[R] Fisher's Exact test
mai at ms.uky.edu
Tue Feb 16 17:46:35 CET 1999
Simon Fear wrote:
> Try this definition of a P value: the
> probability, under the null hypothesis, of observing data as, or more,
> extreme than that actually observed.
I agree with the definition. But IMHO, I think the extremeness should be
defined in the likelihood ratio
sup Lik(Ho)(x) / sup Lik(Ha)(x) is the measure of
extremeness of the observation x, the smaller the more extreme.
Because an observation have small probability under Ho cannot be thought
of as extreme if the same probability under Ha is even smaller,
(think of cont. case or near cont. case) instead it supports Ho.
I also like the idea of having an option in the R function for
two sided P-value, since there seems to be several competing
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
More information about the R-help