"Fixed" assumption of Fisher's Exact Test
A 2x2 contingency table contains counts of variables X1 and Y from retrospective data. The goal to is to test for association between X1 and Y. One of four (25%) contingency table cells has an expected count less than 5, so Chi-square Test of Independence is not valid. Is Fisher's Exact Test a valid alternative approach if the data is retrospective and the X1 and Y counts were not "fixed" (aka, X1...