Averaging averages is only valid when the sets of data — groups, cohorts, number of measurements — are all exactly equal in size (or very nearly so), contain the same number of elements, represent that same area, same volume, same number of patients, same number of opinions and, as w...