Yale University
Department of Statistics and Biostatistics

Bayesian model-building by pure thought: some principles and examples


Monday, November 4, 1996

Andrew Gelman
Department of Statistics
Columbia University

Seminar to be held in Room 309, LEPH, 60 College



In applications, statistical models are often restricted to what produces reasonable estimates based on the data at hand. In many cases, however, models can be restricted based on theoretical arguments alone, in the absence of any data and with minimal applied context. We illustrate with theoretical examples from spatial statistics and time series. These theoretical arguments may have important practical consequences when working with models of open-ended complexity. We discuss briefly the implications in applied contexts including toxicokinetics, election forecasting, and opinion polling.