A few politicians with strong motivations to find any evidence of significant election fraud they can, and holding positions that give them the authority and, indeed, responsibility, to go looking for it have done so, and so far as I know they've all found little enough that I think it'd be fair to round what they've discovered down to "nothing". All evidence I'm aware of is that there's a vanishingly small amount of not-particularly-well-coordinated honest-to-god deliberate voter fraud, plus a somewhat larger but still small amount of accidental illegal voting (I gather this is somewhat easy to do unintentionally if one might claim residence in more than one state) and bad book-keeping or clerical errors. If deliberate voter fraud were a significant problem I would expect these efforts to have turned up far more, even if they didn't find the full scope of it.
> I can't remember the source but one study said there are about 3 million extra people registered in the US (people who are eligible - people actually registered). They may not be acting fraudulently, say if they moved from one state to another, or a different precinct in the state. The 2000 election was decided by 500 votes.
I take your broader point, but do note that the 3,000,000 -> 500 here is comparing apples and oranges, due to the way our Presidential elections work (it wasn't decided by 500 votes nationally)
Yes, 500 votes in a couple precincts of a state.
I can't remember the source but one study said there are about 3 million extra people registered in the US (people who are eligible - people actually registered). They may not be acting fraudulently, say if they moved from one state to another, or a different precinct in the state. The 2000 election was decided by 500 votes. Another study that I can't remember estimate that the same people without ID are, in general, the same people who are not registered to vote.