Thad Dunning, Professor

Closed (1) The destruction of historical archives and the consequences for inference in social science research

Check back for status

Why did some countries democratize earlier than others? Why did some governments incorporate indigenous populations into early state- and nation-building projects while others did not? The answers to these questions and many others depend on historical data. Yet, in many cases, access to historical data varies based on deliberate or chance events. Authoritarian leaders may destroy materials that are viewed as threatening to their survival, or natural disasters--such as floods, fires, and earthquakes--may eliminate certain historical records. This project evaluates how the destruction of select archives has shaped both the cases (e.g., time period, country) that social scientists study and thus, the conclusions they ultimately draw from their research.

The URAP for this project will work closely with the researchers to compile a database of existing social science research on historical instances of democratization in Latin America. The URAP will also assist in the collection of information on the destruction of archives, libraries, and government repositories in the region.

Day-to-day supervisor for this project: Anna Callis, Graduate Student

Qualifications: Existing coursework in history, political science, or applied economics, as well as knowledge of Excel. Advanced Spanish language ability (reading and writing) and prior experience with Hathitrust preferred.

Weekly Hours: 3-5 hrs

Off-Campus Research Site: Remote



Closed (3) Paving the way for the rise of outsiders: candidate and voter behavior in the era of political disillusionment?

Check back for status

Outsider candidates have been present in politics around the globe for decades. The most well-known outsider candidates who have won office share a populist style and have adopted extreme positions on policy issues, both on the right and on the left of the political spectrum (e.g., Hugo Chávez in Venezuela, Viktor Orbán in Hungary, Alberto Fujimori in Peru, among others, including Donald Trump in the U.S.). This empirical observation has, in turn, led the literature on outsider candidates to focus almost exclusively on describing and studying the consequences of anti-establishment/extremist candidates (e.g., Levitsky et al. 2016; Carreras 2012; Serra 2018). However, these studies have overlooked important variation between outsider candidates. Notably, the last presidential elections in France, Brazil, Colombia, Argentina, among other countries, all included moderate candidates running for office, who did not campaign on a populist style nor on extremist positions. This project seeks to fill in this gap by studying the strategic decisions adopted by candidates running as independents or with newly created parties (i.e., new entrants) in terms of where to locate along the policy dimension and which type of rhetoric to employ when appealing to voters.

URAP students will be asked to take part in one or more of the following three different parts of the project. First, they will help constructing an exhaustive database with a set of variables describing the profiles and contexts in which candidates running as independents or with newly created parties (i.e., new entrants) have risen in Latin America since the third wave of democratization. Second, students will help with the elaboration and implementation of an expert survey that will complement the database collected in the first part of the project. Finally, in the third part, students will be asked to collect and translate public speeches by outsider candidates running in Latin America, and to analyze them using topic modelling among other natural language procedures for text analysis. These three parts of the project will provide students with an overview of the different stages involved in quantitative research designs: e.g., definition of coding procedures, data collection, data cleaning, text and data analysis, among others.

Day-to-day supervisor for this project: Surili Sheth, Graduate Student

Qualifications: Given the tasks involved in all four parts of the project, students who are interested in applying should have some prior experience and familiarity in data-driven research. Some knowledge in statistical packages (either R or Stata), text analysis and machine learning is preferred but not required. Language skills: (any level of) Spanish or Portuguese is also a plus.

Weekly Hours: 6-8 hrs

Off-Campus Research Site: Remotely. Tasks will be done remotely (everything can be found online). Students are expected to attend virtual meetings that will take place every other week.

Closed (4) Understanding the link between self-help groups and women's political participation in Bihar

Check back for status

Does the presence of self-help groups increase women's political visibility? Because of the presence of quotas for local village council seats (panchayat seats), we wouldn't expect the number of women who win to increase -- as reserved seats are only for women, un-reserved seats are generally though of as "men's seats." However, we may expect to see some increases in the number of women running for reserved seats, or in the number of women voting, in areas where SHG presence is high. We have some datasets over three election cycles in Bihar with this information - it needs to be cleaned. We want to scrape data on SHG presence, clean it, and merge it with the elections data.

The undergraduate first and foremost be scraping data from a government website (after we talk through specifications of what the data should contain and what is on the government portal) and cleaning it. The student will also clean and merge this data with other datasets that I already have (elections data)., Ph.D. candidate

Qualifications: Able to write code to scrape data from government website (using whatever language the student is comfortable with, after we have discussed the specifications), data cleaning in Stata and analysis in Stata or R, fluency in Hindi. A computer science student who is interested in social policy projects in India, and who is fluent in Hindi, would be a good match for this project.

Weekly Hours: 6-8 hrs

Off-Campus Research Site: Online, but it is on Bihar / will be working with Bihar elections and government SHG program data