Thomas Lumley • 1/5/2026

Simulation and CLT

The article details a statistical simulation to analyze a new robust location estimator that finds the densest half of a data distribution. It implements the estimator in R, tests its asymptotic normality via the Central Limit Theorem using histograms and Q-Q plots, and compares its efficiency (variance) to the sample mean and median. It concludes by identifying the estimator as the known 'Tukey's shorth,' which has cube-root asymptotics and low efficiency at the Normal distribution.

0 comments

#R Programming #Central Limit Theorem #Statistical Simulation