Politics: The 2004 Exit Poll Discrepancy

Note: This posting has had some significant updates appended to it.

All over the world, when democracy is put to the test in an election, one of the ways to verify that the election is fair is with exit polls. Exit polls aren't meant to replace voting, but help to support its legitimacy. As you probably know by know, there are troubling discrepancies between the exit polls and the actual vote results in our last election.

What's troubling is not that there are discrepancies. That is to be expected. What is troubling is the exact nature of the differential, and the seeming lack of official response to its existence.

What is somewhat troubling is that the discrepancies are large in the battleground states, and small in the non-battleground states. And not a small difference, in many non-battleground states, the exit polls are off the final results by less than a single percentage point, in some battleground states, the exit polls are off by almost 10%. This suggests that something unknown and unprecendented is happening. Since all of the battleground state discrepencies favor President Bush, the blogosphere is running rampant with cries of "fraud". I'm not ready to make that claim yet, however...

What is very very troubling is that this isn't getting the official or journalistic attention it seems to deserve. These numbers are way off base, and the explanations put forth so far don't seem to hold water. I remain hopeful that a full investigation will be conducted, and that a real explanation will come forth, but until then, I'm am not at all surprised that there are those who question the legitimacy of the election results. In other countries, exit polls discrepancies of this sort have been cause for the U.S. to so strongly question the validity of the election results that we have pressured the winner of the election to step down.

We need to demand, if we want to keep our democracy, that this be investigated and satisfactorily explained.

For more detail, here is an excellent paper on the subject, by Dr. Steven Freeman, which covers the data I have cited above in more detail.

Update: The situation is worse than I thought. One by one, the results of the election in each state are going to be certified. As soon as that happens, the computer memories will be cleared and the paper records of the votes and ballots will be destroyed. In many states, this will start happening as soon as tomorrow. While there are people asking for a recount, there are in many cases large mandatory filing fees.

So, in what I consider a truly bizarre turn of events, if you're as concerned about the integrity of our elections as I am, I suggest you contribute some money to the Green Party's presidential campaign, now that it is over. The Green Party and the Libretarian Party have joined together to ask for recounts, starting in Ohio, since the Democratic campaign seems unwilling to do so.

Update #2: The first official overturning of results due to faulty electronic voting machines has occured. Apparentlly anyone in this Indiana county who voted a straight Democratic ticket got all their votes cast for the Libretarian candidates instead. I'm wondering what will happen if it's found that this is a nation-wide problem, because this might well be the glitch that accounts for the exit poll differential.