So would the election ever pass as a scientific analysis of voter opinion? Is there a monster margin of error? It would seem that with a race this close, and a voting system this complicated and in places, antiquated, it's simply impossible to determine with statistical significance who wins.