QQ Plot Challenge

Is the sample reasonably consistent with normality?

80
5 1000

Changing the sample size generates a new sample from the same underlying population. Try both very small and very large samples.

Question: 1
Sample size: 80
Score: 0 / 0

The plot compares the observed sample quantiles with theoretical quantiles from a normal distribution.

Is this sample reasonably consistent with a normal distribution?

Advanced: from samples to sampling distributions

Keep the population from this question fixed, but now imagine repeating the experiment many times. Change the sample size and watch what happens to the mean.

Your challenge: how large does the sample need to be before the sampling distribution of the mean looks reasonably like the distribution expected under the Central Limit Theorem?

80
5 1000

Central Limit Theorem

The grey bars are simulated sample means. The dashed curve is the normal sampling distribution expected from the CLT, when the usual finite-variance conditions are satisfied.

Simulated sample means CLT prediction

Law of Large Numbers

This is one long sequence of observations from the same population. The line shows the running mean as more observations are added. When the population mean exists, compare it with the horizontal reference line.

Running mean Population mean

A QQ plot does not prove that a population is normal. The question is whether departures from the normal reference pattern are large enough to be visually convincing. With very small samples, strong departures can also be difficult to see.