|
Activity Number:
|
232
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #307821 |
|
Title:
|
Synthesis of Bayes Factors and Estimation for Large Samples: Adaptive Hypothesis Testing Intervals (AHTI)
|
|
Author(s):
|
Luis R. Pericchi*+ and Maria E. Perez
|
|
Companies:
|
University of Puerto Rico, San Juan and University of Puerto Rico, San Juan
|
|
Address:
|
P.O. Box 23355, San Juan, PR, 00931-3355,
|
|
Keywords:
|
Adaptive Hypothesis Testing Intervals(ahti) ; Synthesis ; Bayes Factors ; Estimation ; Hypothesis Testing ; Objective Bayes
|
|
Abstract:
|
As a first Synthesis, we propose a calibration of p-values and significance levels that convey specific guidelines in how to diminish the alpha-levels as the sample sizes grows, to alleviate the discrepancy of Bayes Factors and Classical Testing (assuming it is accepted that type I error is not fixed but can diminish as evidence accumulates). This decrease in alpha-levels is appropriate for testing with huge sample sizes. There is a second Synthesis involved very important to unify Bayesian model selection procedures, the Synthesis between Probability Intervals and Bayes Factors: A pervasive idea in the whole of Statistics is to reject null hypotheses when point nulls are outside intervals. This is wrong (from a Bayesian point of view) when confidence levels are held fixed.
|
