Croker Sack

"Democracy is the theory that the common people know what they want, and deserve to get it good and hard." — Henry Louis Mencken (1880-1956)

Saturday, March 12, 2005

King County didn't reconcile ballots and voters

King County did nothing resembling a bona fide reconciliation of their ballots and voters prior to certifying their election returns on November 17 (and their subsequent recounts).

Today at Sound Politics Stefan Sharkansky posted copies of the files he received from King County which explain what they did.

The explanation of their reconciliation process demonstrates that any explanation at all for the discrepancy between the number of signatures in the poll book and the number of ballots inserted into the AccuVote machines at the polling places was accepted as a good enough explanation.

Take the example given on page 3 of that 5 page explanation in which the polling place at Bothell Regional Library was found to have 32 more ballots inserted into the AccuVote machines than there were signatures in the poll books. Since no precinct officer reported having seen provisional ballots inserted improperly into the AccuVote machines, the election office employee who canvassed that precinct’s records didn’t note the discrepancy as resulting from “PBAV” or “provisional ballots run through AccuVote.” Instead, the canvasser came to that same conclusion but noted it as 31 “no label” ballots – meaning that there were no provisional ballot envelopes with the stick-on labels containing voter identification information, but nevertheless that had to be where the extra “voterless ballots” came from.

Ironically, if the election office employee doing that canvassing had been more careful, the “variance” between the number of signatures and the number of ballots counted might have been reported as zero rather than one. For Bothell Precinct 01-0254, which used that library polling place, the cover sheet with the canvasser’s notes has been posted at Sound Politics. [Update 3:52 PM -- After looking again, I'm not sure I understand what the canvasser was doing. There were 179 ballots counted on election night -- 177 went into the AccuVote machine at the polling place, and 2 were "add-ons." There were 171 signatures in the poll book. That leaves 8 more ballots than voters, yet the canvasser surmised that only 4 provisional ballots were improperly inserted into the AccuVote machine. It seems that 8 were improperly inserted into the machine at the polling place. Perhaps the canvasser guessed wrong in concluding that only 4 "no label" provisional ballots were improperly inserted into the AccuVote machine. Either that or some other problem with the records at that polling place exists. If only 8 provisional ballots were issued, and if 4 of them were properly sealed into envelopes, then the canvasser's surmise that 4 were improperly put into the AccuVote machine makes sense. Where, then, did the other 4 ballots that were included in the 179 election night count come from? I have put this note in brackets and left my original thoughts alone, since I really don't know what the canvasser was doing with the available numbers -- nothing adds up the way the canvasser indicated it does.] The cover sheet shows that the canvasser found 179 ballots had been cast, but only 171 signatures were recorded in the poll book (“WANDA 171”). The canvasser apparently should have noted +8 for the variance in that precinct, but noted only “+7.” Then, instead of making an “adjustment” to subtract 31 from the total 32 “variance” to arrive at a net “variance” of one for the polling place, the canvasser could have subtracted 32 to arrive at a “variance” of zero.

Of course, the variance between the number of signatures in the poll books and the number of ballots was still 32, but no such variance was reported to the canvassing board.

Instead, the canvassing board was told (to the extent that they read any of the stuff made available to them) that the “variance” for the precincts using the Bothell Regional Library polling place was only one more ballot than the number of voters who signed the poll book.

That is worse than no reconciliation at all, because the canvassing board members would have been misled about the true magnitude of the discrepancy. They would have believed that the number of illegitimate ballots that appeared in the vote count under irregular circumstances (that is, in violation of the laws intended to ensure that our elections are decided by the legal votes cast by eligible voters who vote only once) was much smaller than it actually was.

The “reconciliation” done by King County in essence accepted as sufficient any seemingly plausible explanation for the presence of illegitimate ballots, then “adjusted” their report so that it didn’t show any problem at all with those illegitimate ballots.

The courts are supposed to presume that election officers and canvassing boards do their jobs in a careful and lawful fashion, but what are the courts supposed to do when it is obvious that the election officers didn’t do their jobs carefully, and the canvassing board members (other than, presumably, Dean Logan) were kept in the dark about the magnitude of the problem?

The canvassing board was required to certify that the county’s election returns were a “full, true and correct representation of the votes cast." In a sense, they did – except that the illegitimate nature of many of those votes was hidden from them. In short, their election returns weren’t a “true” statement of the legal votes cast.

It is no wonder, then, that there is a large discrepancy in the number of ballots cast and counted and the number of voters known to have participated in the election. When that discrepancy was first noticed at the end of December by Stefan Sharkansky, the official party line in response was that the process of updating voter registration records to show the date of the last election in which voters participated wasn't supposed to show a match between voters and ballots.

Now, the original records are finally beginning to see the light of day, and it appears that the discrepancy is real -- and that the reason the updating of voter records after the election showed a discrepancy is because there was a discrepancy.


Blogger Iguana said...

Good analysis Micajah. It's startling to me that Logan doesn't seem to feel it is necessary to explain with specificity what happened in these examples. "Human error" is not enough of an explanation.

March 13, 2005 1:57 AM  
Blogger chew_2 said...


I had a hard time understanding the reconciliation spread sheet and notes also. (I do know that the "32 no labels" was for the whole Bothell Library polling place not just for precinct 1-0254. Also the cover of 1-0254 shows a handcount of 180 for signatures, not the 171 wanda count)

However, in general it seems like the total number of ballots without signatures was only 216, which seems pretty reasonable to me.

Do you disagree with this statement from the "Summary of the Reconciliation"?

"Polling places that showed a negative variance have a sum total variance of “-158”, indicating that the sum of provisional ballots issued, add-ons, and the AccuVote machine tabulation was 158 less than the sum total of ballots signed for in these polling places.

Likewise, polling places that showed a positive variance have a sum total variance of 216, indicating that the sum of provisional ballots issued, add-ons, and the AccuVote machine tabulation was 216 more than the sum total of ballots signed for in these polling places. The combined negative and positive variance is 374. Based on 334,185 polling place ballots cast, this represents a variance of .0011, or just over one-tenth of one percent."

Granted this is only for the polling place votes, but I would expect the absentees would show far fewer errors.

March 14, 2005 2:28 PM  
Blogger Micajah said...


My basic problem with the reconciliation done by King County is that they don't attempt to reconcile the number of qualified electors who are known to have voted and the number of ballots included in the vote count. (And, to add insult to injury, they didn't comply with the statute which required their precinct election officers to count the signatures in the poll books and compare that number to the ballots in the ballot box/AccuVote machine.)

Instead, what they do is attempt to account for all the ballots issued and to ignore how those ballots got into the vote count.

The 216 figure is the number of ballots that are in the count despite the absence of any explanation for their origin. In other words, they appeared as if by magic. In short, they were not recorded as having been issued to anyone, eligible voter or not.

The 660 "Adjusted #" is not a true reconciliation of legitimate votes and voters. It is simply a subtraction of the number of ballots in the count which are thought to be provisional ballots improperly inserted into the AccuVote machines at the polling places. If they are correct in their deduction, then those ballots aren't forged copies of ballots -- they are ballots which were issued to people who were not listed in the poll books as voters eligible to be issued regular ballots. They didn't belong in the vote count until those people had been confirmed to be eligible voters.

Subtracting ballots based on a deduction that they had been issued as provisional ballots but voted as though they were regular ballots doesn't reconcile legal votes and voters. It only tends to explain where those illegitimate ballots in the count probably came from.

Knowing where they probably came from doesn't transform them into legitimate ballots. They are still ballots which were not legally cast and counted.

March 14, 2005 5:31 PM  

Post a Comment

<< Home