Graeme Blair bio photo

Assistant Professor
Political Science

      E-mail     Google Scholar       Twitter

Selected Research Projects

On the Geography of Assets and Citizens: How Proximity to Oil Production Shapes Political Order

I propose in this book project that physical proximity of assets to people often enables local groups to interrupt production and tax collection from those assets through protests, sabotage, and strikes. The state is then forced to bargain with the groups to avoid an interruption to state revenues, and to offer concessions such as a share of revenues derived from local assets. When bargaining fails, violent conflict often results. I test this theory using a multi-method research design based on novel, fine-grained data on the geography of oil exploration and production and civil wars around the world, case evidence from oil production interruptions in Nigeria from 2006 to 2009, and a large original survey of civilians in the Niger Delta oil region in Nigeria on how civilians living near oil fields collaborated with the armed groups to interrupt oil production.

Writing: book manuscript in preparation

Funding: International Growth Centre, National Science Foundation

Motivating Community-Minded Behaviors

Joint work with Rebecca Littman (Yale) and Betsy Levy Paluck (Princeton).

Social scientists have long sought to explain why people donate resources for the good of a community. Existing explanations do not fully account for why people adopt novel community-minded behaviors, which may carry special costs and risks. In a field experiment in Nigeria, we tested two campaigns that encouraged people to try reporting corruption by text message. Psychological theories about how to shift perceived norms and how to channel behavioral momentum drove the design of each campaign. The first, a film featuring actors reporting corruption and the second, a mass text message reducing the effort required to report, caused a total of 1,181 people in 106 communities to text, including 241 people who sent concrete corruption reports. Psychological theories of social norms and behavior change can illuminate the early stages of the evolution of cooperation and collective action.

Where: Akwa Ibom, Bayelsa, Delta, and Rivers states in the Niger Delta region, Nigeria

Partners: Integrity Nigeria, MTN Nigeria, iROKO TV, Magic Movies

Writing: working paper under review

Does Community Policing Build Trust in the State and Reduce Crime?

I lead a six-country study (“metaketa”) to test a common intervention that implement community policing program consisting of two components: (1) a community engagement program to solicit information on community problems from citizens and transmit information about police programs to citizens; and (2) a problem-oriented policing program, in which police address problems identified through community engagement programs directly with small, dedicated budgets or indirectly with the assistance of other public and private agencies.

For more, see EGAP Metaketa IV: Community Policing.

Where: Santa Caterina, Brazil; Medellin, Colombia; Monrovia, Liberia; nationwide, Uganda; Sarghoda Division, Pakistan; Sorgoson, Philippines

Funding: U.K. Department for International Development

Methods for Studying Sensitive Topics in Surveys

Joint work with Kosuke Imai (Princeton), Jay Lyall (Yale), and Yang-Yang Zhou (Princeton).

The study of public opinion fundamentally depends on honest responses to survey questions. If a survey respondent feels pressured to respond in ways that please the interviewer or protect her own safety, social scientists will not be able to identify the true population distribution of preferences. If that respondent feels great pressure, she may not respond at all, posing an additional challenge to inference. These issues affect our ability to learn about issues from support for insurgent groups to voter turnout.

Three survey question methods – list experiments, endorsement experiments, and the randomized response technique – aim to address this problem by obscuring the truthful response of individuals and thus offer privacy protection to respondents that may encourage honest responses and lower non-response rates. Unlike with direct questions, researchers cannot study the responses of individuals, but each technique remains useful for identifying estimates of the population distribution of responses.

For more, see our project site,


  • Political Analysis (Blair and Imai 2012) describing regression and list experiment diagnostics for several common designs. pdf
  • Journal of the American Statistical Association (Blair, Imai, and Zhou 2015) describing many randomized response designs, power calculations, and regression methods. pdf
  • American Journal of Political Science (Blair, Imai, and Lyall 2014), we present new graphical and statistical methods for comparing and combining dat from the two types of experiments with empirical applications from survey in Afghanistan. pdf


  • list package for R for analyzing list experiments
  • rr package for R for analyzing randomized response technique items

Improving Research Designs in the Social Sciences

Joint work with Jasper Cooper (Columbia), Alexander Coppock (Yale), and Macartan Humphreys (Columbia).

The evaluation of research depends on assessments of the quality of underlying research designs. Surprisingly, however, there is no standard definition for what a design is. We provide a framework for formally characterizing the analytically relevant features of a research design. The approach to design declaration we describe requires defining population structures, a potential outcomes function, a sampling strategy, an assignment strategy, estimands, and an estimation strategy. Given a formal declaration of a design in code, Monte Carlo techniques can then be easily applied to a design in order to diagnose properties, such as power, bias, expected mean squared error, external validity with respect to some population, and other “diagnosands.” Declaring a design in computer code lays researchers’ assumptions bare and allows for clear communication with funders, journal editors, reviewers, and readers. Ex ante design declarations can be used to improve designs and facilitate preregistration, analysis, and ex post reconciliation of intended and actual analyses. Design declaration is also useful ex post however and can be used to describe and share designs as well as to facilitate reanalysis and critique. We provide an open-source software package, DeclareDesign, to implement the proposed approach.

For more, see our project site,

Funding: Laura and John Arnold Foundation and seed funding from Evidence in Governance and Politics

Writing: working paper


  • DeclareDesign for declaring and diagnosing the properties of research designs (for R)
  • fabricatr for imagining your data before you collect it (for R, Stata forthcoming)
  • estimatr for fast estimators for social scientists (for R)