Let's revisit Friday's P-I article about King County's groaning discrepancy "Election expert praises King County: Record-keeping called very accurate". This article is being trumpeted by some of the more prolific fabulists from the lunatic-fringe blogs as a "definitive debunking" of the so-called "discrepancy myth", or something.
Unfortunately for the Gregoire fan bloggers who want you to believe that our election system was "a model to the rest of the nation and the world at large" the article has a massive doughnut hole where the custard-filling is supposed to be.
[Former King County Elections Director Bob] Bruce was quoted recently as saying the discrepancy in the 2000 presidential election was less than 20 votes. But yesterday he explained, "We're not talking about crediting. We're talking about the variance between the number of names in the poll book and the number of people who voted," a reflection of actual election-return accuracy.Gregoire's delirious fans read this as conclusive exculpatory evidence. Those of us who inhabit the world of facts see this as the doughnut hole that it is: If this year's ballot/vote credit discrepancy of 1800+ is incommensurate with the 2000 reconciliation discrepancy of 20, then what is the 2004 reconciliation discrepancy that is commensurate with the 2000 number? The article doesn't say. As far as I can tell, King County has never released this number nor has it released any documents with precinct-by-precinct ballot reconcilation.If that type of discrepancy turns up in a precinct-by-precinct reconciliation of the election canvassing process, Logan said, "then we run (returns from) that polling place again" during the canvass to find the error.
Furthermore, if King County did perform a valid precinct-by-precinct ballot reconciliation in each of the three counts, then how did the number of counted ballots increase from 898,238 in the first count, to 898,574 in the machine recount, to 898,633 in the manual recount (the latter is before the 566 "Larry Phillips" ballots that brought the final total to 899,199)?
Punchline: King County's explanations to date for its groaning 1,800+ discrepancy simply don't hold water.
I'll have more on the King County discrepancy story in the coming days. Stay tuned.
Posted by Stefan Sharkansky at February 20, 2005 11:45 AM | Email ThisI see a lot of people talking about the total number of votes changing, but it is the actual number of ballots cast that is important. There are a lot of people that only cast a vote in the presidental race.
You have to look at the actual physical number of ballots and compare to the number of voters. Not the number of votes cast for a particular race.
Posted by: Vince Callaway on February 20, 2005 12:14 PMBruce, who was King County elections superintendent and later the records and elections director for 13 years, until 2002, said that discrepancy wasn't even calculated when he was there because it wasn't necessary, and it "should not be an issue" now.
"We never bothered doing a comparison (of the variance with those of previous elections) because we never needed to," Bruce said. "But I would guess that it would probably have been about the same number" as the 2004 variance.
Bruce was quoted recently as saying the discrepancy in the 2000 presidential election was less than 20 votes. But yesterday he explained, "We're not talking about crediting. We're talking about the variance between the number of names in the poll book and the number of people who voted," a reflection of actual election-return accuracy.
Seattle Times 2/20/05 "Changing Numbers"
State and county elections officials say the numbers were up in the 2004 election because workers did a better job of detecting and fixing mistakes voters made on their ballots — mistakes that prevented those votes from being read properly by machines.
In a sense, they suggest, they've traded one kind of accuracy for another. (what?)
"We in effect bend over backward for the voter," said John Pearson, deputy state-elections superintendent. "There is a trade-off in terms of the numbers not balancing every time [in recounts]. I think the trade-off is worth it."
(my paraphrase)
By using ESP or "psychic linking" the intent of the voter was successfully determined to a larger degree this election.
Reminds you of the husband that voted for his dead wife because he knew who she would want to vote for; doesn't it.
The apparent discrepancy in King County's numbers is between the number of ballots cast and counted and the number of voters known to have participated in the election.
There were apparently 1800+ more ballots in the vote tallying process than there were voters known to have cast ballots.
It's not a discrepancy between votes and voters. It's a discrepancy between ballots and voters.
Follow the links Stefan provided, and you should be able to see that the numbers he put in the entry are the numbers of ballots -- not the number of votes.
For the final King County report, there were 899,199 ballots counted. Included in that number are the 21,297 ballots on which no vote for governor was marked.
Enhancing a ballot or duplicating it in order to make it possible for the tallying machine to "see" the vote doesn't increase the number of ballots. It only increases the number of votes and simultaneously decreases the number of "undervoted" ballots.
