I'm using Fisher's exact test for some of my own data and occasionally get p-values of 1.000 in Stata. I'm wondering how to interpret this. At first I thought this was a mistake, but thinking a bit more, I think it's possible - am I correct in interpreting this as "the probability of a contingency table as or more extreme than the one observed is essentially 1"? This generally happens with the proportions in the two groups are very similar. As with a p-value of 0.000, is there a convention for rounding these when presenting them in tables? Should I round down to 0.999? Or just leave it at 1.000?
Hi, a p-value of 1.000 is very unusual but it is possible. It is probably ok to round down to .9999 for presentation purposes, and you are reaching the same conclusion.
Thank you, this is helpful.