Christopher Lucas, Professor Dr. Richard Nielsen, Margaret E. Roberts, Brandon M. Stewart, Dr. Alex Storer, and Professor Dr. Dustin Tingley presented a paper entitled Computer assisted text analysis for comparative politics, 26 April 2014 at the New Faces in Political Methodology conference, sponsored by the Penn State Quantitative Social Science Initiative, and held at Penn State University in State College, Pennsylvania, USA.
Here is the abstract:
Comparative politics scholars are well poised to take advantage of recent advances in research designs and research tools for the systematic analysis of textual data. This paper provides the first focused discussion of these advances for scholars of comparative politics, though many arguments are applicable across political science subfields. With the explosion of textual data in countries around the world, it is important for comparativists to stay at the cutting edge. We situate recent and existing tools within a broader framework of methods to process, manage, and analyze textual data. While we review a variety of analysis tools of interest, we particularly focus on methods that take into account information about when and who generated a particular piece of text. We also engage with more pragmatic considerations about the ability to process large volumes of text that come in multiple languages. All of our discussions are illustrated with existing, and several new, software implementations.
The paper focuses particularly on Structural Topic Model, a form of unsupervised topic modeling.
The paper describes a number of studies involving automated analysis of legal texts, and provides an extended example involving analysis of Islamic legal texts (fatwas).