Posted by: Micajah on February 20, 2005 02:09 PM"Most Washington counties inspect every ballot before it's counted the first time. King and a handful of others don't.
When the governor's race forced a recount last November, those counties — not surprisingly — accounted for a disproportionate share of changes in the vote totals."
I propose that the problem with ballots and voters is much higher than any of us will ever know.
The vast majority of provisional ballots were counted in 2004. This runs contrary to the experience of most of the other states and their experiences with provisional ballots. Ohio accepted something like 20%.
If the state and the counties cannot find deceased voters, felons voting illegally and voters not having residence in Washington state, how can we be confident that they were able to determine that the vast majority of the provisional ballots cast were valid?
We will probably never know the truth.
Posted by: jaybo on February 20, 2005 03:17 PMProvisional ballots are verified in essentially the same way that absentee ballots are verified.
Recognizing that approximately 350 provisional ballots were improperly inserted into vote tallying machines at the polling places in King County (rather than into signed envelopes for later verification), the ballots were supposed to be rejected if they were cast by people who weren't registered to vote.
No matter what the percentage may have been, those which were accepted were supposedly cast by people who were registered to vote -- but whose names weren't on the poll books at the polling places where they went to vote.
Provisional ballots aren't part of the "donut hole," since they can be matched up with the voters who cast them.
The "donut hole" is the absence of any documented explanation from Dean Logan's crew which demonstrates that the number of regular ballots cast at each polling place matches the number of voters' signatures in the poll books.
Posted by: Micajah on February 20, 2005 03:37 PMLogan has known since 29 or 30 December that there is an apparent shortage of signatures in those poll books.
He has never, to my knowledge, even claimed that the apparent shortage isn't a real shortage.
Instead, he has said that the shortage of signatures might have resulted from the issuance of ballots to people at the polling places even though those people didn't sign the poll books.
To your knowledge, has anyone asked Logan to show what his pre-certification reconciliation efforts revealed?
Has anyone ever asked any member of the canvassing board what pre-certification reconciliation information was provided to them before they certified their election returns as true and accurate statements of the numbers of legitimate votes cast and counted?
Posted by: Micajah on February 20, 2005 03:42 PMDoc is full of himself, based on reading his posts re: LA County voter irregularities and this blog for this topic - just wants to hear himself bloviate and see if he can trip up whoever or push some buttons. I doubt if he sincerely cares about corruption and voter fraud - he may say he does, but look at his arguments - smoke and mirrors (heavy on the smoke - like hemp & THC). His Democrat Socialist buddies will likely get a rude awakening in 2006 election if not before. Look at the big picture - one day at a time.
Posted by: KS on February 20, 2005 03:50 PMThank you for the compliment. I've tried to figure out what the law says (or could be said to say), and sometimes I've hit it right. But, as they used to say where I grew up: Even a blind hog roots up an acorn occasionally. So, I can't take too much credit -- otherwise I'd have to throw out the new hat I just bought.
Posted by: Micajah on February 20, 2005 04:08 PMI watched virtually every aspect of the election in Jefferson County. Incredibly well run. Experienced election staff...all 2 of them with 18,772 ballots and 18,772 voters names (incl the 3 ACP's).
I can feel my anger build whenever I hear about KingCo minimizing this reconciliation process. If they reconcile all the precindt pollbooks....they should be able to reconcile the entire County as these poolbooks should account for all ballots that precindt was issued AND all ballots they received at that precindt (incl. absentees, provisionals etc.)
EVERY BALLOT ISSUED TO THAT PRECINDT AND EVERY BALLOT RECEIVED BY THAT PRECINDT!!!!!!!!!!!
What is so complicated about this Micajah???
The depositions ought to smoke it out.
I wouldn't doubt that Logan & Huennekens had staff change these pollbooks in their fury to reconcile AFTER THE DEADLINE! We'll see.
Posted by: Mr. Cynical on February 20, 2005 04:49 PMYou said,"provisional ballots aren't part of the donut hole since they can be matched with voters".
You have to assume this one thing, somebody filled out all of those ballots, whether you can now find a voter for them or not. I think that I can say with confidence that they did not fill themselves out. therefore, you could say that they are also ballots without a "legal voter".
