This article considers the coverage probabilities of the approximate confidence intervals of Chen [1] and Wilson [2] and Wald [3] for the binomial parameter p. Also the Clopper-Pearson “exact” interval [4] is considered. Wilson’s interval (called Score confidence interval [3]) had the best performance, since it yields coverage probabilities close to nominal confidence levels, even for small sample sizes. Chen interval [1] has similar behavior as the Score interval [3]. However, Chen interval [1] has somewhat larger and more nearly conservative coverage probabilities, especially for small sample sizes.

confidence interval, binomial parameter, normal approximation.