Naked Statistics: Stripping the Dread from the Data

by Charles Wheelan

large, biased sample is arguably worse than a small, biased sample because it will give a false sense of confidence regarding the results.