Again, I will restate what I said in a different way. If provisional ballots were not matched correctly, then they would in fact be ballots without legal voters. I am beginning to wonder what process they used to make this match. Remember, the number of provisional ballots accepted in this state was unusually high.
That should be a red flag in and of itself.
Posted by: jaybo on February 20, 2005 05:13 PMBelieve me, the medical profession would never let you all operate no matter how many texts you read. Never.
Leave the opinions to the experts.
KS you shouldn't double post your diatribes. They appear disingenuous.
Posted by: Doc on February 20, 2005 05:18 PMBut that's where the similarity ends.
The Milwaukee Journal-Sentinel, a paper that endorsed Kerry, has led the push for investigation, even though most of the questions revolve around Democratic precincts in Milwaukee. The city's mayor, the police department, and even the FBI have gotten involved. On Feb. 9 a joint legislative committee voted unanimously for a state audit.
Of the evidence that has turned up, my favorite is the case of "Marquis F. Murff," who registered and voted twice, once using a nonexistant address and once using the address for St. Catherine residence, a home for women since l895.
How the irregularites may have affected the election is not the main issue in Wisconsin; it's the possibility of massive vote fraud in a state once known for good elections. And they've got political blogs there too, but the talk isn't all about which party is the crookedest, or who'll gain politically from the mess. You see a lot of comments about how bad it looks to the rest of the country, or to their kids.
On Wisconsin.
Posted by: George on February 20, 2005 06:10 PMKeep it up, my friend!
Dick...er, Doc Believe me, the medical profession would never let you all operate no matter how many texts you read. Never.
Leave the opinions to the experts.
So why are you commenting?
Posted by: marks on February 20, 2005 07:08 PMSeriously, I doubt that very much. But you must think so, with all the attention you are giving this site, which is your prerogative.
Doc, do you happen to live in Georgia?
You arguments appear disingenuous to me, as you appear to be saying that you are making a house call on the lunatic fringe here at SP to straighten us all out.
Please send your bill to CG, temporarily located at the governor's mansion.
Doc, it may be that the people here are the only ones saying it. Perhaps we are the only ones that aren't a part of the support structure that keeps Sims/CG in place, and perhaps we are ones that have chosen to educate ourselves on the issue.
Gee, voters cast ballots. Workers make sure ballot counts match voter counts, certify, and forward onto HQ. Then, during recounts, we don't see magical mystery ballots appear. Makes sense to me, but as you will attest, I am no expert. Of course, argue against the reconciliations and documenting the results onsite. Leave the loophole open. That'll make things better, particularly if you are a tax happy liberal.
Saw the article in the Times today on elections. Dean Logan doesn't want to stay up late, nor does Bill H. Found out that election workers are getting better at fixing mistakes. (One wonders whether a mistake might be considered to be voting for someone not a Democrat. Surely, that must be fixed, right?) And the funniest damned thing was said repeatedly. They don't want to do it BECAUSE IT WOULD COST MORE MONEY!! Since when was there anyone in public government in King County that wanted to save the taxpayers money? A 6 billion viaduct, a 4 billion suspension bridge on 520, Sound Transit, Monorail? HOT lanes? Fuel surcharges? Levys? MVETs? That, right there was the tipoff that these folks are lying. Ron Sims wanting to save money. How funny!
And just as funny, the Times article focused exclusively on tabulation methods, differences inherent in them, and came up with some margin of error statistics on that basis. To the exclusion, I might add, to the dead, the illegal out of state, 'undocumented alien,' double voters, "found votes," and such. Guess excluding things that are only found here, the Times still says that the margin of error is greater than the margin of victory in this race.
It's amazing to me to see what a low standard people are setting here on the subject of following rules, expecting competence, and expecting a fair election. One wonders to what degree these apologists are willing to accept such things if it goes against their candidate. One suspects they would whine loudly. (I cite, for example, the Gannon 'issue' where no one broke the law, and yet we hear the rising crescendo of weenies whining about it.)
'night all,
Steve
Posted by: Patches Pal on February 20, 2005 08:40 PMThe bars where you get your expert opinions must be real dens of iniquity. But you don't have to spend such a sordid life. Try your downtown library for an hour, read the RCWs and WACs and County Codes (oh, I forgot to ask if you are sufficiently literate, but will give the benefit of the doubt). THERE IT IS! The votes must reconcile at EACH PRECINCT, before they ever go into the seething mess at County level - where the law requires that they also reconcile.
