Nonsense. Elections are not random samplings, they are enumerations, for which the "margin of statistical error" is zero. Inaccuracies in the enumeration can be caused by fraud (Florida 2000, Georgia 2002) or human error. If "honest-error-induced" inaccuracies were knowable as a percentage of the total vote, and if they were randomly distributed across the whole electoral universe, then it would be possible to estimate the margin of statistical error within that total of randomly distributed honest counting errors. But they aren't, and so it aint.
Shane Mage
"When we read on a printed page the doctrine of Pythagoras that all things are made of numbers, it seems mystical, mystifying, even downright silly.
When we read on a computer screen the doctrine of Pythagoras that all things are made of numbers, it seems self-evidently true." (N. Weiner)