I'm absolutely happy that the margin (for almost all the races) was wide enough that it was not only outside the margin of error, but outside the plausible margin of any kind of fraud.
All the same, I'm developing an idea about crowdsourcing some post-election audits to try and see if we can generate enough data to either confirm that there were irregularities - or make a convincing argument that there were not.
I'd love to toss out a challenge to the conservative bloggers who tried to raise the ACORN/registration fraud as a meaningful risk to the integrity of the election.
First of all, I think it's likely that every state has a significant number of 'zombie registrations' - registrations from people who have died, have moved out of state, or registered illegally either deliberately or inadvertently.
Given a set of assumptions about the quality of the work that ACORN did, you can either assume that they did great audits of the cards turned in by their low-wage registration workers, and thus created very few zombies, or that they didn't (either deliberately or through lack of resources, processes, or ability) and so created a lot of zombies.
So it's reasonable to assert that there's a 'zombie issue'
Now the question is where there are exploits we can devise that use those zombies in a plausible way to commit voting fraud.
I'll exclude from this list encouraging people who shouldn't be registered but are from voting. That doesn't rise to the level of systemic fraud on one hand, and to the extent that we accept Patterico's argument that this represents a large number of illegal aliens who have registered and voted - it represents broader policy issues.
So how would you audit for voting fraud like this??
I'd like to throw down a challenge to the people who care about this to work with me in devising ways that we can do some proof tests to see if the problem exists, to try and sample it's extent.
Lots of knowledgeable people have looked at this and made the flat statement that it isn't an issue.
If you think it is - and I'm in the 'it might be' crowd - then let's get some folks together and figure out how to demonstrate it.