[ESS] [R]- McNemar with unequal sample sizes

Martin Maechler maechler at stat.math.ethz.ch
Sat Dec 17 11:39:17 CET 2005


Please use R-help for questions on R !!

>>>>> "Christian" == Christian Montel <christian.montel at eligo.de>
>>>>>     on Sat, 17 Dec 2005 10:05:00 +0100 writes:

    Christian> Hi, i need to test the equality of proportions in
    Christian> a paired case, like here:
	
    Christian> after before + - + 10 14 - 5 53

    Christian> So usually i use the mcnemar.test(stats).

    Christian> Due to mortality, I now got the problem that my
    Christian> sample sizes in both factors are not equal any
    Christian> more -- there are less cases in the "after"
    Christian> condition. I already learned that Ekbohm (1982)
    Christian> and Marascuilo adressed the problem conceptually,
    Christian> but I'm not that good at writing my own
    Christian> functions.

    Christian> Can anyone give me a hint what
    Christian> function/test/package in R to use?

    Christian> Thank you very much for your help, Christian




More information about the ESS-help mailing list