ABSTRACT
The bootstrap method differs from the traditional parametric approach to inferential statistics because it uses sampling with replacement to create the sampling distribution of a statistic. The bootstrap method doesn’t take a random sample from a population in the same way as that used in our previous inferential statistics. The bootstrap method is useful for reproducing the sampling distribution of any statistic, e.g., the median or regression weight. The basic idea is that conclusions are made about a population parameter from a random sample of data, but in which a sampling distribution of the statistic is generated based on sampling with replacement. The factors that influence the shape of the sampling distribution are therefore important because it is the bootstrap estimate from the sampling distribution that allows us to make an inference to the population.
