Bootstrapping is an approach to properties of statistics, such as sampling variances, standard errors, and confidence intervals, that does not rely on a particular assumption about the shape of the distribution around a given statistic. Bootstrapping is therefore said to be a nonparametric approach to statistical inference. It can be particularly useful when the researcher does not know the theoretical distribution of a given test statistic or when no such distribution exists.

Bootstrap methods for evaluating statistics rely on data-based simulations wherein the observed data stand in for the population of interest. Measures of uncertainty around a statistic that are obtained via the bootstrap therefore might be thought of as being drawn from samples of a given sample, as bootstrapping is a computationally ...