More on the Folly of SoundPolitics.com

I've been spending the last few weeks monitoring what's going on over at SoundPolitics.com, the rather vituperative right-wing site currently dedicated to getting a revote for so-far losing Washington goobernatorial candidate Dan Rossi. They have been able to focus attention on the problems that occur in every election, but become so acute when there is a close result. I think that's a responsible form of advocacy, and as I keep saying, electoral reform is an issue that has cried out for address since before 2000.

Unfortunately, led by a fellow named Stefan Sharkansky, they seem intent on going beyond mere questioning, and making a series of rash declarations that either had to be immediately rescinded, or that were shown to be wrong and were offhandedly corrected. That they do this repeatedly is what's disturbing. First it was the military ballots going out late, then the vote discrepancy in King County being actually more voters than votes--which it turned out not to be. It seems the recurring problem is that allegations and conclusions are not validated or corroborated in any way before publication. Sharkansky is getting the ear of the national right wing echo chamber, and his lack of professionalism and standards of reportage apparently don't bother the likes of Brit Hume and NRO's The Corner.

So this week the latest breathless allegations emerged: King County is still misrepresenting their voter-to-ballot discrepancies.  What's the angle? According to Stefan, the County failed to account for precincts in which a larger number of voters than ballots existed. Thus, rather than canceling each other out (more voters in one precinct being washed out by more ballots in another precinct), each failure to have one-to-one correspondence would result in two errors.

There's nothing inherently wrong with deciding that a voterless ballot is as bad as a ballotless voter, but there are a couple major problems with his conclusions:

1. Stefan does not have an accurate listing of voters by proper precinct, nor fully descriptive vote counts at the precinct level for all three canvasses.
2. Because of 1), assuming that once you're unable to get things to match then no match must objectively exist, is foolhardy. And regardless of whether it turns out to be an accurate statement in most of the questioned records, it's awfully disingenuous to boldly refer to your gemischte analysis as "definitive."

In speaking with King's Bill Huennekens, it's clear to me that the only valid replication of King's voter discrepancy information is one that utilizes pollbook records at the precinct level. The publicly available files are rolling records, not snapshots of Nov 2. Perhaps thousands have been added, removed and modified since then, particularly with respect to precinct address. So for Stefan to call his results conclusive--much less definitive--is absurd.  Having voters listed as voting without the same number of ballots, seems like a problem. But there are several possible human error answers--right polling place, wrong precinct, for instance. Some stations have multiple precincts to serve, and if you sign the pollbook for the wrong precinct, that will show up in the pollbook discrepancy analysis. But when using just the voter file as Stefan is, that will come up as two errors: Voter A in Precinct X voted, but there's no ballot. And over in Precinct Y, there's a ballot with no eligible voter for it.  Voters could have moved, signed the pollbook and not voted, or be erroneously listed as having voted. For instance, illegal provisional votes are supposed to lead to a striking of the flag from the voter's name that shows he voted. Any failures to remove that flag will create a phantom voter.

Carla at Preemptive Karma has done an excellent job of sleuthing out some of the stated facts from the responsible officials, and while you can't take what they say at face value all the time, they are often able to provide verifiable information that Stefan seems to lack because he has set himself up as adversarial to King's officials.  His specific claims in this latest salvo are not only unverifiable, there doesn't seem to be much point--since the files he has cobbled together are not going to allow real validation of King's numbers with any precision.

So since this seems like a dead horse to beat, let's concentrate more on what seems to be building as the main Rossi contention--more illegal votes than the margin of victory. Forgetting for a moment that this does not appear to be sufficient to lead to annulment of the election result, it's also a weak path to take, IMO. Republicans are jazzed about having identified 737 specific ballots that are bogus.  Of those, 437 were errantly fed provisional ballots (from Pierce and Stevens Counties in addition to King), and 284 who were either felons or dead. Notice how Pierce and Stevens Counties provide 20% of the provisionals, when I know they're not 20% of the total registered voters in those three counties.  That would seem to indicate that King's ratio on that score is not as bad as it may have appeared. Which made Carla and I wonder--just what ARE the numbers and ratios for other counties? The revelations on that score have been spottily reported in the media. So we're going to take a cross-section of Washington counties, and see if we can't get their findings on their own illegal and invalid votes. Stay tuned.

