Dear Allstat Members,
I was asked to provide method/reference for calculating
the pooled mean difference between two proportions
and its confidence interval , when there
are two groups (e.g. two treatments), and the outcome is binomial
(e.g. success vs failure), and when the
data were collected from more than one centre.
I know that there are methods for odds ratio or for risk
ratio, but I am not aware of any method for risk difference.
I appreciate any suggestion.
Thank you.
Akos Pap
statistician
Bayer Vital GmbH, Leverkusen
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|