Chen [1] proposed a practical approximate confidence interval (CI): where  is a Bayes estimate of the Bernoulli parameter π, and   is an approximate 100 (1 − 2α) percent CI for π. This CI is interval valued, equivariant, symmetric, and it possesses uniform asymptotic coefficients that are very close to the nominal ones. We have done a simulation study that indicates that, for moderate n (n ≥ 50) and (0.10 ≤ π ≤ 0.90), the maximum absolute difference between the actual confidence level of  and the nominal one is less than or equal to 0.025.

Bernoulli parameter, confidence interval.