How to Make the World Add Up : Ten Rules for Thinking Differently About Numbers

by Tim Harford

Sampling error is when a randomly chosen sample doesn’t reflect the underlying population purely by chance; sampling bias is when the sample isn’t randomly chosen at all.