I have contingency table with 2 treatments (t1, t2) and 3 stages (s1, s2, s3) as following:
T1 T2
s1 x1 x2
s2 y1 y2
s3 z1 z2
I got a significant interaction when i did the fisher exact test. However, now i want to know where is the significant interaction. I tried fisher.multcomp (RVAideMemoire) and pairwise_fisher_test (rcompanion). However, it gives me the values only for different stages (for example, recovery T1 vs. growing T2) between the two treatments. I would like to know the results between the two treatments for a same stage (for example, recovery T1 vs. recovery T2). Can someone help me in this? Thanks a lot.
It depends what you mean by "between the two treatments for a same stage". For example, would comparing the two counts in one stage make sense ? That is, if the counts for the table for the first two rows were S1: 10, 100; S2: 1, 10. Does the difference in 10 and 100 make that stage significant, or does that fact that S2 has the same proportion change the conclusion ?
Thank you for your reply Salvatore. I would like to know if the proportions between treatment is different (so in your example: if 10/100 is different than 1/10). As I have a table 2 x 3, I thought about doing a pairwise analysis by comparing the proportion of one of the three stages over the 2 others with another fisher's exact tests...unless there is specific psothoc tests for that.
You can do that, if it makes sense for your purposes. It's a "one-versus-rest" approach.
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology