Friday, December 26, 2008

Obama sweeps Seattle (election analysis part 1)

In my post-election post, I mentioned that I wanted to add to my early observations about the results on November 4.  Now that I’ve downloaded the King County precinct data from the King County Elections website, and begun processing the voluminous information therein, I’m ready to start the more detailed look at the election.

The “e-canvass” data file created by King County is, in a word, huge.  1,570,184 records huge.  That’s due to the level of detail in the data records.  Even an uncontested race, such as House Position 1 in my own Legislative District, generates 16 records for each precinct in the LD:

• 1 for the count of Registered Voters in the precinct
• 3 (Absentee, Polling, and Total) for Times Counted, the number of ballots cast in the precinct
• 3 (Absentee, Polling, and Total) for Jamie Pedersen, the one and only candidate on the ballot
• 3 (Absentee, Polling, and Total) for Times Blank Voted, the number of ballots left blank for this race
• 3 (Absentee, Polling, and Total) for Times Over Voted, the number of ballots with more than one marked choice (in this case, both Jamie and a write-in)
• 3 (Absentee, Polling, and Total) for Write-in, the number of write-in votes cast in the race

Every additional candidate adds another three records to the total for the race.  The presidential race, with eight tickets on the ballot, generates 37 records ... in every single one of the 2526 precincts in King County.  That multiplies out to over 93,000 records just from the presidential election.  And each ballot in the county covered 30-some additional offices, measures, initiatives, amendments, and levies.

Large as it is, the e-canvass dataset is only about three-quarters the size of the equivalent 2004 file.  That’s largely due to the ridiculous “top two” primary system we’re now saddled with, which limits all races except the presidential race to two candidates on the general election ballot.  Also, King County now has about three percent fewer precincts than it did four years ago.  That’s an issue I’ll be returning to in a future post.

In light of the national outcome of the 2008 general election, the results here in King County—particularly in the city of Seattle and the six Legislative Districts that fall at least partially within it—aren’t all that interesting.  This is an overwhelmingly Democratic city, in a definitively Democratic county.  Still, we do have some interesting issues to discuss.  Let me start by correcting a statement I made in the post before this one—contrary to my (and Mayor Nickels’s) belief, Dubya actually won two precincts in Seattle in 2004.  We were both aware of Broadmoor, the gated golf course community where Kerry was soundly defeated 57%-42%, but we’d missed the close defeat (51%-48%) in nearby Madison Park.

No such blots on the city in 2008.  Obama got 53% of the vote in Broadmoor to 46% for McCain, and he breezed to a 60%-39% margin in Madison Park.  I’ll offer a more detailed look at the Seattle presidential vote in a future post, but at this point let’s just whet your appetite with the citywide totals:

As I continue to examine the King County election results, I’ll hit a number of themes.  At the moment, I plan to present the above presidential results broken out by Legislative District.  The city of Seattle contains the entirety of the 36th, 43rd, and 46th LDs, most of the 34th and 37th LDs, and a small portion of the 11th LD.  Another topic of interest is registration and turnout, the principal targets of the Obama and Combined campaigns this year.  For this topic, we’ll also want to look at the last presidential cycle in 2004.  After all, you can’t say much about how good, bad, or indifferent your turnout is unless you have something else to compare it with.

Of course, we weren’t voting only for a president this year.  After the 2004 cliffhanger, the gubernatorial rematch between Chris Gregoire and Dino Rossi also drew great interest.  It turned out vastly better (and much, much quicker!) than last time, and it’s quite apparent that the Governor’s share of the vote increased all over the state.  In addition to the improvement in her vote percentage, I’ll explore the possibility that fewer voters opted to sit out the governor’s race this time around.  It’s little more than a surmise on my part, but I recollect that there was widespread dissatisfaction with the choices we had been presented four years ago, which could have translated into a relatively high proportion of voters skipping that part of the ballot.  It may not be possible to separate that effect from the possible opposite influence of the influx of new voters whose only real concern was voting for Barack Obama, but I do want to explore it.

I’d also like to look at LD and precinct-level patterns in the 7th Congressional District, which means adding portions of the 32nd LD, just north of Seattle, to the analysis.  Did Congressman McDermott lose any precincts to his hapless Republican “opponent”?  I also want to look at House races nationwide, to determine whether Jim really did draw more votes than every other winner save one.  Even if he didn’t, he’s certainly right there at the very top of the heap.  I also intend to look at what was surely the most interesting state legislative race in King County, the selection of a successor to Representative Helen Sommers in Position 1 in the 36th District.

Additional avenues for analyses may present themselves as I continue to look at these King County precinct numbers.  More to come!