[R] suggestion for proportions

[csrabak]

Em 8/9/2011 05:24, Viechtbauer Wolfgang (STAT) escreveu:
> I assume you mean Cohen's kappa. This is not what the OP is asking
> about. The OP wants to know how to test for a difference in the
> proportions of 1's. Cohen's kappa will tell you what the level of
> agreement is between the two tests. This is something different.
>

Indeed it is different.

> Also, the OP has now clarified that the data are paired. Therefore,
> prop.test() and binom.test() are not appropriate. So, to answer the
> OPs question: yes, mcnemar.test() is what you should be using.
>

While it seems a consensus has been reached in the understanding of OP's 
wish, I still not completely sure, as the word "accuracy" has been used 
as synonymous of /proportions/ for the two tests.

Also as the description of "2 different diagnostic tests" has been 
written by the OP I offer my interpretation that the agreement between 
tests is to be considered.

OTOH if the accuracy can be assessed against a golden standard, I would 
suggest approaching the problem with confusion matrices.

HTH

--
Cesar Rabak