Do your expert opinions come from French lawyers?
Let the other readers decide for themselves if they believe it to be disingenuous. That's my story (or diatribe if you will)and I am sticking to it. Have a nice day.
Posted by: KS on February 20, 2005 09:40 PMSo I'll ask it again, Mr. Goldstein, if the discrepancy can be explained by some simple tallying problem with the pollbooks vs. the crediting process, then how come King County has not come forward with the answers that would quickly lay this matter to rest?
Given how partisan all of King County's actions have been from November to the present, there's no question that if King County had a way to quickly put the matter to rest, Dean Logan would come forward with it just like they did with all of the magical mystery ballots.
Since KC has not come forward with a quick explanation, reasonable Washingtonions (which you will not find at HorsesAss.org) can reasonably conclude that there is no simple answer to the 1800 vote discrepancy.
Oh and, of course Goldstein never answered the question.
Several points.
1) During the week of January 4-6, 2005, King County Elections conducted a partial "reconciliation" of precincts in order to determine why thousands more ballots were counted than the number of credited voters. They (KC Elections) did not do a reconciliation of all precincts. They picked which precincts they would look at.
2) Observers were not allowed to watch this process up close. For example, observers were kept 30 feet away from the crediting "wand" that scanned in the bar codes next to the names in the poll books.
3) King County Elections officials refused to inform the observers of which precincts they were looking at, while the process was taking place.
4) During this process, King County determined that 348 provisional ballots had been improperly fed through AccuVote machines at polling places on election day.
5) During this process, the Elections workers were marking the poll books with new ballot totals, new check marks, and so forth, and were in many cases crossing out "bad" information from the first audit (which had been conducted, and completed, of every polling place shortly after the November election).
6) During this process, King County Elections officials revealed that they had just scanned in some voters (in order to give them credit for voting) even if those voters did not sign the poll book. In other words, in some situations, during the first full week of January, when it appeared that a person had voted--despite the lack of a signature in the poll book--he or she was suddenly and belatedly given credit for voting.
7) The poll books were stored in the "vault" in the Administration building. The "vault" is a lockable room, but the door was left open and the vault was left unguarded for hours at a stretch, with no supervision, despite a sign on the door that indicates that the door must be closed at all times.
8) Weeks later, Mr. Logan of King County Elections revealed that about 250 of the 348 ballots had been verified as having been cast by legitimate voters.
Logan's problem is Logan...
Trying to be a true Dem. Party "team-player".....while believing the accolades other inept County Auditors have showered on him.
Dean was always the "EXPERT" on elections. REALLY??
Lets take a typical Bank reconciliation as an example.
At the end of the day, the individual tellers have to balance each of their "cash drawers" against the total credits and debits for that day. So, what they do is add all of the receipts up and check them against the total cash. As long as they both match, they are good. Specifically, if John Doe came in with a deposit of $100.00, but the teller mistakenly credited his account for $110.00, that's okay. Because on that same day Jane Doe came in to withdraw $100.00 from her checking account but was only given $90.00.
I mean, let's be honest, this happens all the time. I'm sure that we can all remember a time when this type of experience happened to us and be honest with yourself, you really didn't think it was a big deal, did you? As a matter of fact, when the bank manager informed you of their accuracy rate, you felt pretty foolish for even bringing it up.
The following was an attempt at satire, I hope you enjoyed it.
Posted by: jaybo on February 21, 2005 09:26 AMAt that time elections, if any, would be 100% accurate, because there would only be one socialist dictator "running" for office, and everyone would be credited as voting.
Provisional ballots should have been put into provisional envelopes. Many people put their provisional ballots directly into the AccuVote machines while they were in the polling places. Since King County provisional ballots were identical to normal ballots, these were counted and added to the universe of valid ballots. It was discovered later on, of course, that many such ballots were cast by people who should not have voted (for various reasons) and that the invalidity rate of the 348 ballots was almost triple the invalidity rate of provisional ballots that were properly put into provisional envelopes.
