Abstract

In an attempt to ensure that the probability of a Type I error is exactly α, the p-value obtained through the permutation approach turns out to be the most reliable especially when complete enumeration is feasible for small sample sizes. The problem with the permutation approach has been high computationally demanding in terms of space and time complexities especially for fairly large sample sizes when complete enumeration is not feasible. The bootstrap approach is adopted as a viable alternative to the permutation approach especially when the numbers of observations are fairly large for a two - sample testing problem. Illustrative implementations are achieved and leads to the computation of accurate p-values. 

Keywords: bootstrap, permutation, Type 1 error, p-values, observations, replications