Abstract:
|
One of the key foundations of frequentist statistics is statistical decision theory and associated concepts of procedure optimality. Brown was arguably the leading decision theorist of his generation and so had a profound impact on this part of the foundations of statistics. In this talk, however, I’ll focus on his less well known impact on foundations, through his work on conditional frequentist analysis and unification of frequentist and Bayesian methodology. One of his earliest papers was on the conditional level of Student’s t test and one of his latest was on post model selection inference, with many profound papers in between dealing with crucial issues involving conditioning, including how to develop conditional frequentist methods of testing that are identical to Bayesian methods. We will review some of the highlights of Browns work in these areas.
|