Paul Crowley comments on Consider Representative Data Sets