2009/Statistics

From SIPB Cluedumps

[edit] Statistics and the Non-Conflict between Bayesians and Frequentists

Date: October 13, 2009, at 4:30 PM

Presenters: Keith Winstein (keithw)

Location: 4-231

Notes: Statistics and the Non-Conflict between Bayesians and Frequentists (slides)

Abstract: I'll go over the building blocks of statistics and why you often hear about the conflict between "Bayesians" and "frequentists." Focusing on simple examples, I'll explain each camp and why I think they aren't really in disagreement. Hopefully we can get the audience shouting about this non-conflict. I'll also talk about some of my work on measuring the performance of confidence intervals and p-values, how you can make $800 million because of a lousy approximation, and "Why Most Published Research Findings Are False."

Bio: Keith Winstein '03 is an associate member of SIPB, on leave from his Ph.D. at MIT. He works as a reporter at The Wall Street Journal in Boston, covering science and medicine.