The latest Elway Poll shows Sen. Patty Murray with a comfortable lead over all rivals.
Last week KING-5/SurveyUSA had Dino Rossi leading incumbent Democratic Sen. Patty Murray by a stunning 52-42 percent, a margin both camps dismissed as an outlier. Well, now that the latest Elway Poll has Murray leading Rossi by an even more impressive 51-34 percent margin, it’s easy to understand the insiders’ reluctance to accept SurveyUSA numbers at face value.
SurveyUSA had Murray only garnering 73% of both self-identified Democrats and Liberals, and a paltry 46% of metro Seattle voters… unimaginably low numbers come November. By comparison, Elway found Murray enjoying the support of 92% of Democrats, and 73% of Seattleites.
Further, Elway examines Murray’s job performance ratings, a number often looked to as an indicator of electoral strength, and while incumbents never want to come in under 50%, Elway points out that “Murray’s ratings have historically been mediocre,” and that her current 48% rating falls only slightly below her 17-year average. So Republicans shouldn’t take much encouragement from that either.
What does this all mean for Rossi, who was widely expected to announce his candidacy by the end of April, and who we know to have had lengthy meetings and conversations with consultants and fundraisers in recent weeks? Well, there’s simply no strong evidence that Murray is quite as vulnerable as Republicans would wish her to be. Sure, this is just one poll amongst many, and the election is still a long ways off, but as Elway concludes:
The world is going to turn a few times before voters actually cast ballots, and campaigns make a difference. But no matter how these findings are sliced, Patty Murray appears to be in a formidable position.
No doubt Rossi is still getting a lot of encouragement to enter the race from the NRSC and the Republican consultancy class, even in the face of these daunting numbers, so I hope he takes seriously a bit of free advice from someone with nothing to gain from his decision one way or the other. Nobody has ever profited from underestimating Patty Murray… except, you know, the dozens of pollsters, fundraisers, consultants, media buyers and other political professionals who have made millions off previous failed attempts to unseat her, and who would make millions more off of your campaign.
Remember, politics is just as much a business as real estate, so don’t trust their salesmanship any more than you would trust your own.
Update: [Darryl] I’ve done some further analyses here, including a combined analysis of the two polls today and last week’s SurveyUSA poll.
There is a new Rassy poll out today, too, and it shows Murray beating Rossi. I think the Elway poll probably gives a more accurate picture of the current state of the race and I suspect that Rossi will soon announce that he has decided against challenging Murray.
His announcement will be good news for everybody except Rob McKenna.
what does kemper think?
Rossi will decline to challenge Murray because of the liberal State Supreme Court.
What makes Stu’s poll the “legitimate” one and SurveyUSA’s the outlier? Why couldn’t it be the other way around?
Pete @4,
Because the two sides have unpublished internal polls that don’t match the SurveyUSA results.
Could it be due to the track record of the election outcomes in this state?
The only “poll” that matters.
Which is why Reaganism is DEAD!
@4 . . . did some research.
Stu polled 405 registered voters. Stu also uses live callers.
SurveyUSA polled 517 likely voters (note that the Democrat source who tipped off the Slog claimed it was a 1-night survey with a “small sample,” but it was a 4-night survey with and 517 people is probably slightly-above average). SurveyUSA, like Rasmussen, uses IVR. SurveyUSA’s sampling methodology is also unorthodox. Most outfits polling likely voters develop a “sample space” of likely voters and then randomly select names from that group to survey. SurveyUSA appears to have developed a sample space of people (not even registered voters), interviewed 850 of them, and then decided that 517 of those interviewed were likely voters. I don’t know which method is more accurate, though it’s interesting that no other firm (that I know of) does it this way.
If we were closer to the election, the likely voter poll would be a better one to use.
I think the CW right now is that Republicans are more “energized” than Democrats, but there’s still plenty of things that could unfold before November to change this. So if a pollster builds a turnout model based on current CW, it’s just as likely that model will be accurate as it will be inaccurate. So there’s an argument for the registered voter poll.
