Who Won the Election?

You would think that this had been settled by now. In case you were not sure - Bush won. I bring this up because a couple of posts on the new Huffington Post say that Kerry won. In addition to the usual suspect, John Conyers, Jim Lampley, a sports writer, says that the exit polls were correct and that the voting machines were all fixed. When a rival blogger pointed out the assessment report by the exit polsters themselves, he responded that the pollsters got it wrong because they assumed that the votes were accurate. I can't resist a challenge like this. Here's my take on why the vote could not have been fixed. Since I live in Ohio and it was THE swing state, I will limit my observations to my own state.

First, it should be pointed out that the polls leading up to the election had predicted a Bush victory. The site ElectoralVote.com showed Bush in the lead almost continuously from the Republican convention to the election. Kerry did do slightly better than the polls had indicated, taking Wisconsin by a razor thin margin. In fact. Bush did not win any states that he had not been leading in the polls prior to the election.

Given this, why didn't the exit polls raise a red flag when the first results started coming in? The answer lies in the undecided voters. The expectation was that they would all break for Kerry. When Kerry got off to an early lead in the exit polls, the pollsters chalked it up to the undecided voters finally making up their minds.

So the advance polls match the actual election count. Is there anything else that we can learn from the polls? Yes. The results of the polls varied from pollster to pollster. Each poll uses a different method for conducting its questionnaire. The poll that came the closest to the actual election result was totally automated with pre-recorded questions. This meant that every person polled got exactly the same questions with identical emphasis. It also means that the people being polled were not giving their responses to a live human. As polarized as the electorate was and as strong as some anti-Bush sentiment was, this could throw off the results. Both Bush and Reagan before him always had a core of "closet voters" - people who voted for them but would not admit it in public.

This was also a likely factor in the exit polls. The majority of pollsters were fairly young women - an demographic group associated with Kerry. All it took was three Bush voters out of 100 people polled to skew the poll. The biggest error rates were at precincts in large cities which is where all of Kerry's support was. This would also affect the Bush voter, making him less likely to announce his vote to a possibly hostile audience.

So there are the arguments for why the exit polls were wrong. What about the actual vote? How hard would it be to change the results?

Let's start with punched cards and optical mark ballots. These are collected at the precinct then taken to the county Board of Elections and counted. The count is done by card readers attached to a PC which does the actual tabulating. The PCs are stand-alone. In order to change the results you have to have physical access to the PC.

When the votes are tabulated, the Board of Elections is watching. The Board is made up of equal number Democrats and Republicans. Also complicating things, Ohio law calls for rotating the candidate's names so that they are in a different order from one precinct to the next. This is so that no candidate gains votes from being the first name on the ballot.

So in order to throw the election, you would have to enter a secret code into the PC doing the counting. You would have to do this in from of witnesses from both parties. Also, since the PC only knows which columns have been marked, not which candidate is which, you would have to tell the PC which column was to get the extra votes for every precinct. And those witnesses are still watching.

And you would have to arrange for this to happen in every county meaning that you would have to have dozens of confederates.

The touch screen machines are a different matter. There are two general types of touch screen voting machines. One uses a touch-sensitive CRT like the self-checkout lanes use. The others have membranes and lights for each position.

The CRT-based machines could have hot-spots on the screen that, when touched, would let you override the counts already entered. The membrane ones would have to have a different code. Either way, according to the theory, a Bush operative would have to have some time alone with the machine in order to punch in the over-ride codes. Just like before, this is not allowed so you would have to be very careful. You would also have to get your hands on every machine. And you would have to have operatives in every precinct.

I suppose alternatively, you could give the override codes to some normal voters and have them vote late in the day. You would have to send several voters to each precinct to be certain that they would have access to all of the voting machines.

The number of people needed to pull this off is staggering. And it would take this effort. The analysis of the exit polls showed that there was no different between different types of voting machines. If one type was fixed, they all were.

So, hundreds or thousands of Republicans were in on it. Not one suffered remorse and confessed. This was done in front of hundreds or thousands of Democrats but not one noticed anything at the time.

Plus, every voting machine or tabulator in Ohio has hidden cheat codes. This includes the numerous ones that are a decade or more old.

Or the exit polls were wrong for the reasons I gave above.

Which is more likely?