|
Abstract:
|
In sample surveys, the most commonly used confidence interval for proportions adds and subtracts the margin of error from the survey weighted estimator. This “Wald-type” interval can be problematic when the sample size is small, often exhibiting undercoverage and interval endpoints that are not between zero and one. Recent literature has addressed this problem and proposed potential solutions, but more work is needed. We will discuss some of the methods that have been proposed and the challenges that remain. Participants will have the chance to discuss their research and experience with computing confidence intervals for proportions in complex sample surveys.
|