All that said, the most informative way to look at this would be to not dismiss either poll. The SUSA poll is an idea of what would happen if the election were held today. However, the election is not going to be held today. In six months, Sen. Murray can campaign to swing voters and get ready to turn out plenty of Democrats, and exogenous events could occur that drive swing voters to Sen. Murray, depress the number of Republicans planning on voting, or increase the number of Democrats planning on voting. If all of these things happen, then Stu’s poll would probably be right on the money.
SurveyUSA sometimes has poll results that are way off (and they appear to get pulled off their website too at some point). I’m not sure what they’re doing, but I’ve started to take their results with very large grains of salt.
Two points:
a) Elway has long been the best pollster in Washington. It’s WA’s equivalent of California’s Field Poll.
b) This poll still understates the support Patty will garner in Seattle when the real voting starts. 73% is well below what she’ll get in Seattle. If you look at WA-07 (generally equivalent to the city of Seattle) in previous cycles, you’ll see that Patty got 78.4% against Nethercutt (and Chris Gregoire got 72.5% against Rossi) in 2004. And that Maria Cantwell got 79.3% against Mike! McGavick in 2006. And that Gregoire got 77.2% in the 2008 rematch against Rossi. Add at least 5% to Patty’s percentage in Seattle.
Goldy, the Slog article about the SUSA poll said both sides’ internal numbers show it tight, not Murray with a big lead. That would seem to make them both outliers, which is what I was getting at in my question. My dissertation @7 sorts it out, I think.
Moving away from margins of error and polling methodology, my instinct says that the Elway Poll is the trustworthy one here. I guess my question was aimed at finding out if you were also basing your conclusion intuition, or if you had some kind of quantitative reason.
Also, do you know what Stu’s track record is? Everybody around here seems to respect him, and I’m not intending to impugn him at all. But oftentimes he seems to produce results that are somewhat outside of the range of the other polls. I’m wondering if anybody has kept track of his polls from October-ish and compared them to the actual results.
Pete @10,
I didn’t mean to imply that we should assume Murray has a big lead were the election held today, based on Elway’s Poll alone, but rather that within the context of other published and unpublished polls, there’s no reason to believe that Rossi would win, even by a slight margin, were the election held today.
The main point being, Murray isn’t nearly as vulnerable as the SurveyUSA poll would suggest, and Rossi would be making a mistake merely averaging it in with the others as the crosstabs are bizarrely out of whack both with other polls this cycle, and with historical results.
Who knew John Elway would become a pollster when he retired from the NFL.
Slippery Pete @ 10,
Here is a graph of the polls in the Gregoire–Rossi race from September to Monday, 3 Nov 2008. (Here is a longer series that ends on 1 Nov 2008.)
The final tally was 48.27% Gregoire to 46.35% for Rossi. If we normalize the numbers (i.e. make them add up to 100%) so that they can be compared to the graphs, it is Gregoire 51.0% to Rossi 49.0%.
It looks to me like Elway was overly optimistic for Gregoire in 2008. The last polls by Rasmussen, Strategic Vision, SurveyUSA, and Washington Poll were all right on the money.
@13. Thanks, that’s good info. I’m not sure why the Elway poll was so far away from the others. It looks like Elway and Washington both used the same methodology . . . multi-day survey of registered voters using live interviewers. Not sure what would make Elway’s results so far away from the actual results, the cluster of likely voter polls, and the Washington poll that used similar methodology.
Still, even if we spot Republicans five points to compensate for the “Elway effect,” that’s still 51-39, Murray over Rossi, based on this poll.
“What makes Stu’s poll the “legitimate” one and SurveyUSA’s the outlier? Why couldn’t it be the other way around?”
Well, the numbers indicating remarkably less support in Seattle, and among Democrats, would be a red flag that something was wrong, I would expect.
There’s been a lot of experience with polling over the past 3/4 century or so, and they have usually been able to get pretty accurate results with a relatively small sample. But there’s always the chance that a handful of calls just happen to catch extremists, which skew the results, hence the “outlier” concept.
