When reporting the results from a Fisher's exact test, do I just report it like (fishers exact, p = .02) or is there a test statistic for Fisher's exact test?
$\begingroup$
$\endgroup$
1
-
1$\begingroup$ The test statistic is effectively the likelihood for the observed table. In a one-tailed test for a 2x2 you could arguably take one of the cell counts as a test statistic. I'd lean toward just giving the p-value but in the 2x2 case some people also report the odds ratio when they do Fishers exact test $\endgroup$– Glen_bMay 2, 2015 at 3:35
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
What exactly you will report in writing up your results will depend on your field and the journal you are submitting to. We cannot exactly answer that. However, your suggestion seems reasonable. Note that Fisher's exact test does not have a test statistic (like $t$ in a $t$-test): What is the test statistic in Fisher's exact test?.
answered May 1, 2015 at 20:16