With less than 24 hours left to go in the King County executive race between Susan Hutchison and Dow Constantine, the latest SurveyUSA poll commissioned by KING 5 showed Susan Hutchison trailing Dow Constantine, 53-43. A more detailed cross-tab of the responses, collected between October 30 and November 1 from 614 likely and actual voters in King County, can be viewed here on SurveyUSA’s Web site. SurveyUSA claims a plus or minus four percent margin of error in its method.

The Constantine campaign today is attempting to use the SurveyUSA poll  –  and the recent Washington Poll — in the same way the hawkers of weight loss miracles use before and after shots as “evidence” of cause and effect. If the poll does end up reflecting anything close to the final vote count, it will be proof positive for the Democrats that an avalanche of special-interest funded attack mail and negative ads has been worth the dollars funneled by liberal causes.

But the SurveyUSA and Washington Poll surveys should not be placed parallel to earlier polling, or even the data released previously by SurveyUSA. In past samples, only the single issue of candidate preference was put before voters, while in this eleventh hour snapshot the idea of party affiliation was the subject of two out of the four questions. According to SurveyUSA, the four questions asked by its callers were:

1. If the election for King County Executive were today, would you vote for ... (choices rotated) Dow Constantine? Susan Hutchison?
2. Do you view Susan Hutchison as a Republican? A Democrat? Or an Independent?
3. When you think about voting for county executive, how important is it for you to know the candidate's party affiliation? Very important? Somewhat important? Not very important? Or not at all important?
4. When did you decide who you would vote for in the general election for mayor?

Although questions are rotated (not asked in the same order to every person surveyed), the accuracy and usefulness of polling can become diffused and depreciated when the overall scope is diversified as appears to be the case in this latest SurveyUSA effort.

It is worth noting that The Washington Poll released last week that showed Constantine taking a large lead also asked multiple questions including those focused on other races and ballot issues, and was difficult to put much faith in because of the results came from oversampling King County within a statewide sample. Jim Miller raises some interesting points to ponder about the oversampling claims of The Washington Poll in his prediction post at Sound Politics.

The presence of the fourth question concerning the mayor’s race (we are left to assume it is the Seattle race between Mike McGinn and Joe Mallahan the survey implies), and the comparative sample sizes between it and the headline question about which candidate voters prefer to be the next County executive, could be a reason to view the most recent SurveyUSA results with skepticism. While 614 responses were used to report a 53-43 prognostication of a Constantine victory, the sample size in the question concerning the mayor’s race was 526. The ratio of Seattle’s voting population to that of the surrounding county is roughly 1 to 2; Seattle voters are approximately one-third of the total in King County. If the sample size used for the mayor’s question was a subset of a larger group of responses that was demographically representative, we would expect to see the sample used for the first question be closer to 1,000, not just north of 600. SurveyUSA confirmed by e-mail that the smaller sample was taken from within the 614 used for the Hutchison-Constantine choice question, indicating a large Seattle bias in the polling.

Of course, the ultimate measure of the poll’s effectiveness will be the results of the actual vote count. As of 8:00 p.m. Monday evening, the King County Elections Office was reporting that only 24 percent of all ballots had been received. (You can check daily statistics here.) Both candidates will be getting out the vote in earnest from sunup to sundown tomorrow to mop up the remainder of the 56 percent participation expected by County election officials. This is the first year in which King County elections are conducted entirely by mail-in voting. All ballots must be postmarked by tomorrow (November 3^rd) in order to be counted.

For the address and hours of your nearby post office, check the USPS Web site. Ballots can also be dropped off several 24-hour drop boxes in locations around the county. To find the one nearest you, check this page on the King County Elections site.

If you haven’t already voted make absolutely sure that you do and encourage friends, family, and co-workers to do so as well. Exercise your right to choose King County’s direction for the future.