R help - Nov 2009 - Sample size calculation for paired proportion testing?

I would appreciate any help on this problem. I need to perform a sample size
analysis for a study comparing the performances of 2 different methods of
diagnostic classification.  Assume that method 1 has an accuracy of p1
against known truths (a reference classification, as multiple categories),
and method 2 has an accuracy of p2. A study is being planned in which N
subjects will be independently classified by these two methods, and I need
to calculate how many subjects N are need to determine if p2 > p1 for a
given power. I’m thinking of this as a paired proportion testing problem
(p1-p2) (McNemar’s test?). How can I do the power analysis in R?

Thanks
-- 
max

	[[alternative HTML version deleted]]