OT: This Week In Voting Machine Security 3431

Tim Streater

Representatives of the political parties may be present at the count. Also, representatives of the political parties get to review ballot design and other things prior to the election.

There is still some scope for small scale fiddling. As in the UK, postal ballots are the most vulnerable. But there is also the trick of getting fake people to vote. As you know, in the US voter participation is low. Also, the precincts periodically post of a list of who has and who hasn't voted yet. Getting people to vote on behalf of people who haven't yet voted and are unlikely to vote is always possible.

Other schemes are by reducing voter turn out in certain precincts. If party X controls the county (which runs the election) and precinct A is known for voting heavily for party Y, then in principle the county could mess things up in precinct A, by say, providing less convenient polling places or not even enough poll workers so lines are long. This is a common accusation in the US. Indeed, this is the form of the accusations of fiddling in Ohio, 2004. Another things that adds to the confusion (and the suspicion of fraud) is that people lie to pollsters in a fairly systematic way. On the whole, people vote more conservatively than they tell pollsters. Think of the Tory's surprise victory in the early 1990s.

In the past, there have been some counties and precincts that have been under so strong control of a party that despite the safe-guards they could take full control of the count. If you find precincts that tend to always report late, and then produce balance tipping results, these are good candidates for investigation. Thirty years ago, this kind of thing happened often. Indeed, the 1960 presidential election was determined by a few late reporting precincts in Chicago, which at the last minute seemed to produce exactly the numbers that Kennedy needed to beat Nixon. I don't know what procedural changes were insbreastuted to bring that sort of thing to an end (or at least to dramatically reduce it).

Going much further back, the election of 1876 was a true disgrace. The winner (Hayes) won fewer popular votes than Tilden, but Hayes won in the electoral college. Tilden's campaign tried to work around that by bribing electoral college electors in Florida where the vote had been very close. What appears to have been exceptional in that election is that Tilden got caught.

The general rule, unsurprisingly, is that the closer the "true" election the easier it is to fiddle.

Sure you do. It's just that in the UK it is such an established tradition that people accept it as one of the quirks of the electoral system. In the UK, there can be pretty wide variation in population from one consbreastuency to another. This makes rural (more likely Tory or LDP) votes count more than urban votes.

I'm a big supporter of using a preference voting system (though not proportional representative).

As I said in a previous post, people can't really cope with an election result being within the margin of error of the tallying system. We can do what we can to reduce that margin of error, but I don't think it can ever be eliminated.

-j

-- I rarely read top-posted, over-quoted or HTML posts My Reply-To address is valid.