|
Activity Number:
|
379
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #310272 |
|
Title:
|
Permutation Test-Based Significance Levels for Repeated Measures
|
|
Author(s):
|
Katie Spencer*+ and Daniel Reyner
|
|
Companies:
|
Zimmer, Inc. and Zimmer, Inc
|
|
Address:
|
PO Box 508, Warsaw, IN, 46581-0708,
|
|
Keywords:
|
repeated measures ; distribution free ; permutation test ; model assumptions ; clinical trial ; 2 group comparison
|
|
Abstract:
|
Repeated measures analyses are common to clinical trials which seek regulatory approval. However, violation of model assumptions related to the distribution of residual error may require abandonment of the statistical model. This paper provides an example of the use of permutation tests to generate measures of statistical significance in repeated measures analyses. This method allows for a prespecification of the statistical model which will be used in final analysis, achieving estimates of statistical significance which are distribution-free. The example focuses on comparisons of repeated measures between two groups.
|
