Today's article in the Seattle Times, describing how ballot checks vary in Washington's counties, has some numbers that, indirectly, support my rough estimates on the amount of "distributed vote fraud" in Washington state.  (Unfortunately, the table containing the numbers does not seem to be available on line, so you'll have to buy a copy of the newspaper, or take my word for it.)

(Missed my posts on distributed vote fraud?  It isn't a new idea, though my phrase for it may be.  What I mean by distributed vote fraud is the vote fraud committed, not by party officials or candidates, but by individual voters acting on their own, for example, non-citizens voting simply because they want to.  I described it here, gave it a name here, and wrote a disclaimer on my estimates here.)

Now the numbers.  The table shows, for 22 of Washington's 39 counties, the total number of rejected ballots, and, of those, the number rejected for mismatched signatures.  The Seattle Times uses the table to make an important point about the varying rejection rates in Washington's counties, but we can also use it to check my estimate on fraudulent votes.   In those 22 counties, which hold most of Washington's population, 2828 votes were rejected for mismatched signatures.  Since about 3 million people voted in November, that means that 1 in a 1000 Washington state votes was rejected for this single reason.   Most of the counties make an effort to contact voters when there is a signature mismatch, and to correct it, if possible, though the efforts vary with the county.  Given that, I think it fair to conclude that nearly all of those 2828 votes are in fact fraudulent.

But that wasn't the only reason that votes were rejected.  Those same 22 counties rejected 19,577 ballots totally.  Not all of the ballots were rejected because of fraud.  Some, for instance, were rejected because a voter did not sign their ballot, which is absent-mindedness, not fraud.  But many of them must have been fraudulent votes, cast by, for example, people who were not registered.  So we can say that the election officials detected somewhere between 2828 and 19,577 fraudulent ballots, just in those 22 counties.

Did election officials detect all the fraudulent ballots?  Of course not.  In fact, there is good reason to think that election officials didn't even detect all the signature mismatches.  The clerk who checks the signatures has to balance Type 1 and Type 2 errors, though I doubt many of the clerks think of their decision in exactly those terms.  If a clerk decides, after the comparison, that the signature is invalid when it is really valid, then the clerk commits a Type 1 error.  If the clerk concludes that the signature is valid when it is really invalid, then the clerk commits a Type 2 error.  Most clerks will, I am sure, prefer to make Type 2 errors and let some fraudulent signatures slip through, rather than reject even a few valid signatures.  (And perhaps we should want them to commit many more Type 2 errors, though that is a separate question.)

The same kind of thinking is found, I am sure, in other places where election officials try to detect fraudulent ballots.  The clerks who check signatures will be inclined to let dubious signatures pass because it is less trouble for them.  Officials in other parts of the system will make similar judgments, for similar reasons.

In my October and November posts, I offered this guesstimate for Washington state.  There were somewhere between 3,000 and 30,000 fraudulent votes cast in the November election, giving a net gain to the statewide Democratic candidates of somewhere between 1,000 and 10,000 votes.  I think these numbers on rejected ballots show that my guesstimate is plausible.

The distribution of the rejected ballots supports my argument that most of the fraudulent votes cast last November were cast for Democratic candidates.  King County, with about a third of the state's population, had 1,976 of the 2828 ballots rejected for signature mismatches.  The Seattle Times thinks that shows uneven application of the law.  That may be part of it, but I think it also shows that there are more crooks in my home county than in the state as a whole.

Cross posted at Jim Miller on Politics.