Interactive · from Volume 1

Sample size and stability

The same population every time. Draw a small sample repeatedly and the estimate wanders. Draw a large one and it barely moves at all: at n = 2000 the cloud is dense and the estimate sits on the truth.

This sample--

n--

Spread so far--

Sample size

Draw again

Every estimate you have drawn, against the population value.

Keep going

Measurement error and attenuationAdd noise to either measure, watch r pull toward zero.Within and between subjectsOne nested dataset, three different correlations.