On Tue, 22 Jul 2008 08:28:35 -0700, Turner, G. Marc wrote:
>Mike Palij wrote:
>> (e.g., what is a nonparametric t-test?)
>
>It actually seems to be "common" in biostats and, from what little I
>know, is an alternative name for the Fisher-Pitman permutation test.

I don't believe that they were referring to the Fisher-Pitman permutation
test, instead, the reference that they give is for an article that uses
generalized estimating equations (GEE). If you have Jstor access see:

*  GEE with Gaussian Estimation of the Correlations When Data Are Incomplete
* Stuart R. Lipsitz, Geert Molenberghs, Garrett M. Fitzmaurice, Joseph
Ibrahim
* Biometrics, Vol. 56, No. 2 (Jun., 2000), pp. 528-536
* Publisher: International Biometric Society
* Stable URL: http://www.jstor.org/stable/2676997

The approach described in this paper I believe would allow the original
researchers to use all of the cases even if they had dropped out by
posttest.
I don't think that the Fisher-Pittman test allows for missing data.
Moreover,
I believe that the Fisher-Pittman test is sensitive to inequality of means
and
of variances, so one would have to follow a specific procedure to seperate
tests of means from tests of variances.

-Mike Palij
New York University
[EMAIL PROTECTED]



---
To make changes to your subscription contact:

Bill Southerly ([EMAIL PROTECTED])

Reply via email to