Dr. Teddy Seidenfeld

Biography

Teddy Seidenfeld works on the foundations at the interface of philosophy and statistics, addressing problems involving multiple decision makers. One notable collaboration with M.J. Schervish and J.B. Kadane from Statistics at Carnegie Mellon University explores relaxing the norms of Bayesian theory to permit a unified standard for individuals acting as separate decision makers within a cooperative group. This work contrasts with the strict impossibilities posed by conventional Bayesian frameworks. Seidenfeld's collaboration with Larry Wasserman examines the short-run consequences of utilizing Bayes' rule to update a shared set of expert Bayesian opinions based on shared information, focusing particularly on anomalous cases termed 'dilation'. This phenomenon illustrates how new evidence can increase uncertainty regarding an event of common interest, indicating the extent of probabilistic disagreements among experts. His current collaborations with Kadane and Schervish include developing theories on the indexing degree of incoherence in non-Bayesian statistical decisions, representing coherent choice-functions using sets of probabilities, and investigating scoring rules for probabilistic forecasts. Seidenfeld has also contributed to the development of finitely additive expectations for unbounded random variables. His papers and recent presentations are concentrated within this evolving topic area.

Research Interests