Skip to main content

Research

CSSS faculty and graduate students collaborate to develop cutting-edge statistical methodologies for the social sciences, and to advance social science research more generally. Current areas of research strength include statistical analysis of social networks, population projections, spatial statistics, and the statistical analysis of data generated by social media and other electronic data sources. Many faculty members adopt a Bayesian approach, which allows for quantification of uncertainty in statistical estimates. Another feature of CSSS research is its attention to the visual presentation of quantitative results. We are also committed to principles of reproducibility in scientific research.

Notable current projects include: the development of latent space and latent class models for social network analysis, which help us understand the structure and dynamics of phenomenon such as diverse as the spread of disease, social influence among students, and international trade and conflict; the development of probabilistic population projections for all countries in the world; extension of Bayesian hierarchical registration curve methods to analyze life course trajectories of deviant behavior, including marijuana use; and multiple projects utilizing data from mobile phones and Twitter to recover population characteristics, communication patterns, and migration flows.

New Bayesian Projections of World and Continental Populations. Building on methodology developed by CSSS Founding Director Adrian Raftery. Figure published in Science (346:234-237 (2014).)