Interpreting post hoc multiple comparison for RxC Fisher's test

[monsterpie]

Hi everyone,

I have a conducted a Fisher's exact test with Monte Carlo simulation which resulted in a significant p value. My row variable, which is the # of people, has 4 levels (0, 1-50, 51-100, 101-150) and my column variable has 5 levels (AccountA, AccountB, AccountC, AccountD, AccountE).

To identify the significant pairs between groups, I've conducted multiple comparisons using a 2x5 tables (e.g., 0 vs 1-50, 0 vs 51-100 etc.) with Bonferroni adjustment, and the adjusted Bonferroni p values are below. I am unsure of how to interpret the Bonferroni p values...for example, how would the adjusted p value for Test 1 (0 vs 1-50 people) be interpreted? 

[StatDave]

You interpret them just like you would the raw p-values, but now you are controlling for the familywise error rate and can be more confident that you do not falsely declare some tests significant in the set of tests. See the Overview section in the MULTTEST documentation for basic discussion and more in the Details: p-value Adjustments section.