Re: Bonferroni correction

tennisfan wrote: 
Thanks for your reply.

Basically I'm running 4 separate regressions with the 4 DV being the
total score and the 3 subscales. Yes these 3 subscales make up the
total with the exception of 4 items.

My predictors are my grouping variable and a few other covariates. I am
not putting the subscales and total score into the equation together.
Looking at t-tests the total score and subscales are all significant.
The range of correlations between the total score and subscales is
.33-.88 (p<.001).

I hope this clarifies my analysis. Any advice on whether to correct my
p-value for these 4 separate analyses would be appreciated--or if it
would be too conservative.

Thanks,
Mary


Here are a couple suggestions from a paper in the "Basic Statistics for Clinicians" series in CMAJ. (http://collection.nlc-bnc.ca/100/201/300/cdn_medical_association/cmaj/series/stats.htm)

"We can also specify, before the study is undertaken, a single primary outcome on which the main conclusions will hinge. A third approach is to derive a global test statistic that combines the multiple outcomes in a single measure. Full discussion of these strategies for dealing with multiple outcomes is beyond the scope of this article but is available elsewhere [18]."

Reference 18 is:

Po*** SJ, Geller NL, Tsiatis AA. The analysis of multiple endpoints in clinical trials. Biometrics 1987;43:487-98.

--
Bruce Weaver
bweaver@xxxxxxxxxxxx
www.angelfire.com/wv/bwhomedir
.

Quantcast