But more likely, I’m thinking that in urban areas the inability of polsters to call cell phones is putting the entire sample methodology in danger. Younger people are foregoing land lines altogether in lieu of a cell phone. How can you keep having accurate sampling if an entire demographic, one which is growing over time and which is tending to support one party over another, is excluded?
While Rossi might take a tiny bit of solace from the fact that Murray, an incumbent, only has 51% in this poll with six months to go before the general election, he should also note that he would have to get EVERY ONE of the undecided votes, PLUS another 1% + 1 vote, in order to win.
I don’t doubt that he can gain some ground between now and November, but I’d be surprised if he got over 45% of the vote in the election.
@16: I believe the general rule of thumb is that any incumbent over 50% is safe, barring some unforseen occurrence.
When I wrote about “spotting Republicans 5 points” @14, it made me think about the “house effects” Nate Silver has written about over at 538.com. So I figured I’d look up the most recent poll in this imaginary race, adjust the results for each outfit’s respective “house effects” that Nate has calculated, and share the results. I opened up a spreadsheet, typed in the published results from the most recent Rasmussen, R2K, and SUSA polls, and the Elway results, then went to 538.com. Apparently, Nate beat me to the punch and wrote an entire article on the weird polling going on here:
http://www.fivethirtyeight.com.....si-51.html
Not even he can make sense of it. It’s interesting that he said SUSA is a good pollster, which contradicts what some of the commenters have thought. But he did notice a few flaws with this particular SUSA poll.
Even though Nate said it wasn’t worth trying to analyze it further, and he’s infinitely more knowledgeable in this area than me, I continued with my original idea. I took the most recent results from Rasmussen, R2K, and SUSA from here. Then I normalized the results to ignore undecideds. Then I adjusted based on the house effects Nate has calculated this cycle, which is R+5.5 for Rasmussen and D+4.4 for R2K. Neither SUSA nor Elway have house effects calculated by Nate, but we’ll say Elway’s is D+5 based on the rather unscientific method of comparing their final 2008 gubernatorial poll to the actual results. For SUSA, we’ll say their house effect is 0 because (A) Nate says they’re typically quite good and (B) Nate’s most recent calculation of pollster-introduced-error shows them to be the second best outfit of all. Making those adjustments, I came up with the following levels of support for Rossi’s imaginary run against Sen. Murray: 43.5% (Rasmussen), 55% (SUSA), 48.5% (R2K), and 45% (Elway). So actually, Elway’s poll might be pretty close. Both Rasmussen and R2K are within Elway’s MoE of +/-5% (once you account for their respective house effects). This would mean SUSA really is an enormous outlier. In sum, I would say that Murray leads Rossi by about 55-45%, which I think most people here would think is reasonable. I have no idea what makes that SUSA poll such an aberration though, and now I’m really curious to know what went so wrong for them.
[Deleted — Off topic]
[Deleted — Off topic]
Slippery Pete, that’s interesting.
I may be biased, but I don’t think this particular R2K poll was nearly as D-favorable as Nate suggests. Looking at their crosstabs, I noticed that that poll had extraordinarily strange results in the 7th Congressional District.
If you look at the historical numbers since 2004 (shown in the second point of my comment #9 above), you’ll get a good idea of what might be expected here in McDermott’s district — somewhere close to 80% for Patty.
The R2K poll shows Murray-Rossi-Undecided in WA-07 as 63-31-6. That’s 15% low for Patty.
That anomaly, all by itself, should add about 1.8% to the Murray percentage — rather than 1/9 of the state, I called it 12% of the state, since Seattle’s turnout is higher than the overall turnout. Adding 1.8% to Patty also subtracts 1.8% from Dino, so I’d suggest that the WA-07 error places Murray’s R2K margin around 3.5% lower than the poll “should” have produced.
Today’s Rasmussen has it 48-46 for Murray. I suspect that is pretty close. The other candidates are longshots.
Wednesday, May 05, 2010
I suspect Rossi will soon announce.
