In 2000, Democrats.com was the largest blog fighting the stolen election in Florida. In 2004, we were the largest blog fighting the the stolen election in Ohio. But in 2008, I'm not very worried about another stolen election. I'm not saying Republicans won't try, because they certainly will. I'm just saying they won't succeed. Why?

Stealing elections isn't easy. Voting systems vary by state, even by county. Some are all paper (like Oregon's vote-by-mail) or old-fashioned mechanical (like New York's levers). Optical scanner systems produce a paper trail that can be manually recounted. Only touchscreens are black boxes, and honest election officials can keep evil hackers away if they try hard enough.

As a result, no one can press a secret button somewhere and steal all 50 states. There are just a few states (like Georgia) with a uniform touchscreen system. In other states, Republicans would have to conduct covert operations in key counties. If Republicans really planned to do this on a significant scale, they would have stopped Democrats from winning the House and Senate in 2006. There's no evidence they tried, and even if they did they certainly didn't succeed.

We're less than a week away from the 2008 election. If Republicans want to steal it, they will have to engage in fraud on a scale so massive it would be impossible to hide.

1. Obama is way ahead in the Electoral College, where victory means 270 votes. My favorite site, fivethirtyeight.com, gives Obama a lead of 348.2-189.8=158.4

2. Obama's lead in the battleground state polls is large and steady. Obama's lead in the states he needs for 270 is at least 8%. In other words, to lose the election, Obama would have to drop 8% in 6 days, which isn't going to happen.

Bloggers like Brad Friedman and the Election Protection Wiki are collecting a lot of evidence of small-scale fraud by election officials and problems with voting machines. (Natasha Chart has great tips for protecting your vote.) But are these problems large and widespread enough to steal the election?

In both 2000 and 2004, the election was so close it was decided by a single state - Florida in 2000, Ohio in 2004. In 2008, Obama's 158-point lead is so large that it won't come down to one or two states so it wouldn't matter if Republicans stole both Florida (27) and Ohio (20).

Obama is way ahead in all the Kerry states. To reach 270, he only needs to add Iowa (where he leads by 10.3%), New Mexico (also 10%) and Colorado (8%). So McCain would have to steal those states too.

Moreover, Obama has solid leads in two more red states, Virginia (7.3%) and Nevada (6%) - so McCain would have to steal those too.

Stealing elections isn't easy. And in 2008 we have a very powerful tool to detect whether a state has been stolen: pre-election polling averages.

Pre-election polls are pretty accurate. The closer you get to Election Day, the more accurate they become. I have not seen any expert opinions, but I suspect an average of final pre-election polls will be more accurate than exit polls. Why? In states that allow early and no-excuse absentee voting, a high percentage of votes will be cast before election day; those votes simply cannot be measured with exit polls. Also exit polls aren't magical; they have their own margins of error, just like pre-election polls. Margins of error are always a function of sample size; the larger the sample, the smaller the MoE. The combined sample size of pre-election polls may be larger than exit polls, so their margins of error would be smaller. (This is all speculation; as I said, I have not seen experts compare the accuracy of pre-election polls with exit polls.)

Scott Althaus looked at the accuracy of the battlegrounds in 2004, using pre-election data from the final week. The inaccuracies were balanced - 6 states for Bush, 6 states for Kerry. They were also generally small; in only 4 states - WV, FL, WA, NV - were the inaccuracies greater than 2%.

In 2008, we have a lot more pre-election polling data to work with. How much more? Ask Karl Rove:

There has been an explosion of polls this presidential election. Through yesterday, there have been 728 national polls with head-to-head matchups of the candidates, 215 in October alone. In 2004, there were just 239 matchup polls, with 67 of those in October. At this rate, there may be almost as many national polls in October of 2008 as there were during the entire year in 2004. 

And to compound our polling bliss, we also have several websites combining that data to produce highly accurate averages. For example, here is the Pollster.com average for Virginia:

For McCain to steal Virginia when Obama is leading by 7.8%, McCain would have to flip 7.9% of the votes. No one in the political world would believe election day results that were so far off from the polls.

Certainly the Obama campaign will have county-by-county, even precinct-by-precinct projections to compare to the "reported" results. (Here's a great example of county-by-county analysis, in this case Oregon's early voting.) That will allow them to pinpoint the discrepancies, and demand hand recounts in those counties or precincts. And if the hand recounts proved fraud, McCain would be caught red-handed. Instead of going to the White House, McCain would go to jail.

Update 1: The graph below shows the gap between the national pre-election polls and the final results. In 2004 (pink), Kerry was projected to lose by 1% and actually lost by 2.5%. In 2000 (gray), Gore was projected to lose by 2% and actually won by .5% due to the last-minute discovery of Bush's DUI. In 2008 (blue), McCain is 6% behind - past trends show a net change of more than 2.5% is extremely unlikely.