Can you please explain the difference between those two things? I am a Canadian biochemist/biologist and really haven't a clue about American politics, although I find them interesting 
Exit polls - they ask something like 1 in 10 people, after they vote, WHO they voted for. It's incredibly accurate throughout history, as you get a population sample of nearly 10%. They use to it announce the winner the minute the polls close, and they are nearly ALWAYS spot on with their predictions.
Well, in 2004, john kerry had a good lead in FL, and a great lead in Ohio, based upon these exit polls. This would have given him the victory. But, suddenly the new electronic voting machines (the prez of this company is a strong repub campaigner who voewd to help put Bush in office), well these machines said Bush (who I voted for actually) won with a nice large margin in both places.
Happening in Florida would be a 1 in 50,000 chance. In Ohio, the odds were even higher. Together? It was in the billions. Then, a programmer came fwd, and passed a polygraph (lie detector), claiming he was paid by Bush'd brother jeb (the FL governor) to write software which reverses electronic voting machines (strangely, if you reverse the results, they line up with exit polls perfectly).
A FL detective vowed to break the case open, but then drove across state lines to Georgia (no autopsy state) and beat himself up, bound his wrists, then cut his wrists while several strange men hung outside his hotel room. Case died with the lawman. But anyone with half a brain and an understanding of statistics will look at the 2004 election with disgust.