A binomial test assesses whether a count of successes in independent binary trials is consistent with a specified success probability. Under the null hypothesis, each trial is modeled with the same probability of success; an exact test calculates probabilities from the binomial distribution rather than a large-sample approximation. section 4.2.3
Small clinical trial
If 14 of 20 patients recover and the historical recovery probability is 50%, a one-tailed test asks how likely 14 or more recoveries would be under that baseline. The exact probability is approximately 0.057. section 4.2.3
