Statistical correlations like theories must be stated (a) in a series of subjectively testable operations, and (b) using the most parsimonious sequence, (c) with the most rational incentives.
Most of the time, a statistic is being used to lie or counter a lie. Like truth and falsehood there is no reason to state the truth other than to counter a falsehood. Nor any reason to state a falsehood other than to counter a truth.
Source date (UTC): 2018-06-04 13:05:00 UTC
Leave a Reply