A New Test for 2 × 2 Tables

IF an experiment yields results in the form of a 2 × 2 table: where m and n have been fixed in advance, a test for deciding whether there is evidence of association between the attributes A and B and the mutually exclusive and exhaustive attributes P and not-P has been given by Fisher¹. This test, on the null hypothesis of no association, associates the results with a probability m!n!r!s!/N!a!b!c!d!.