(Cross-posted at SSP)
(Note: To counteract widespread speculation, we should remember that the Brookfield totals are not in dispute without a widespread conspiracy. The result of +7500 Prosser was reported yesterday; the question is if they were recorded in the totals yet or not.)
I'm kind of skeptical about the 14,000 new votes supposedly discovered in Waukesha County today. As such, I decided to quickly do some investigation into the state of the county.
Per WisPolitics at 1:18 PM on Election Day:
In Waukesha..."everything else doesn't seem to be that busy."
[[Deputy City Clerk]] Kozlik said final turnout should be in the 20 to 25 percent range, which is on par with other spring elections with a statewide race on the ballot.
Another source reported Waukesha turnout might be "as high as 35 percent".
Officially, turnout (pre-Brookfield) was 42% (per county site) - 110643 votes. After Brookfield, it should roughly be 47%. That's astonishingly high turnout for a spring election. The state average was 33% (Dane County - Madison which reported super-super-high turnout was 49%.) Turnout might have picked up over the afternoon/night, but that raises a major major warning bell for me. If there were so many voters, why didn't anyone see them?
Now, Kathy Nickolaus, the county clerk has been criticized for the poor security of her computer systems, in which election data is literally kept on her personal computer. The (pretty conservative I would assume) Waukesha County Executive Board reprimanded her for this and ordered an audit in January. Also see http://www.jsonline.com/... and http://www.jsonline.com/...
She was also under criminal investigation in 2002 until granted immunity related to a case (in which Prosser was also involved) where Republican legislative leaders had their staffers (paid by taxpayers) work on the campaign. Per State Senator Chris Larson, "Heard rumor the Waukesha County Clerk was also the tech point person for Republican caucus during the scandals 10 years ago." Confirmed by WSJ
In short, it's eminently possible that something fishy is going on. Completely possible that it's legitimate and human error, but this doesn't exactly pass the smell test with flying colors for me.
The main thing really suspicious for me is the reported 47% turnout. That's basically the same level as Dane, and we heard stories about that all day long. But Waukesha supposedly had the same level of turnout, but nobody seemed to notice it.
More info should come in soon at any rate.
UPDATE 1:
The other two main conservative suburbs in Wisconsin had similar turnout
Washington reported 46.13% turnout.
Ozaukee reported 45.70% turnout
I can't find any election-day reports whatsoever as to their turnout. Compare to estimates of 20-35% in Waukesha
What I'm trying to figure out (made very hard by lack of precinct data in Waukesha - something the conservative county officials criticized her and a main reason I'm suspicious since we have nothing but the county official's word that the city of Brookfield wasn't counted before) is where exactly the 47% turnout came from. Waukesha city itself only had 31.4% (per official results)
Also changed the title
UPDATE 2:
Essentially, I'm looking for more results of voter turnout during the election and comparing to actual #s.
Appleton (Outagamie County, home of Prosser) - 15% turnout by noon, 45% estimated. Result: 41.2% actually voted
The high turnout was noted during the day
Green Bay - No firm #s, but using same standard (they had 10% at 10AM,20% at 2PM, so I'll call it 15% by noon, same as Appleton), 45% estimated. Actual result county-wide: 'More than 43%.'
High turnout also noted
In short, it does seem to me at first glance that estimates of turnout during the day were fairly accurate (if slightly exaggerated.) This only makes the Waukesha results seem even more suspicious - they estimated 20% first, with 35% at most... and then results were 47%?
UPDATE 3
City of Wausau (Marathon County): No exact predictions, but high turnout observed. Actual result: 38% county-wide, roughly 40% in city (36% in one end, 44% in the other)
Fond du Lac County -40% predicted. Actual #s: 26,121 votes from official results. Voting-age population:72,807 -> 36% turnout
CAVEAT: This does not necessarily mean anything. I found a report of 20% turnout in Wausau and Stevens Point (both were really 40%.) But the report doesn't really cite anything and may just be based off of the 20% pre-election estimate. Also, Milwaukee City had 25% estimated vs. 38% actual.
It's possible that the Waukesha officials just picked 20% as estimated turnout since everyone in the state was saying that before.
Anyone from Waukesha here who can tell us their experiences?
