent data science techniques that lend themselves to analyses of the governance structures of institutions and organizations. Several examples using data science to analyze legal, political, and social institutions are introduced. Then we sketch how specific data science techniques can be used to study important research questions that could not (to the same extent) be studied without these techniques. We conclude by comparing the main strengths and limitations of computational social science with traditional empirical research methods and its relation to theory. All this is amended by links to literature and Internet resources and to the most relevant text mining tools and download sources, showing how to get started with data science methods independently.