So we're in a strange limbo now, between what should happen -- the SoS certifies the result on the 30th, Gregoire is governor-elect, she's inaugurated on January 12 -- and what the GOP may do to gum up the works and delay the inevitable.
Rather than a full examination of the overall results, this report is something of a miniature. I'll concentrate largely on the disputed King County ballots. For the "big picture", refer to my report from last Wednesday or some of my earlier diaries.
More material below the fold...
But there is much additional interesting information to be gleaned from those ballots, a look inside the counting and canvassing process. For that, we have to look somewhere other than the Secretary of State's website. This time, the really useful data were reported by the King County elections office. The bottom part of the linked page gives full detail on the disposition of the newly-counted ballots.
Recall that the only known characteristics of the 735 ballots in this collection were that a) they were absentee ballots, b) they were postmarked on or before November 2, and c) no digitized signatures were available to match against the ballots.
Our first observation from King County's report on the disposition of the ballots is that only 566 of them were found to be valid, countable ballots. Nothing is said about the reasons for rejecting the other 169 ballots. However, this appears to be an extremely high number of invalid ballots -- fully 23% of them were rejected from the count. Now, as we'll demonstrate in a moment, King County does not report on rejected ballots in its tallies. Thus, except here, we cannot know how many ballots were submitted but didn't make it into the denominator of Ballots Counted. It couldn't possibly be anywhere near what we see in this group of ballots -- the county's turnout was 83%, so even if (implausibly) every single registered voter tried to cast a ballot, no more than 17% of them could possibly have been found invalid.
Thursday's King County Ballots Counted total is 566 higher than Wednesday's, indicating that King includes write-ins, undervotes (aka blanks), and overvotes in that quantity. I haven't studied county reporting criteria, but I think it's not impossible that other counties might differ in the categories of ballots included in that denominator. For instance, they might not include overvotes as valid, countable ballots.
Among those 566 valid ballots, the three named candidates drew 528 votes. Libertarian Ruth Bennett got more than twice her expected number, while the percentages for Rossi and Gregoire were lower than expected from their overall county results. Valid non-candidate votes came to 6.7% of the valid ballots, far more than the 2.5% seen in the overall King County results, and above even the highest such proportion in any county (Clark County, at 5.5%). Fully 6.2% of the 566 valid ballots displayed no choice at all in the governor's race.
Gregoire's vote count wasn't significantly lower than expectation based on overall KC results, but the differences for Rossi (low), Bennett (very high), and not-candidate (extremely high) were significant.
One point about the overall manual recount...
Soon after the King County results came in, the Republicans began screaming about a post-certification change in the vote count in Thurston County, implying that if they could add a Gregoire vote after their totals were certified, then strong-Rossi counties should be able to do the same thing.
Well, I don't know the story behind the Thurston situation. I've assumed all along that the SoS website only reported certified results -- that's why they never showed the Wednesday King County totals, with Gregoire leading by 10. I rechecked my day-by-day snapshots from the hand recount, and indeed Thurston County showed Gregoire gaining 15 new votes on December 15th, and then changing it to +16 on December 20.
But that's not the only change in a candidate's votes in a county. Jefferson County reported Rossi at +5 on December 9, at +6 on December 10 and thereafter. On those same dates, Pacific County reported no new Bennett votes, then +1 for her. In Whitman County's initial report on December 14, Gregoire added 4 new votes; on December 16, that count became +7.
The story isn't over yet. More to come, no doubt, in the way of frivolous Republican lawsuits (or at least frivolous threats of frivolous lawsuits). And, once I've had a chance to catch my breath, I plan to draw some bigger conclusions about what the recount process shows us here in Washington and across the nation.