Sunday, January 18, 2009
Presidential percentages (election analysis part 2)
In my first post about the results of last November’s election, the topic was the overwhelming Obama/Biden victory in Seattle. This time, we’ll take a look at turnout and the presidential vote in King County, and set the scene for a more detailed look at its Legislative Districts. First, though, I’ll look at some of the pitfalls and difficulties resulting from choices made by analysts, even in something as (seemingly) simple as reporting the percentage of the vote won by the candidates.
Let’s start by setting the scene geographically, with a description of King County in terms of Legislative Districts. The state of Washington is composed of 49 LDs, defined after the 2000 Census to contain very close to identical populations.
As the state’s largest county, King County contains, in whole or in part, 17 of those Legislative Districts. The main portion of the map on the left displays the 13 LDs that fall entirely within King County—the 5th, 11th, 30th, 33rd, 34th, 36th, 37th, 41st, 43rd, 45th, 46th, 47th, and 48th. As for the four partial-King LDs, the 32nd LD has about 80% of its population in King and 20% in Snohomish. A bit over 1/3 of those living in the 31st Legislative District reside in King, with the remainder in Pierce County. The 1st LD is largely in Snohomish, with only about 1/6 of its voters in King County. Finally, the mountainous northeastern corner of King County holds a tiny portion (under one percent) of the voting population of the 39th Legislative District; you have to look at the inset map at the bottom of the image to see the 39th, as well as the full extent of the 5th, 31st, and 45th LDs.
The Washington Secretary of State and King County Elections agree that there were 1,108,128 registered voters in King County, and that 930,038 of them voted in the November election. That’s an 83.93% turnout—lower than the 84.61% for the state as a whole. If you subtract out King County from the state total, the other 38 counties combined to turn out their voters at a rate almost exactly a full percentage point higher than King’s (84.92%). That might be an argument in favor of all-mail voting, a practice in which King County is joining 37 other counties—every county but Pierce—in 2009.
While the state and the county agree on the above numbers and on the number of votes for the presidential candidates (as do I, from my e-canvass analysis), they don’t agree on the percentages earned by the candidates. To understand why they differ, it’s necessary to look at all categories of votes. In addition to the votes cast for one of the candidates on the ballot, we have to decide how to deal with:
- Blanks—where voters did not select any of the candidates in the race
- Overvotes—where voters (mistakenly, I assume) chose more than one candidate in the race
- Write-ins—where voters cast their ballots for someone other than the names on the ballot
| Category of vote | Ballots | SoS % | KCE % | my % |
|---|---|---|---|---|
| Obama/Biden | 648,230 | 70.30% | 69.97% | 69.97% |
| McCain/Palin | 259,716 | 28.17% | 28.03% | 28.03% |
| Nader/Gonzalez | 6,982 | 0.75% | 0.75% | 0.75% |
| Barr/Root | 4,058 | 0.44% | 0.44% | 0.44% |
| Baldwin/Castle | 1,471 | 0.16% | 0.16% | 0.16% |
| McKinney/Clemente | 1,199 | 0.13% | 0.13% | 0.13% |
| Harris/Kennedy | 217 | 0.02% | 0.02% | 0.02% |
| La Riva/Puryear | 159 | 0.02% | 0.02% | 0.02% |
| (Blank) | 2,946 | ---- | ---- | ---- |
| (Overvote) | 635 | ---- | ---- | ---- |
| (Write-in) | 4,425 | ---- | 0.48% | 0.48% |
| Denominator | 930,038 | 922,032 | 930,038 | 926,457 |
The Secretary of State ignores write-in votes, whereas King County and I include them as valid ballots. All of us drop the blank ballots and overvotes—none of which can be considered as valid votes in the race—from our calculations. Where King County Elections errs, in my opinion, is in their failure to present the specific denominators of each race on the ballot. Although they calculate percentages using what I believe to be the appropriate denominator, that value isn’t displayed in their tables. Instead, they display the total number of ballots (930,038) as if it is the denominator for every statewide race, even though each race actually ended up with a different denominator after removing the blanks and overvotes.
Obviously, the differences between the the SoS percentages and mine are very small. In the overall scheme of things, the outcome of the presidential race in King County looks the same no matter which way you go—Obama/Biden kicked ass, with (rounded off) 70% of the votes to McCain/Palin’s 28%. Washington’s largest county was also its most Democratic at the presidential level, coming in just slightly ahead of San Juan County. At least, that’s the case if you believe the Secretary of State; without easy access to San Juan’s count of write-ins, King’s lead can’t be confirmed. Even the 2/3 of King County that isn’t Seattle went very solidly for Obama. While, as I reported last month, Obama received 84.3% of the vote here in the Emerald City, his percentage in the rest of the county came to 62.0%. Only two counties, Jefferson and the aforementioned San Juan, beat non-Seattle King County.
In upcoming election analyses, I’ll examine the LDs within King County, and then move on to the gubernatorial race and comparisons with 2004. Stay tuned!
