Taillard's statistical test is designed for the analysis of 2×2 contingency tables. Sample A contains a successes out of na observations, and Sample B contains b successes out of nb observations. All observations are assumed to be mutually independent random variables.
To perform the test, we additionally assume that the success rate of A is higher than that of B (i.e., a/na > b/nb).
The one-sided Taillard test computes the probability of the null hypothesis stating that the success rate of A is not strictly greater than that of B.
The two-sided Taillard test computes the probability of the null hypothesis stating that both success rates are equal.
It has been shown in Taillard et al. that this test is more powerful than Fisher’s and McNemar’s exact tests.
Fill in the following form, knowing that the condition a/na > b/nb should be respected.