Anyways, I have tons of cosmology homework to work on and one sleepless night worrying about this election is enough, so that's it for the updates.
UPDATE 4 There is currently another Rec List diary noting 7500 votes is just enough to put it out of the free recount range. My reaction to that is that the author might be getting ahead of themselves with the allegations; absent a vast conspiracy, Brookfield did have Prosser+7500 (I saw a news article on it from just after election day.)
The issue is how we know it hadn't already been counted since the County Clerk doesn't give precinct-level data and has everything on her personal computer. I suppose it would be plausible (IF we assume fraud, which is not at all given or even that likely) that she could 'select' the 'forgotten city' as one that had close to that.
But thinking about it, that explanation doesn't seem to hold water. 0.5% is roughly 7400 votes, and Kloppenburg had about a 100-200 vote lead before that came out. if I had been in charge of such a fraud attempt, I would have picked a number more like 8000-8500... so that it would be more safely outside that range, since it's going to move from recanvassing or etc. Such an attempt would be like trying to 'fix' an election so you win by 10 votes... possible, but why bother when you could go for a safer number that wouldn't look much more suspicious?
UPDATE 5 I've sent emails to TPM and the Milwaukee J-S editors laying out the quick argument that something is weird with reported turnout in Waukesha and asking them to investigate a bit by interviewing poll workers and voters. Hopefully one of them will follow up (they did respond to my last tip), and we'll have more info.
UPDATE 6 I've noticed that Ramona Kitzinger, the Democratic county vicechairperson and member of the canvass board said that it "went over everything" and the numbers were correct, so I'm feeling less suspicious than before. Still, I would be in favor of a county recount. I did google Kitzinger extensively to verify that she was a normal Democrat local organizer, and that did appear to be the case (there have been cases of LaRoucheites and etc. taking those posts.)
Also, another much more cautious and nicely reasoned rec list diary on the same issue but a different focus
UPDATE 7 An excellent diary shows that Microsoft Access automatically saves, and hence Ms. Nickolaus's story is inconsistent. As a result, I'm now much more soured on her story.
Right now, considering all the evidence that's come in, and trying to consider stories consistent with everything that's likely to be true, I think that there are two main possible scenarios. I would rank their likelihood as roughly 90% chance for #1, 10% chance for #2.
1. The city of Brookfield actually was left out, but County Clerk Nickolaus was dishonest over the reason. As a quite conservative and partisan election official, this is my speculation, but it is quite possible that she was paranoid over possible Democratic voter fraud, and hence held back a few thousand votes so (in her mind) we wouldn't know how much to steal. I cannot personally think of any other plausible scenario in which the vote is legitimate, but still consistent with reasonable facts. The caveat with this scenario is that if this was the case, it would be much simpler and more plausible for this to be reported on election night.
2. Fraud exists on a much larger scale than just the County Clerk. Either the voting machine code is suspect, or several of the precinct-level officials are involved as well. It's impossible that any fraud is Ms. Nickolaus's doing alone, as the county-level canvass ensures that her spreadsheets agree with precinct-level data. What had to be done if fraud exists would be to disperse Brookfield's votes throughout the other municipalities, then set the city's results to 0 to have a legitimate excuse (and fool the Dem observer.)
In such a case, IF we assume fraud existed (tenuous assumption, of course), it's extremely unlikely that this was an isolated incident, but rather something that the infrastructure existed for (you don't just get people together to commit election fraud over one night. It has to be planned over time.) This would also explain why they took the risk - because they were afraid of being caught if they had manipulated the county results already on election night (and would explain the much-higher-than-reported turnout), with the inevitable recount looming, and sought to provide a buffer large enough so that it would not be needed.
However, this scenario also has major caveats - holding together such a conspiracy would be difficult indeed, and also, again, it would be sheer incompetence if moving outside the 0.5% margin was their goal, as the 7500 votes almost exactly do so (and the needle can still shift either way - my analogy of trying to steal an election so you win by 10 votes. Possible, but why?) And finally and most unlikely, people who didn't vote would be recorded as doing so in the databases - something that would be found before long.
In short, something fishy seems likely to be going on. I'm guessing paranoia much more likely than fraud, but who knows?