Unintentionally or not, Pierce County Auditor Pat McCarthy made a mistake by putting her name on a voter survey included with the ballots mailed to voters this month.

But she'll make a bigger one, I think, if she doesn't reverse her decision to wait several days before releasing the full initial results of the county's first ranked-choice voting election Nov. 4.

Note: I asked McCarthy by email today for a response to the concerns I raised here. Unfortunately, her late-afternoon email was garbled due to a formatting error. My former colleagues at the TNT will post her comments as soon as they are received.

[More:]

McCarthy currently plans to release on election night only the totals of first-place rankings candidates receive in the RCV contests. These include races for county executive, County Council, sheriff and assessor.

I don't think it makes any sense for the auditor to arbitrarily release only part of the preliminary results on election night. Voters understand full well that first-night results can change, especially after enduring the aftermath of the 2004 governor's race in Washington.

And because McCarthy herself is a candidate for county executive, she needs to avoid any possible perception that she would be privy to the full results before other candidates and the public.

McCarthy needs to be purer than Caesar's wife at this point. That means releasing all the results, right away.

The Washington Coalition for Open Government (I'm a board member), state Auditor and former Pierce County auditor Brian Sonntag, and RCV proponents agree on that score.

In following discussion, "ballot image file" refers to a computer file that contains all the information from voter ballots used to determine the results of the RCV election. The file includes each voter's rankings of the candidates in each RCV race.

Here's former Republican state Rep. Toby Nixon, president of WashCOG:

If the ballot image file can be made available, then it should be.
Hopefully, it is in the form of a database, rather than (or in addition to)scanned ballot images. The question then is whether TNT or others have software that can process that database and apply the same algorithm used by the auditor to determine the results. There no reason why the algorithm
can't be run against the database every day to produce an interim result.

Here's Brian Sonntag's reaction:

If returns are available .... and being published .... how do you NOT
publish all you have?

McCarthy has said the delay was advised by this was the advice of election officials in San Francisco, who have run several RCV elections now. They said early RCV results could change dramatically as additional ballots come in after election day, causing voter confusion.

Here's a full-length rebuttal by Richard Anderson-Connolly, a University of Puget Sound professor who was instrumental in the effort to get RCV on the county ballot last year:

The decision by the Pierce County Auditor to delay the release of the results of the ranked choice voting counting algorithm until Friday, November 7 is not merely an instance of a lack of openness in government but is also a violation of the Washington Administrative Code.

According to WAC 434-250-020(5) “'Tabulation' means the production of returns of votes cast for candidates.” Because the votes cast for candidates are determined by the RCV counting algorithm the reporting of only first rankings does not satisfy the requirements of tabulation.

It is important to recognize that in an RCV election a vote is not equivalent to a voter’s first preference unless that first preference is either the winner of the race or the second-place finisher. Thus the posting of first choices for eliminated candidates on Tuesday, Wednesday, and Thursday evenings is not "the production of returns of votes cast for candidates" as stated in WAC 434-250-020.

To see that first preferences are not necessarily equal to votes we can assume that on election night the count of first choices gives 20% to candidate A, 35% to candidate B, and 45% to candidate C. The first rankings for B and C are indeed equivalent to votes because A will be eliminated and either B or C will win the race. The first rankings for A, however, are not the votes cast by the voters who ranked A as their first choice. The votes for these voters will their second choices, either B or C, or else exhausted, if a second ranking was not used.

The claim that first preferences are votes before running the algorithm would be equivalent to claiming that a vote in the primary for a candidate that did not make the top two should be counted as a vote in the general election until Friday night.

The auditor's office indeed recognizes that voters may have three rankings but have only one vote. The Pierce County Canvassing Board Policy Manual states, "Based upon established rules the algorithm eliminates losing candidates and transfers the vote [emphasis added] to remaining candidates based upon the voter's first, second, and third ranking" (p. 53). The manual properly refers to the vote in the singular.

WAC 434-250-150 permits the auditor “to modify the requirements of this chapter in order to accommodate the requirements of a ranked choice voting election.” This section can not be used to justify the delay in the proper tabulation of RCV results because it does not permit a delay for an arbitrary reason but rather for one that is a requirement of ranked choice voting. The software purchased from the voting equipment vendor produces a tabulation of the RCV results in a short amount of time. This file can easily be uploaded to the internet along with the plurality results.

The reason for the delay is not a "requirement of a ranked choice voting election" but rather a decision to follow the precedent established by the Department of Elections in San Francisco. A second reason offered for the delay was the possibility of "swings" in the outcomes of the RCV races (Blue Ribbon Review Panel, May 9, 2007). There is no evidence that such "swings" occur more frequently in RCV than plurality races but, regardless of this possibility, changes in the lead have never been used as a reason to delay posting plurality results. In fact neither reason constitutes a proper basis to violate the Washington Administrative Code.

When Amendment 3 was passed by the voters Pat McCarthy said that she would run the best RCV election in the country. The lack of transparency and the refusal to follow Washington State voting regulations will prevent her from realizing this promise.