2014 Modern Modeling Methods (M3) Conference
Pre-Conference Workshop: May 19, 9:00-5:30 pm
Bayesian Methods for the Social and Behavioral Sciences
Dr. David Kaplan
Dr. Kaplan is Professor of Quantitative Methods and Chair of the Department of Educational Psychology at the University of Wisconsin – Madison, and holds an affiliate appointment in the Department of Population Health Sciences. Dr. Kaplan’s current research focuses on the development of Bayesian methods applied to a wide range of education research settings. His specific interests include: Bayesian model averaging; objective versus subjective Bayesian modeling; Bayesian posterior predictive causal inference; and Bayesian approaches to problems in large-scale survey methodology. Dr. Kaplan’s collaborative research involves applications of advanced quantitative methodologies to substantive and methodological problems in international comparative education. He is most actively involved in the OECD Program for International Student Assessment (PISA) where he has served on its Technical Advisory Group and currently serves as Chair of its Questionnaire Expert Group. He also sits on the NAEP Questionnaire Standing Committee. Dr. Kaplan is a Fellow of the American Psychological Association (Division 5) and was a Jeanne Griffith Fellow at the National Center for Educational Statistics. Dr. Kaplan received his Ph.D. in education from UCLA in 1987.
Bayesian statistics has long been overlooked in the quantitative methods training for social and behavioral scientists. Typically, the only introduction a student might have to Bayesian ideas is a brief overview of Bayes’ theorem while studying probability in an introductory statistics class. This is not surprising. First, until recently, it was not feasible to conduct statistical modeling from a Bayesian perspective because of its complexity and lack of available software. Second, Bayesian statistics represents a powerful alternative to frequentist (classical) statistics, and is therefore controversial. Recently, however, there has been great interest in the application of Bayesian statistical methods, mostly due to the availability of powerful (and free) statistical software tools that make it possible to estimate simple or complex models from a Bayesian perspective. The orientation of this workshop is to introduce practicing social and behavioral scientists to the basic elements of Bayesian statistics and to show through discussion and practice why the Bayesian perspective provides a powerful alternative to the frequentist perspective. It is assumed that workshop attendees will have a background in basic statistical methods up to, and including, regression analysis. Some exposure to multilevel modeling and factor analysis is desirable.
Morning topics include:
1. Major differences between Bayesian and frequentist paradigms of statistics, with particular focus on how uncertainty is characterized;
2. Bayes’ theorem;
3. Bayesian model building and model evaluation;
4. Bayesian computation.
Afternoon topics include:
1. Bayesian analyses using R;
2. An example of Bayesian regression;
3. An example of Bayesian factor analysis;
4. An example of Bayesian hierarchical linear modeling;
5. Wrap-up: Relative advantages of the Bayesian perspective.