I am pleased to have the opportunity to welcome you to the Center for Applied Statistics website. This is a chance to tell a wide audience what the Center for Applied Statistics does here at Washington University.
Let me emphasize the importance of statistics to an academic institution. Statistics is literally the lingua franca of twenty-first-century academic research. It is essential to nearly every field that uses data, from cinematography to history, from astronomy to physics, as well as everything in between. Statistics is now a common language across the humanities, social sciences, and natural sciences, and therefore no great university can afford to be without a
strong research and teaching component in modern statistics.
While we are based in Arts & Sciences, we work with University scholars in the Brown School of Social Work, the Medical School, and the Olin School of Business. Some examples of ongoing applied work that faculty housed in the center or associated with the center are conducting include:
- Creating dynamic models of why and how democracies are internally violent/ repressive;
- Improving methods designed to identify test items that are biased against persons in particular groups (e.g., women, ethnic minority groups);
- Figuring out why U.S. cabinet-level political leaders stay on the job for only two years on average;
- Applying circular data models to estimating violent attacks in Iraq;
- Understanding the causes for Chinese civil wars in 2600 B.C.;
- Estimating fan support for the designated hitter rule in baseball;
- Judging the efficacy of AIDS assistance programs in Africa; and
- Comparing Alfred Hitchcock’s camera style to other directors.
Examples of theoretical work include:
- Improved algorithms for Bayesian nonparametric estimation of random effects;
- New methods to identify cheaters (and other types of aberrant responders) based on their pattern of responses to test questions;
- Hypothesis tests incorporated directly in the specification of competing prior distributions;
- Nonparametric clustering algorithms with mild assumptions;
- Bayesian lassos for model specification;
- Path dependence as an empirical Markov process; and
- Missing data imputation for categorical variables.
These are only a sample of current activities, but they provide an informative view of our research at the moment.
Given the level of importance for statistics on campus, what role does the Center for Applied Statistics play? As the only unit exclusively dedicated to statistics, we are the intellectual home for statisticians at Washington University. We provide a central point for quantitative scholars— our associated faculty and students cover nearly every discipline at the University— to collect and interact. Our research has appeared in Journal of the American Statistical Association, Multivariate Behavioral Research, Psychometrika, Journal of the Royal Statistical Society, Journal of Statistical Software, Journal of Computational Graphics and Statistics, and other methodological outlets. We also host an annual conference every April.
Our faculty—including four jointly appointed faculty with the Department of Psychology, the Department of Political Science, and the Law School—are currently supported by five National Science Foundation grants as well as other sources of funding. In our original plan from 2007, we are scheduled to double in size. We’re joined every year by two “distinguished statisticians in residence.” The visiting statisticians are top scholars in statistics who spend the first and last week of the semester here at Washington University giving lectures and setting up an agenda for the time in between with a local faculty member and graduate students. These are highly topical and very technical offerings.
Our regular curricular offerings include undergraduate statistics courses and graduate-level courses such as Categorical Data Analysis, Factor Analysis, Multilevel Models, Panel Data Models, Time Series Analysis, and Generalized Linear Models, as well as other methodological courses cross-listed with departments.
Overall, statistical research is an exciting area at the University, and there is a large demand worldwide for statisticians in government, industry, and academia. Furthermore, statistics is the ultimate interdisciplinary field, allowing trained methodologists to work with a wide variety of coauthors and colleagues. I hope that this overview gives some insight to our activities and mission.
Jeff Gill, Professor of Political Science
Director, Center for Applied Statistics
Vice President, Society for Political Methodology