This website was developed for a June 2010 NSF sponsored workshop to introduce political scientists to text analysis. The workshop was a success and we hope to find funding to hold another one in the near future. John Wilkerson and Phil Schrodt
Welcome to the Tools for Text Workshop, made possible with support from the National Science Foundation (SBE), the Society for Political Methodology, and the Center for American Politics and Public Policy at the University of Washington.
We will cover a range of approaches that address unique annotation objectives.
- Data Prep – joint presentation
- Manual Annotation – Stuart Shulman, University of Massachusetts, Amherst
- Unsupervised Learning – Brandon Stewart, Harvard University
- Supervised Learning – Stephen Purpura, Cornell University and John Wilkerson, UW
- Dimensional Scaling – Will Lowe, University of Maastricht
- Behavior and Prediction – Phil Schrodt, Penn State
We assume that most participants possess limited background knowledge about text annotation methods but have specific reasons for wanting to know more about them. The first day of the workshop will be devoted to presentations that introduce the topics. Dress comfortably – it will be a long day!
The second day of the workshop is be devoted to computer lab work where groups of about 15 will rotate through 5 training different training sessions. To minimize delays due to network and configuration issues, we will be using the PCs in the lab (rather than personal laptops).
One important take away from the conference is that none of these methods offers a silver bullet. Each has its particular strengths and limitations. In addition, it may quickly become clear that some of the methods are not applicable to your current project. However, there is no doubt that these methods can open up a new world of research possibilities, and in this respect we hope that the workshop will have a detectable impact on your longer term research agenda.
There will be legitimate differences of opinion among the presenters about these methods. We will do our best to identify the sources of these disagreements but ultimately it is up to participants to decide which considerations are most relevant to their own research objectives.
We are pleased that you are coming to Seattle and look forward to a hopefully stimulating, and certainly fast-paced workshop!
Phil Schrodt John Wilkerson