I am involved in the following projects.
Wordscores' uneven success in applications has generated some scepticism. Lowe (2008) demonstrated bias in the method but also identified the situations where its underlying assumptions are reasonable approximations to a statistical ideal point model. One assumption that is not addressed in this analysis, shared by other political text scaling methods, is that an actor's expressed position directly expresses her preferences and is unaffected by strategic considerations. I am currently working on a reformulation as a hierarchical measurement model. Allowing predictors at the level of ideal points will allow the model to be better integrated into strategic models of political language
Collaborators: Ken Benoit and Michael Laver
Based on close analysis of a small number of speeches the former president of Taiwan Chen Shui-bian has been characterised as being both vague and increasingly provocative on the issue of Taiwan independence during his presidency. This work tries to place Chen in a broader account of how a rational leader should vary the level of rhetoric in speeches depending on external events, his own agenda, and audience preferences in an environment of incomplete information. We test this account using computerized content analysis to captures the rhetorical themes of Taiwan independence, Taiwanese identity and democratic reforms.
Collaborators: Jon Sullivan
From the project page:
"This dataset contains information about organised armed groups that are identified by documentary, media, and other sources to be pro-government but are not part of the state's regular security forces. The information covers the period from 1981 to 2007 across all countries, and is not limited to ongoing conflicts. The main unit of observation is the pro-government armed group. The dataset is part of a wider project funded by the Economic and Social Research Council (ESRC). The associated research will focus on the ways in which governments use these informal groups or militias, their origins and termination. The dataset will allow researchers to more accurately study how these groups affect the outbreak, duration and nature of conflicts and the likelihood of a successful peace agreement following conflict."
Collaborators: Neil Mitchell and Sabine Carey
A database of implementations of the Rome statute that establishes the International Criminal Court. The database connects political and legal information about states with detailed analysis and legal commentary on national implementing legislation. The project is a collaboration with Nottingham's Human Rights Law Centre and the ICC via its Legal Tools Partners program. A subset of this data is now publically available.
Collaborators: Olympia Bekou and Emilie Hunter