Sampling distribution

Also known as: Finite-sample distribution

The sampling distribution of a statistic $T$ is the probability distribution of $T$ induced by repeatedly drawing random samples and recomputing $t$. Mental picture: histogram of many simulated $\hat\theta$ values.

Bias, variance, and MSE are properties of this distribution. The CLT approximates the sampling distribution of $\bar X$ by a Normal.

← Back to CS Glossary