--TJ

Comments

[scottd]

David Goldstein has already worked out provisional ballot error rates for four counties:

http://www.horsesass.org/index.php?p=384

KingCo's results were not remarkable.

Last night at a town hall meeting, Dean Logan disclosed that KingCo workers had identified many of the voters who cast the 348 provisional ballots in King county. This included 250 legally registered voters:

http://seattletimes.nwsource.com/html/localnews/2002163225_logan28m.html

[G Davis]

I'm glad someone has undertaken the daunting task of trying to find out what the other counties numbers look like. Good for you and Carla and whomever else is involved.

I heard a story somewhere (newspaper or testimony perhaps) when this whole mess started about an auditor in one of the small counties that had problem ballots and simply placed a phone call to the voters to have them come correct the ballot. This auditor personnally knew each of the voters so these corrections were simply done.

What sort of effort have sites like SP put into tracking those sorts of actions? What sorts of efforts have they put into the Snohomish County count that is simply a computer tape of vote tallys with no way to verify them at all?

Election reform is absolutely necessary, nationally not just in Washington. The vote is the very basis, the footing for the foundation of our government. Without a clear, concise system doubts will always exist.

Hopefully some of this frenzy will spur our state to make a model for the rest of the nation. It would be worth the angst and money if that comes to be.

Thanks for all the work all of you.

[Torrid]

Thanks for the comments. ScottD, I appreciate the link to Horse's Ass. Luckily for Carla and I, his counties are not the ones that we used for a cross-section analysis, so his are additional rather than redundant. So far I've talked to four counties, and the most notable is Spokane--an unresolved discrepancy of 976, with only about 200,000 voters. That's an error rate more than double King's. And interestingly, it's more voters than ballots, rather than the other way around. I also heard from King spokesperson Bobbi Egan, and she passed along the verification of the 250. It sounded as if they weren't finished checking them, so while I didn't ask, I don't think that means that the other 98 are necessarily INvalid.

GDavis--can you talk further about Snohomish? I spoke to them yesterday I believe; can't remember their numbers (left the papers at the office). Reply here or at our email (alsoalso.also-at-gmail.com) if you can.

[G Davis]

T-The only thing I know about Snohomish County is that they are one of two counties (can't remember the other one, sorry) that voted at least partially via touchscreen. It is my understanding that those two counties asked for and received a waiver in the recounts since they had nothing to recount.

The touchscreen machines produce a cash register type tape as their only verification. So these two counties got a waiver allowing them to simply rerun those tapes through the same machines during the recount process.

In other words, they didn't recount at all...ever.

They may have no method of determining if/verifying that the voters that actually touched the screens were legally registered, double voting, etc. There is nothing produced from the machines except this generic registerlike tape.

I've heard surprisingly little about this. It's one of the problems BlackBox Voting was hollering about in the national election. While BBV is a bit shrill for my tastes, they did bring up this one valuable problem with the *new* systems so taunted by the current admin.

Again, I have only garnered this opinion from what I've read, so I may be completely off base. It, to me however, is certainly worth looking into.

I've got a crew working on my house today, but will see if I can relocate where I saw the info tonight. That might be a starting point for you anyway.

Sorry I can't be of more help.

[G Davis]

Just a couple links to articles on the Snohomish supposed recount. The last link is the memorandum of understanding issed by the SOS office on the subject and goes a bit further into the idea behind allowing Sno and Yakima county to rerun the tapes. This memo makes the tapes a non issue on the surface. What the memo does not address is whether those votes were recorded accurately the first time or if there were any problems with the voters themselves as outlined in the more kneejerk Weekly article linked.

http://seattletimes.nwsource.com/html/snohomishcountynews/2002113871_snorecount09m.html

http://www.seattleweekly.com/features/0504/050126_news_snohomish.php

http://66.102.7.104/search?q=cache:huYrMtAv0pgJ:www.secstate.wa.gov/office/news_docs/Recount/Memorandum%2520of%2520Understanding.pdf+snohomish+county+recount&hl=en

I'll keep my eyes and ears peeled for any other info on the Snohomish situation as it's a major bone of contention with me that any voting method would be so completely unverifiable even to the voter at the moment of voting.

Good luck.