Bootstrapping is a computer—intensive, nonparametric approach to statistical inference. Rather than making assumptions about the sampling distribution of a statistic, bootstrapping uses the variability within a sample to estimate that sampling distribution empirically. This is done by randomly resampling with replacement from the sample many times in a way that mimics the original sampling scheme. There are various approaches to constructing confidence intervals with this estimated sampling distribution that can be then used to make statistical inferences.

Goal

The goal of statistical inference is to make probability statements about a population parameter, θ, from a statistic,

, calculated from sample data drawn randomly from a population. At the heart of such analysis is the statistic's sampling distribution, which is the range of values it could take on ...