Not that I know of, but the point is that our system ought to be robust against historical and statistical flukes. If you're going to decide elections via first-past-the-post at all, you should at least make the election pass a 95% likelihood hypothesis test for difference-of-means or something.