Cover: Bootstrap Confidence Intervals in a Complex Situation

Bootstrap Confidence Intervals in a Complex Situation

A Sequential Paired Clinical Trial

Published in: Communications in Statistics, Theory and Methods, v. 19, no. 1, 1990, p. 181-195

Posted on rand.org 1990

by Sally C. Morton

This paper illustrates the use of several bootstrap confidence interval methods in a complex yet practical situation. The confidence interval for the difference between two treatments in a simplified sequential paired clinical trial, which is analogous to setting an interval for the drift of a random walk subject to a parabolic stopping boundary, is considered. Three bootstrap methods of construction are applied: the accelerated bias-corrected, the automatic percentile, and the bootstrap-t. The results are compared with a theoretical approximate interval. Difficulties inherent in the use of these bootstrap methods are discussed. The automatic percentile method is shown to be the easiest to apply and to work well.

This report is part of the RAND external publication series. Many RAND studies are published in peer-reviewed scholarly journals, as chapters in commercial books, or as documents published by other organizations.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.