Suppose with , and let consistently estimate . A two-sided Wald interval isThe Slutsky theorem makes the standardized error asymptotically standard normal, giving coverage tending to at each regular parameter value. The estimated standard error is , not .
This is an asymptotic construction. For a binomial proportion, the usual substitution gives , which can extend beyond and collapses when is zero or one. Those features explain its poor finite-sample behaviour near boundaries; binomial proportion confidence intervals use other constructions when appropriate.
Articles by others on the same topic
There are currently no matching articles.