Poll Analysis: McKenna takes the lead

A new Elway poll has Washington state AG Rob McKenna leading former WA-1 Congressman Jay Inslee, 47% to 45%. The poll comes on the heels of a Strategies 360 poll showing the race tied up. The Elway poll surveyed 451 likely voters (4.5 MOE) from 18 Oct to 21 Oct.

A Monte Carlo analysis employing a million simulated elections, based only on the responses to this poll, gives McKenna wins 617,196 times and Inslee wins 369,576 times. The analysis suggests that, if the election was held today, McKenna would win with a 62.5% probability and Inslee would win with a 37.5% probability. Here is the distribution of election outcomes:

ElwayOct

This new poll is one of five recent (October) polls trying to assess this race, and they largely overlap:

StartEndSample%%
PolldatedatesizeMOEInsleeMcKenna
Elway18-Oct21-Oct4514.5 45.0 47.0
Strategies 36017-Oct20-Oct5004.4 46.0 46.0
PPP15-Oct16-Oct574 48.0 42.0
SurveyUSA12-Oct14-Oct5434.3 47.0 44.0
Washington Poll01-Oct16-Oct6443.9 47.1 46.3

If you believe the dynamics of this race have not changed much over the several weeks, then a pooled analysis of the five polls can provide additional evidence of the state of the race.

The pooled sample provides 2,712 “votes”, of which 2,487 are for Inslee or McKenna. Inslee receives 1,267 “votes” (46.7%), and McKenna receives 1,220 “votes” (45.0%).

The Monte Carlo analysis gives Inslee 745,625 wins and McKenna 249,491 wins. The analysis suggests that, for an election held now, Inslee would win with a 74.9% probability and McKenna would win with a 25.1% probability:
FIVEOCTPolls

It’s hard to say which of these two analyses better reflect the dynamics of the race. But, it would be naive to deny that this race has tightened up in the past few weeks.

The Elway poll has a pretty small sample, which means there is more sampling error; but, combined with the Strategies 360 poll, it is perfectly justifiable to suggest that McKenna is really in the lead now.

Comments

  1. 1

    Tea for everyone spews:

    Although I fully expect Inslee to win, even if it takes a few recounts, I and my family voted for McKenna. It is time for some new ideas and new blood in state government, and I feel this election will be a mandate on that. Inslee is part of the DC crowd that could not control spending. He would do no better in Olympia, sidestepping the very citizens he is sworn to represent, for his own and his big government ways. I will be there Tuesday to help cheer and congratulate him when he upsets the democratic wasteland in this state.

  2. 2

    spews:

    Gregoire benefitted from Obama’s coattails four years ago. Her own campaign was surprised at the size of her margin of victory.

    Perhaps this is just the pendulum swinging the other way, as Obama struggles with message and with the reality that even if the election is framed as a choice between candidates, rather than as referendum on him, he still doesn’t come off smelling like a rose.

    Obama might still win. But there likely won’t be much downballot effect, except the type he doesn’t want (see Virginia’s Senate race in light of Obama’s Navy comments on Monday evening).

    Inslee’s on his own, and he’ll need to differentiate himself from the way people are beginning to perceive a certain Democrat in the White House.

  3. 3

    spews:

    Bob,

    Ummm…REALITY CHECK.

    First, Obama is going to win in Washington, most likely by double digits. Do you seriously believe there will be no positive down-ballot effect from Obama? Your assertion seems childishly naive.

    Secondly, Obama still maintains a consistent lead over Romney in the state head-to-head poll analysis. Contrary to Romney’s campaign rhetoric, his campaign does NOT have the momentum. The shift we saw after the first debate dried up for Romney, and the race has been relatively stable for weeks—with Obama ahead.

    Tell you what, I’ll try to get a new analysis out this morning, and you try to firm up your grasp of reality….

  4. 4

    db spews:

    I don’t think Inslee has run a very good campaign. He was less known than McKenna, and needed more money early on to broadcast a better biographical ad than he did. I have “liberal” friends who are voting for McKenna, so I wouldn’t be surprised if he wins.

