Institute for Mathematical Behavioral Sciences

Reviewing the literature in many scientific fields these days reveals significant interest in Bayesian statistical methods. Such methods make inferences from data using probability models for quantities we can observe (i.e., the data) and for quantities about which we wish to learn (e.g., parameters, missing data). The greatest advantages of the Bayesian approach accrue in problems with large numbers of unobservables (for example, image analysis or observational studies with missing data). There has been less interest in Bayesian methods areas of science like experimental cognitive science where data analysis is characterized by traditional methods like the analysis of variance. This talk considers roles that Bayesian methods can play in such settings.