The answer to your question about how King County can identify which 250 (approximately) of the 348 ballots were legitimate is complicated, but I will give it a shot:
Say King County discovered that they had 5 more counted ballots than credited voters in precinct X. Perhaps the poll books indicated that the voters failed on five occasions to return a provisional envelope.
Or, perhaps Elections can account for 50 provisional envelopes, but 55 provisional ballots were issued. The 50 could be subtracted, which would leave 5 provisional voters listed in the poll books who presumably fed their ballots into the AccuVote machines. It is then assumed that the discrepancy of five is explained by these five provisional voters who presumably fed their ballots directly into the machines, rather than putting them into the provisional envelopes.
Whatever the method, King County Elections figured out that it happened 348 times (in the precincts they looked at). Incidentally, the observers did not know the methods used to come up with the figure of 348, since they were not allowed to look at the poll books during this process.
That is the easy answer. Now for the more complicated one. The precinct totals change with the migration of provisional ballots during the duplication process. In thousands of cases across King County, provisional ballots were "duplicated" to the precinct where the individual was registered, rather than the precinct where the individual voted.
Therefore, a hypothetical precinct might look like this (bear with me, even a simplified example is complicated):
Let's say that the hand recount shows that there were 90 ballots counted in the precinct, but only 75 voters were credited with voting, for a discrepancy of 15 counted ballots vs. credited voters. In this hypothetical precinct, Rossi received 30 votes; Gregoire received 50 votes; Bennett received 3 votes; and there were 7 others: 3 write-ins, 1 over-vote (multiple votes in one race, which cancels out the vote), and 3 undervotes (the race was left blank).
Poll books show 100 people entered the precinct.
Of the 100, 50 voted with a normal poll ballot (according to the poll book).
Of the 100, 50 voted with a provisional ballot (according to the poll book).
Of the 50 who voted with a provisional ballot, 35 turned in provisional envelopes that contained provisional ballots. Presumably, then, 15 fed their provisional ballots directly into the AccuVote machines.
Of the 35 ballots inside the provisional envelopes, 5 were rejected as invalid (this could happen for a variety of reasons) or they were sent to other counties or states. [Note: I do not know whether King County sends/receives provisional ballots to/from other states. I doubt it. However, I do know that whenever absentee ballots from out of state are dropped off at the polling places, those ballots are mailed to those other states by King County Elections.]
Back to the hypothetical precinct. Of the 35 ballots inside the provisional envelopes, 30 were validated and the voters were credited as having voted.
Of the 30 validated provisional envelopes, 15 belonged in other precincts, so these ballots were duplicated and the new ballots were mixed in with the ballots from the new precinct.
At the same time, of the 30 validated provisional envelopes, 15 belonged in this precinct (for example, a person lost his/her absentee ballot and voted at the polls).
So there is a migration of 15 ballots away from this precinct into other precincts.
Let's say that 10 ballots also came in as provisional ballots from other polling places that were duplicated and credited to voters in this precinct: either from within the county or outside it. (During the recount process, these 10 duplicated ballots were mixed in with the existing ballots.)
Therefore, the net migration is 5 (-15+10=-5) ballots away from this precinct.
To summarize the totals in this hypothetical precinct: the hand recount indicated that there were 90 ballots counted in this precinct, but 100 people walked in to the polling place, and only 75 people were credited with voting in this precinct. Only 5 provisional ballots were rejected. 15 provisional ballots migrated out, and 10 migrated in.
The discrepancy is 15 "mystery ballots" (90 ballots counted minus 75 credited voters). Now, during the reconciliation process, Elections staffers go through the poll book list of provisional voters, and after checking against the registration system, they see that 30 of these provisional voters have been credited with voting. If they add in the 50 normal poll ballots, that amounts to 80 credited voters listed in this precinct poll book. (Due to provisional ballot migration, there are 75 credited voters in this precinct.) Yet 90 ballots were counted.
Assuming that 15 people improperly fed their provisional ballots directly into the AccuVote machines, these numbers make sense, otherwise they do not.
Theoretically, the net migration across all precincts would equal 0, assuming that King County had a net migration of provisional ballots into the county of 0 (which is unlikely). That aside, if the Elections staffers know which 5 provisional ballots were rejected (and I do not know whether or not this is indicated in the computer system) then they can figure out which of the remaining 15 provisional voters presumably fed their ballots directly into the AccuVote machines.
