Using Bootstrap for More Accurate Confidence Interval

This article investigates the bootstrap methods for producing good approximate confidence intervals. The first section introduces the motivated example and compare different approximated confidence intervals with the exact intervals. Section two serves as an overview of different bootstrap methods. Inspired by Efron's paper (1994), this article further explores the possibility to apply the bootstrap method to common odds ratio in a series of 2x2 contingency tables. The third section explains the procedure and gives the results of Monte Carlo simulations.