|
Activity Number:
|
3
|
|
Type:
|
Invited
|
|
Date/Time:
|
Sunday, July 29, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #308069 |
|
Title:
|
On Criteria for a Measure of Statistical Evidence in Clinical Trials: What We Want and What We Don't
|
|
Author(s):
|
Jeffrey D. Blume*+
|
|
Companies:
|
Brown University
|
|
Address:
|
Center for Statistical Sciences, Providence, RI, 02818,
|
|
Keywords:
|
Statistical Evidence ; Clinical Trials
|
|
Abstract:
|
It is important to assess the strength of statistical evidence in a clinical trial. But how statisticians choose to communicate this varies by philosophy and training. Some argue that this is communicated by the p-value, some argue for the posterior probability and some for Likelihood ratios. We propose three evidential quantities that every clinical trial should present: (1) the probability that the study design will generate misleading evidence, (2) the strength of the evidence in the observed data, and (3) the probability that the observed evidence is misleading. We'll argue that no matter what philosophical approach one prefers, only #2 and #3 are relevant at the end of the study and that #1 is always critical in the design stage. We'll explore if current paradigms identify these quantities and why this framework resolves problems of multiple comparisons and multiple looks.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
