We have no conclusive explanation for it, and neither did Peter Callaghan of the Tacoma News Tribune, but the overall trend was decisive enough to be real and clear: The very last voters in the Washington primary election were Republican voters, by a big margin.
The stats, from Callaghan's column today: "On the day after the election, Gov. Chris Gregoire held a 4.1 percentage-point lead over Republican challenger Dino Rossi. But when the final county tallies were in, Gregoire’s advantage had narrowed to just 1.9 percent. Attorney General Rob McKenna’s lead over Democratic challenger John Ladenburg grew from 12 points to 14; U.S. Rep. Dave Reichert’s lead over Democrat Darcy Burner went from 3 points to 4 points." And so on - you'll find the pattern repeated elsewhere.
Explaining this, or working through whatever its significance is, is more difficult than saying it's happened. A Rossi backer suggested that later voters got to know Rossi better and liked what they saw; but that runs into the problems of only a day or two difference in voting time, and into the problem of the voting differential extending across Republican candidacies.
In the last decade over in Idaho, we noted that when that state moved to some-day registration (as opposed to requiring registration in a period that ended several days previously), the Republican percentage rose a notch. Of course, 1992 and 1994 (when this happened) were strongly Republican years anyway for a variety of reasons. But we also found, in precinct breakdowns, that places where same-day registrants were larger portions of the electorate, were the same places where Republicans made their greatest gains.
Exactly what that means isn't at all clear, and your explanation of it may have more to do with what you think of Republicans (or Democrats) than anything else. But there's surely a lesson to be learned in this, if someone can get a handle on it . . .
