Chapter 17. Fisher's Exact Test
17.1 Introduction 317
17.2 Synopsis 317
17.3 Examples 318
17.4 Details & Notes 321
Fisher's exact test is an alternative to the chi-square test (Chapter 16) for comparing two independent binomial proportions, p1 and p2. This method is based on computing exact probabilities of observing a given result or a more extreme result, when the hypothesis of equal proportions is true. Fisher's exact test is useful when the normal approximation to the binomial might not be applicable, such as in the case of small cell sizes or extreme proportions.
Using the same notation as in Chapter 16, you observe X1 'responders' out of n1 patients studied in one group and X2 'responders' of n2 patients in a second ...