An Introduction to the Statistical Bootstrap : Examining Confidence Intervals Using Flight Delay Data
Bootstrapping is an important resampling method with the ability to provide statistical inference with fewer assumptions then most other statistical inference methods. Bootstrapping can be more accurate than classical formula based methods and with the use of a computer can also be quicker to calculate. This paper introduces and describes several popular methods of bootstrapping. In this paper the formula t, bootstrap t, t, BCa and percentile methods are introduced and modeled using a case study on flight delay times. The implementation of these methods in RStudio is discussed and the resulting intervals are compared to each other. The end result is a basic history of the bootstrap and an overview of the methods, with a discussion and example of how and when popular methods would be used.