The method of moments and maximum likelihood were introduced earlier as ways to estimate a single unknown parameter. The first section of this chapter extends these methods to provide estimates of two unknown parameters.
The easiest way to find a confidence interval for a parameter is based on an estimate's standard error (or an approximation) and the assumption that it has an approximately normal distribution. This is called a Wald-type confidence interval.
Although maximum likelihood estimators are close to normally distributed in large samples, their distribution can be far from normal when the sample size is smaller. A better way to find a confidence interval based on a "pivot" is described in the second section.