[plsbenice34]

An invalid statistical argument that appears to support the truth is still invalid. Just like a proof for a mathematical theorem can be invalid even if the statement turns out to be provably true by other means.