If they do not know which 5 provisional ballots were rejected, then it would be difficult to determine which 15 of the 20 non-credited voters had presumably fed their provisional ballots straight into the AccuVote machines.
Very few precincts would yield numbers in which the voters who improperly fed their provisional ballots into the AccuVote machines could be identified, since it could only happen when no provisional ballots were rejected as invalid (or were sent to other jurisdictions outside the county).
There are other complications, too. Some voters simply left the polling place with their provisional ballots in hand. Still others failed to sign the poll books (although King County credited some voters with voting, even if they did not sign the poll books, as long as it appeared that they had actually voted).
Many ballots were mangled by the counting machines or while they were in storage, and thus had to be duplicated during the machine recount. (Thus, ballot stubs of legitimate voters do not always match up with counted ballots.)
This is just a sampling of the issues involved.
Regardless, the only way to conduct a thorough reconciliation process is to check every precinct. King County did not do this (back in early January).
Posted by: Tim B. on February 21, 2005 05:00 PMI think you have missed my intent here. I may not have as much knollege of the use of provisional ballots as I had thought. First of all, from what I have read here and also in the msm, a provisional ballot is to be used by someone who, 1) has lost their absentee ballot, 2) is voting at a precenct other than their own (for various reasons), 3) is at their correct polling place but the data from registration did not get transferred to the polling place, etc. Is this correct?
I also had read here that many ballots were put through the machines without proper verification. Is this correct?
A provisional ballot is supposed to be sealed in an envelope, with the voter's personal info on the envelope so it can be verified. Is this correct?
In the case where those ballots were fed directly into the machines, there is no way to determine if that ballot was legally cast.
I have read that some envelopes were improperly opened before verification of the data on the envelope and then fed throught the machines. Is this correct?
I agree that election officials can verify NUMBERS when it comes to verifying provisional ballots vs the list of voters, but how can anyone tell which of the those many magical mystery ballots were the ones that were legally cast?
If you have 348 ballots and say "we can verify that 250 of them are legal", no can take a particular ballot and say "this one is a legal ballot" or "this one is not a legal ballot", once they are separated from the envelope that is required. Of the 98 still unacounted for (that we know of), how many were for Gregoire and how many were for Rossi?
My whole point here is that this is still a numbers game by King County and does not satisfy the question of who the real winner is.
Posted by: RobertinTacoma on February 22, 2005 09:13 AMAs for the provisional ballots that voters improperly fed into the AccuVote machines, if King County Elections can figure out which voters did this, then they can determine whether or not those particular voters would have been eligible to vote. Apparently Elections staff were able to do this with the 348 ballots, and King County determined that about 250 of the ballots were cast by people who would have been eligible to vote, and whose ballots would have been accepted (presumably) had they been properly put into provisional envelopes. The remaining ballots would have been rejected.
I do not know the answer to your question about whether or not some provisional envelopes were opened (by Elections staff) and counted prematurely, before the voters were verified. If this occurred, I am not aware of it. As far as I know, the problem was individual voters who simply never put their provisional ballots into provisional envelopes on November 2, and instead, fed them directly into the machines at the polling places. In effect, these ballots were counted before anyone could verify whether the voter was registered, and whether the voter's signature matched the one on file.
I think (but I am not positive) that it is theoretically possible to match up particular ballots with voters, since there should be unique ballot numbers written into the margin of the poll books for each provisional ballot that is issued. But the poll workers may or may not have done this in the 348 cases.
Even if it is possible, I do not know whether or not finding the ballots and checking how the votes went would be legal. In addition, some of these ballots may have been duplicated to different ballots had the originals been damaged in handling or during the recount process (in which case, it would be possible to find the original damaged ballots which should be kept in storage, but it would be a hassle).
Furthermore, at this point, there are still thousands of ballots out there that cannot be paired up with credited voters. These are the "magical mystery ballots" (to use Stefan's phrase).
In my opinion, if King County Elections were to conduct a full reconciliation audit in all precincts, they would find many more cases of provisional ballots that were improperly fed into the AccuVote machines. But they might still have hundreds or thousands of unexplained ballots (in which case, the real winner would never be known).
Posted by: Tim B. on February 22, 2005 10:26 AM