    Will it make a difference in how our state is run if McKenna wins? I’m not sure – the legislature is still going to be split closely between the two parties. Our budget is still a mess, and probably will be no matter who wins. The interesting thing is – McKenna could raise taxes to help fund what we need, but I can’t see Inslee ever raising taxes because like most Democrats, he’s been bullied into submission and is too afraid.

    (And I’ve already cast my vote – and I voted for Inslee, just to give you perspective)

  5. 5

    spews:

    Darryl @ 3

    Do you seriously believe there will be no positive down-ballot effect from Obama?

    This is a misstatement of what I wrote. I said @ 2 there “likely won’t be much down-ballot effect”. You perverted that into an allegation that I said there would be none at all.

    I compared my suspicion about 2012 to what happened in 2008. The 2008 gubernatorial result was a surprise to the Gregoire campaign and is generally considered to have resulted from Obama’s very long coattails that year.

    There probably won’t be positive surprises of that nature this year. Go ahead and do your poll analysis, use the NH poll that has Obama + 9 and the OH poll that has Obama up 5 with a D +9 sample, not to mention the polls that use 23% of respondents who already voted when the published rolls say the real number is 11-15% statewide. Not much I can do other than to point out the flaws in the poll data as I see them, and wait 12 more days to learn what’s really happening.

    Here’s my question to you, though: If Romney doesn’t have momentum in OH, how come the OH polls are tightening even as the D + spread in those polls is increasing?

  6. 6

    Liberal Scientist is a Dirty Fucking Socialist Hippie spews:

    @3,5
    Bob, all day long, every day, multiple times per day you spike the ball on the 40th and dance around and taunt people – it’s your MO – we assume you intend this behavior and it’s not some sort of weird, uncontrollable tic.

    It’s startling, therefore, that you whine like a little child when a sane, reasonable, logical person like Darryl calls you out – mildly – on your bullshit.

    Grow the fuck up.

  7. 7

    spews:

    Brown-noser @ 6:

    I’ll take my chances with Darryl. It’s nice that you have his back, and I’m sure he appreciates it, Lib Pinocchio. You should be standing when you do that, though.

  8. 10

    Liberal Scientist is a Dirty Fucking Socialist Hippie spews:

    @7-8-9
    Nice to see someone fellating Dr. Bob besides himself. You two ever consider getting a room?

  9. 11

    spews:

    Bob @ 5,

    “This is a misstatement of what I wrote. I said @ 2 there “likely won’t be much down-ballot effect”. You perverted that into an allegation that I said there would be none at all.”

    WHAT.THE. FUCK. Bob???? Now you are lying about your own words? You literally took your own words out of context. Here is what you said:

    But there likely won’t be much downballot effect, except the type he doesn’t want….

    Inslee’s on his own, and he’ll need to differentiate himself from the way people are beginning to perceive a certain Democrat in the White House.

    You are not just proposing Obama won’t have positive downballot effects, you are saying the downballot effects will be NEGATIVE. Hence, my statement about “no positive down-ballot effect.”

    You are still competing for the title of the worst amateur propagandist IN THE WORLD!!!!

    Regarding your whining about what polls I do and do not include: I have made it clear that I will include ALL polls that meet the published inclusion criteria. That you find it necessary to pick and choose the polls that make you uncomfortable is quite telling about your ability to think rationally about this race…or anything else, for that matter.

    “Here’s my question to you, though: If Romney doesn’t have momentum in OH, how come the OH polls are tightening even as the D + spread in those polls is increasing?”

    If that were true, the most obvious explanation would be that the race was tightening in Ohio.

    But, there is almost no evidence that the race in Ohio is tightening over the past month. In fact, in a few minutes the analysis I did this morning will be posted. It will include a graph showing the polls in Ohio over the past five months (including five new polls in the past three days).

    But, while you are waiting, consider this:

    1. In the past 10 days there have been 11 Ohio polls released. Two are ties, one is essentially a tie at R+0.02%, and Obama leads in the remaining eight.

    2. In October there have been 23 Ohio polls. In all those polls, Romney has only lead in 3: by R+0.2%, R+0.8%, R+1%, R+1%. The two R+1% polls were in early October.

