2

I need to extract data from an article that reports a Fisher’s exact test p = 0.56; odds ratio 95% confidence interval [0.50, 1.2]. Knowing that both groups have 80 individuals is it possible to find out the difference in numbers based on the Fisher's test in both groups?

Silverfish
  • 23,353
  • 27
  • 103
  • 201
Marc
  • 53
  • 3
  • 1
    Welcome to our site! There's no need to write "thanks" at the end of your posts - in fact we prefer that you don't. But you can always thank someone who provides you with an answer by accepting it (pressing the green tick). – Silverfish Apr 29 '16 at 11:33
  • 1
    Are you sure those numbers are correct? I have done a brute force search over the 81 * 81 possible tables and cannot find one consistent with the p and ci you quote. – mdewey Apr 29 '16 at 13:01
  • Well, I got the exact numbers from an article I want to use. Is it commonplace to report the Fisher's exact test with CI of the odds ratio? – Marc Apr 29 '16 at 14:17
  • I also did a brute force search and couldn't find one that matched those numbers. The CI seems way too small if there are only 80 individuals in both groups. I think reporting a CI with a p-value is pretty common – Peter Calhoun May 02 '16 at 03:47

0 Answers0