Remember, Rasmussen had the Mass. Race a deadheat….and Brown beat Coakley by 5.
We need a good battle.
The Key Issues will be:
Deficit Spending
Growing Government
Health Care Plan
Immigration
Murray loses on every one of them.
@22: I think you’re right, but here are a few counterarguments to your point:
(1) It’s possible that Murray’s support really has eroded a little bit in the 7th CD.
(2) The laws of probability suggest that if a candidate’s support is overstated by X% within some subsample of the poll, it is understated by X% in another subsample of the poll. Or put another way, R2K surveyed 600 voters, for an MoE of +/-4%. But their CD07 sample was only 1/9th as large, so the MoE for that subset is about +/-12%. Thus, crosstabs for CD07 are only guaranteed to be within 12 points of their “true result.” So perhaps Murray’s support is overstated in some other CD by the same amount that it is understated in CD07, and the aggregate result of the poll is an accurate reflection of the current state of affairs.
On the other hand, in support of your point, I don’t think Nate was saying this R2K poll in particular inflated the D’s results. If I understand what he means about house effects, then on average, all of R2K’s polls this cycle have overstated D support by 4.4 points. So there is no reason why this particular poll couldn’t understate the D’s support by some amount. Just like he says SurveyUSA has been pretty solid, but their recent Rossi/Murray poll seems to overstate R support by more than a healthy amount.
I don’t care.
My money is still on Dino to win the majority of polls that don’t matter before the poll that counts.
Winning the meaningless polls is Dino’s MO.
[Deleted by Darryl — Off Topic see HA Comment Policy
Dr. Dre will be taking a small vacation from this blog]
25 – Great news..
Patty Murray has not ever been, and is not, vulnerable.
And I still say that the interesting race is whether Dino can win the primary, and I don’t think he thinks he can.
@22
Having seen the other candidates in action, I’d say it’s Murray FTW by at least 10%.
Murray
loseswins on every one of them.Fixed another one for you, KKK.
@27: The humiliation of Rossi losing in the primary would be fatal. His career as a politician with unique ability to snatch defeat from the jaws of victory would be over.
@29
Even if she didn’t you’d be hard pressed to win a state wide race with the 10ther numbskulls the Republicans are fielding.
Patty Murray’s entire political career has been one of confounding her opponents. I think she is considerably more popular in this state than is our governor; I’m not entirely sure why that is, but it might have something to do with the 2004 gubernatorial election spin, by which Rossi captured the media, though not the governorship. Patty Murray’s politics are pretty much this state’s politics–she’s attuned to cross-Pacific commerce, pro-veteran, socially liberal, pragmatic, and a fan of what works. Go for it, Dino.
Yeah.. Funny that the trolls and others keep calling her names and she keeps winning re-election.
This November will be no different as far as I can tell..
Heh. Yeah, funny that.
I’m sure Sen. Murray would be more than happy to add Dino Rossi’s greasy pelt to her collection of Republican scalps.
In 1992 everyone was sure Sen. Brock Adams would crush Murray in the primary in spite of the sexual harassment allegations against him. After he dropped out of the race everyone was sure Rep. Rod Chandler would win against Murray in a walk. In 1998 everyone said 1992 was a fluke and surely Rep. Linda Smith would wipe the floor with a lightweight like Sen. Murray. In 2004 the Democrats were out of power and unpopular, surely with the Osama comments and the vote against the Iraq war this was the year of all years to finally retire Sen. Murray, Rep. George Nethercutt was just the man to do it too.
Actually even though it means Murray won’t be able as much to help Democrats win other Senate and House seats I do hope Rossi jumps into the Senate race. Why you ask? Well given the rest of the GOP field there is the danger that Sen. Murray won’t pay quite enough attention to her own re-election and one of the teabagger midgets like Didier or Benton will threaten to pull off a Brown vs. Coakley. At least with Rossi in the race she’ll take the election seriously and won’t be asleep at the switch. Besides there is nothing the voters hate more than being taken for granted.