    3. In October, Obama has led in 17 polls, all by O+1% or greater. (There were 3 ties.)

    4. In October, Obama has gotten a +5% to +6% about once a week (details: Oct 6-7 O+6%, Oct 12-13 O+5%, Oct 17-20 O+5%, Oct 22-23 O+5.7%) .

    5. In the first 10 days of October, the weighted average of the polls was very nearly a tie. The past 10 days gives a weighted average of O+2.4%.

  10. 13

    Liberal Scientist is a Dirty Fucking Socialist Hippie spews:

    @12

    You’re so manly. Bob can’t go a day without referencing my kids, now you’re insulting my wife.

    Stay classy, Max and Bob.

    I can only assume that your unhinged and completely low-life response to my posts means that you have nothing intelligent to say, and you’re essentially conceding defeat.

    Not that that’s surprising.

  11. 14

    Liberal Scientist is a Dirty Fucking Socialist Hippie spews:

    @11

    OUCH!

    That’s gonna leave marks.

    Bob, can you order your own CT scan? You might want to check for internal bleeding.

  12. 15

    spews:

    Darryl @ 11:

    Thanks in advance for your work on that.

    There’s a National Review piece on Ohio polls today. You probably don’t care so I won’t link.

    I made no suggestion that you exclude any OH or NH poll whatsoever. All I’m doing is looking at the data.

    Do I think it would be useful if you were to comment on turnout models this year vs. 2008? Yes. Do I think it would be useful if you would comment on the potential effect of oversampling of people who claim to have voted but in all likelihood have not? Yes. Would I have less to comment about if stuff like this were addressed in the thread. Yes.

    Chuck Todd was on MSNBC today claiming OH is much tighter than polls indicate. Does this mean anything other than MSNBC trying to gin up controversy? I have no real idea.

  13. 16

    kim jong chillin spews:

    @13

    Oh go fuck yourself you hypocritical crybaby.

    I suupose the intelligent posts you have been leaving like “you two should get room” are somehow classy.

    Don’t cry lib crybaby, I will make sure to give your old lady fare for the cab ride home..along with a huge smile on her face.

  14. 17

    spews:

    Bob,

    “There’s a National Review piece on Ohio polls today. You probably don’t care so I won’t link.”

    I’ve seen it. It isn’t worth the paper its not written on. Unless, of course, like the author you are grasping at straws trying to convince your self that Romney is really doing okay.

    “Do I think it would be useful if you were to comment on turnout models this year vs. 2008? Yes.”

    I disagree. Turnout is important in elections, but there is no reason to think that turnout will result in election results that are substantially different from the polls.

    In any case, there is no objective way to incorporate it into a pure evidence-based model. I’ll also point out that the model was pretty fucking accurate in 2008–in a year where turnout was pretty unusual and unbalanced.

    “Do I think it would be useful if you would comment on the potential effect of oversampling of people who claim to have voted but in all likelihood have not? Yes.”

    Yawn!!! There are dozens of things that I could comment on that would be more interesting—robopolls v. live interviewers, the bias of no-cell polls, the different methods pollsters use to assay “likely voters”, pollster house effects, the problem with live interview polls that make no provisions for Spanish speakers, etc, etc, etc.

    My strategy is to gloss over ALL of these defects, under the assumption that every pollster is trying to get the most accurate possible result. Given enough polls, the errors, biases, and minor defects will balance, and my mean will be correct (even if slightly over-dispersed).

    People that over-analyze individual polls looking for reasons to ignore unfavorable results are usually engaging in self-delusion. (No names mentioned here, BOB).

    “Would I have less to comment about if stuff like this were addressed in the thread. Yes.”

    You are not “commenting” you are babbling about things you scarcely understand, that are largely fed to you by propagandists, and offer minimal to no value in the comment threads.

    “Chuck Todd was on MSNBC today claiming OH is much tighter than polls indicate.”

    I’m sure he consulted Nancy Reagan’s astrologer to find better data than polls for that.

    “Does this mean anything other than MSNBC trying to gin up controversy? I have no real idea.”

    MSNBC is a business. They appreciate you watching.