Here is a nice Java applet.

You can choose one of 8 different distributions for your random variable X. The applet then draws randomly a population of some size N times and computes a histogram of the values (averaged by N).

You will see, that – irrespective from the chosen initial distribution – as the sample size increases, the distribution will converge to a normal distribution with the same mean, but decreasing variance (std dev).

Bunnies & Dragons - and what they have to do with the CLT

… but still without any formulas.

…. and perfectly explained by the way.

The example at the end of the video underlines the fact, that the CLT makes a claim about the population mean and not about individuals!

public/central_limit_theorem_clt.txt · Last modified: 2014/01/03 14:27 (external edit) · []