Bootstrapping
Visualize statistical bootstrapping through resampling with replacement to estimate distributions.
Free theory articleVisualize statistical bootstrapping through resampling with replacement to estimate distributions.
Concept & interpretation
Bootstrapping is a powerful, computer-intensive resampling method used to estimate the sampling distribution of almost any statistic. Instead of relying on complex theoretical formulas (like those derived from the Central Limit Theorem), bootstrapping treats the observed sample as if it were the entire population. By repeatedly drawing random samples from this original dataset with replacement, we can calculate our statistic of interest (such as the mean, median, or variance) thousands of times. The distribution of these calculated statistics—known as the bootstrap distribution—provides a practical way to quantify the uncertainty, calculate standard errors, and construct confidence intervals for our original estimate.