I have been a member of the political science faculty at UNC-Chapel Hill since January 2010. My academic training has been eclectic. I earned my AB in Latin American Studies from Brown University in 1995. After a number of years serving food and drink in Boston, I went back to school to study math and statistics, eventually obtaining an MS in Mathematics from Salem State College (2004), and later an MS in Statistics (2006) and MPhil in Public Policy and Management (2007) from Carnegie Mellon University. In May, 2010, I was awarded a joint PhD in Statistics and Public Policy from Carnegie Mellon. In my dissertation, entitled “Cues and Heuristics on Capitol Hill: Relational Decision-Making in the United States Senate,” I applied tools of social network analysis and collaborative filtering to model cosponsorship and roll-call voting behaviors.
As a major focus of both my teaching and research, I seek to help fellow political scientists become more thoughtful users of statistical methods by promoting statistical reasoning rather than succumbing to a rigid “cookbook” approach to inference. My teaching interests include research design, applied and mathematical statistics, social network analysis, survey sampling, and latent structure analysis. In recent projects, I (1) propose an integrated approach to substantive and statistical significance, (2) demonstrate the advantages of latent class and mixed membership models in using survey data to explore political beliefs and values, ideology, and identity, and (3) analyze textual content and network behavior by political bloggers and users of Twitter to learn about how opinions and reactions to policy issues and events evolve and spread. Please see my current C.V. to see how my research program is progressing.
As part of a broader research agenda on political communication in social and conventional media, I am supported, along with my co-PIs, Noah Smith (Computer Science, Language Technologies Institute & Machine Learning, CMU), Philip Resnik (UMIACS-UMaryland), and Amber Boydstun (Political Science, UC-Davis), by a generous Social-Computational Systems (SoCS) grant from the National Science Foundation (Sept. 2012-Aug. 2015). Internal seed grants from the Odum Institute and the College of Arts & Sciences at UNC have supported my related work with members of the newly formed Social Network Analysis at Carolina (SNAC). Members of this group are currently developing and implementing community detection algorithms for studying networked political blogs. For more information on this project as it develops, see our webpage.
I also serve as chief statistician for Latino Decisions (http://www.latinodecisions.com/), a political polling firm surveying and analyzing Latino public opinion. We recently designed and released an interactive online tool for examining the potential influence of Latino voters in the upcoming U.S. presidential election: http://www.latinovotemap.org/map/. I discussed this tool and the possibility that immigration issues may have electoral consequences, as a member of the panel “Latino Vote Matters: Immigration, Power, and an Interactive Look at the Map,” which may be viewed at http://goo.gl/16b7t. I have also developed models for predicting Latino turnout and vote choice, described briefly here: