A new KING-5/Survey USA poll in the Seattle Mayoral Race was released today. This race has Mike McGinn facing off against Joe Mallahan. Just last week, a Washington Poll poll had Mallahan leading McGinn by 44% to 36% with 20% undecided. The results suggested that Mallahan had an 89.9% probability of winning.
Today’s poll of 586 people (taken over the weekend) shows a tighter race with Mallahan leading McGinn by 45% to 43%. A Monte Carlo analysis (methods) consisting of 1,000,000 simulated elections using the proportions and sample size observed in this poll shows Mallahan winning 635,831 times and McGinn winning 352,638 times. Statistically, the results are a tie. But Mallahan has a small edge with a 64.3% probability of winning; McGinn has a 35.7% probability of winning.
The distribution of election outcomes from the simulation says it better than numbers:
The red bars on the left are Mallahan wins, and the blue bars on the right are McGinn wins.
There are a number of possible confounders here that make this extremely close race even more uncertain. Most obvious is the age discrepancy in support which The Stranger’s Eli Sanders points out may lead to some systematic (statistical) bias:
Remember, too, that SurveyUSA only reaches voters with land-lines, and that some of McGinn’s strongest support is among younger voters—who frequently only have cell phones.
Who knows….
Yeah, that can skew the poll’s accuracy, but I haven’t seen the kind of clear message from McGinn that would cause young voters to coalesce around him or get them to turnout. Typically, turnout of young voters is low, unless you have a presidential election with an RFK or Obama running. So, this poll may not be all that far off.
I think, too, when considering potential turnout you should ask yourself if young voters have a dog in this fight. I’m not sure very many of them care about whether a tunnel is built. I doubt they care very much about the run-of-mill city hall issues. This isn’t a race they’d see as affecting them very much.
RR McGinn has had a huge set of youngish volunteers, turned on by his environmental message and experience, for them “Sierra Club” is about as validating as being a ww2 veteran was for older voters in the 1950s 1960s etc. (your set??) and recently the campaign morphed into McGinn as antiestablishment/ neighborhoods/enviro versus mallahan = puppet of power elite. At this point your failure to see the coalescing and turnout that already happened in the primary with McGinn’s appeal t younger voters is basically ignorance of facts. It’s only grown as the financial impact of the tunnel hits home in the McGinn and O’Brien ads and yes, young voters today get municipal finance and know the older voters want to stick THEM with the tunnel overruns, so it’s bad financially as well as a bad tunnel design from a mobility point of view and bad for the environment and it’s a continuation of the power elite’s generational failure to not put transit and the environment first.
My god, there isn’t a bigger kind of issue at that could possibly “generate” more excitement among younger voters in local races. The connection of urban things to the larger environmental issues is “it” — local politics is now about the planet.
The older folks not getting it also pretty motivational, that’s the whole point grandpa. Young people feel that the tide is rising,
and we’d better start swimming or we’ll sink like a stone
for the times — they are a changing.
(Used to be your song, right?)
McGinn will take the 43rd and maybe one other on the north end. the rest will go to Mallahan.
Let’s see, your Monte Carlo has predicted that Burner would win (twice), that I-1033 would fail to get on the ballot and yet you still use it. Either your system is flawed to the 10th degree, or you are flawed to the 20th degree (I am leaning towards the latter).
Nah, IAFF Fireman, it is your understanding of statistics that is flawed.
IAFF “Fireman” @ 5,
“Let’s see, your Monte Carlo has predicted that Burner would win (twice)”
See…here’s your problem, Squirt. You don’t actually research the facts…or you just make shit up (I’m leaning toward the latter).
Fact: In 2006, I was not doing Monte Carlo simulations of polling results. Here is my final poll report in Nov 2006 for the WA-08 race.
Fact: In 2008, the last Monte Carlo analysis I did gave Reichert a 95.2% probability of winning.
“that I-1033 would fail to get on the ballot and yet you still use it.”
Wrong again, Squirt.
Fact: I did no Monte Carlo analyses for projecting whether or not I-1033 would qualify for the ballot.
“Either your system is flawed to the 10th degree, or you are flawed to the 20th degree (I am leaning towards the latter).”
…so says the “genius” who literally just got every fact wrong in his comment. (*Snicker*)
Turns out the problem was his perception of reality (refer to @7), something not at all uncommon with wingnuts.
I predict: Mallahan 51%, McGinn 48%, 1% other.