SurveyUSA has just released a new poll for the Washington state gubernatorial race. The poll of 634 people shows Governor Christine Gregoire leading Dino Rossi 50% to 46%. The 4% spread is an improvement for Gregoire from the 1% spread in the early April SurveyUSA poll and the 1% spread in a late March Rassmussen poll.
Clearly, the current SurveyUSA poll results are within the 4% margin of error for the poll. What this really means is that the probability of Gregoire beating Rossi is less than 95%, based on what we can tell from a relatively small sample. But we can do better than to just dismiss the poll as “a statistical tie.” Instead, we can estimate the probability that Gregoire should win, given the information available from a sample of this size. This can easily be calculated mathematically. But it is more intuitive to simulate elections and simply count the number of Gregoire wins and Rossi wins.
I simulated a million gubernatorial elections of 634 voters each, where each person had a 50% chance of voting for Gregoire, a 46% chance of voting for Rossi and a 4% chance of voting for neither.
The results give Gregoire 859,651 wins and Rossi 132,432 wins. This suggests that Gregoire has something like an 86.7% probability of beating Rossi if the election were held right now, and Rossi has a 13.3% chance of beating Gregoire.
Here is a plot showing the distribution of votes in the million elections:
The area to the right of the red line are wins for Gregoire and those to the left are wins for Rossi.
Things may change between now and November, but if I had to bet the (server) farm on a candidate right now, my money would be on Gregoire.
Darryl. You have a gigantic brain. I love it when you break it down like this, I bet we hear “statistical dead heat a thousand times…until the Elway poll comes out later this month.
PS – this poll was definitely taken before Rossi’s laughable transportation proposal. He’s probably in the low 30’s now.
I wonder if such polls can reflect voter turnout. Obama is going to be on the ballot, and a lot of people are going to vote this time around that didn’t the last gubernatorial election, or any election. The majority of those Obama voters will pull the lever for Christine, I suspect.
KING5 + Robert Mak + Robo-polling which only calls old people with land lines = PURE IDIOCY
What happens to the model if you you make 4 groups instead of three.
Core votes that will never change …. for each party.
Total choosing party-core vote
This model assumes that undecided will break as the non core.
And Gregoire has just started her campaign, and as governor she can be in the news any times she chooses. Rossi had to sweat blood to get his laughable and widely criticized (as “loony”) transportation non-plan into the news. Now that he’s fallen on his face will they be so quick to publicize his next inanity? I think not. Lou Guzzo has a better chance of winning.
BTW, I am sure she stiffed the Sonics because it would play in E. WA.
SeattleJew,
“What happens to the model if you you make 4 groups instead of three. Core votes that will never change …. for each party. Total choosing party-core vote”
SurveyUSA didn’t identify these numbers except that they identified 2% who would vote for someone else and 2% who were undecided.
“This model assumes that undecided will break as the non core.”
It does. If the election were held now, and that 4% (other plus undecided) break completely for Rossi, the race would end in a tie (i.e. 50%-50%). If the 4% break completely for Gregoire, she would win with a 97.7% probability.
If only the 2% undecided break completely for Rossi, Gregoire has a 72% chance of winning. And if the 2% undecided break completely for Gregoire, she has a 94% chance of winning.
MakeItSo @ 4 has a good point. Polling using the telephone is becoming less and less accurate, due to the number of people who are ditching their land lines entirely and using their cell phones instead.
Fortunately for the Democrats, it appears that the young people tend to be the ones most likely to not have a land line at all. They tend to be overwhelmingly Obama supporters, who will vote for Gregoire once they learn a little bit about who Rossi and the Washington State Republicans really are.
Reminds me of the 1932 election, when the polls showed Hoover winning overwhelmingly right up to the day of the election. The problem was, the only people who could afford phones at the time were pretty rich, and weren’t badly impacted by the Depression. Everyone else had lost their phones along with their cars and their jobs, and they voted for F.D.R.
Did you run a Monte Carlo simulation on whether Hillary or Obama wins the nomination for the Demcorats?
rhp6033,
Wasn’t that the 1948 election, with Harry Truman holding up the headline showing Dewey defeating him in ’48?
I am glad to see that this poll favors CG but I never count on these polls to indicate a winner – the only poll that counts is in Nov.
Survey USA routinely uses automated phone-calling and random digit dialing, asking people to “push 1 for Yes, push 2 for No” etc. I received one of these calls once and just hung up.
Polling on the cheap. It can’t be reliable without a database of actual registered voters.
Obama at the top of the ticket should help. Same poll shows him leading McCain by 13% (Clinton only leads by 3%). Of course, if Clinton steals the nomination and overturns the will of the people, many would be Gregoire supporters may not even bother to vote out of disgust.
@4,12
Yup!
@7 Darryl
Hmmm, still, couldn’t you use a fudge factor for the minimal estimate pf those absolu7tely committeed? Perhaps some fraction of those who calim party affiliation?
I think this is important because the model you use assumes an equal probaability of all demo and all rep votes being labile. I doubt that is true but really have no idea what then the factor would be for each party.
What happens to the model if you just assume that 30% of the lectorate is D and 30% is R?
Obviously there would be no effect in the trend, but your estimate of certainty would, I think, be much lower and therefore more realistic.
One reaso0n I think this is a critical issue jow is that I think we are going to se a decrease in firm R this year.
Darryl–
WTF?!
How do you find the time to spend on meaningless minuscia like this??
Did you do the same thing with the Polls in 2004?
How accurate were they?
My take is few who voted for Rossi last time will vote for Gregoire this time.
Many who voted for Gregoire will vote for Rossi this time.
Gregoire MUST attract new, uneducated voters.
The most reliable poll of recent elections has been Rassmussen.
This just out today:
Friday, April 18, 2008
The Rasmussen Reports daily Presidential Tracking Poll for Friday shows John McCain leading Barack Obama, 48% to 42%. The presumptive Republican nominee also leads Hillary Clinton 50% to 41%.
Let’s see how Rasmussen views the Governor’s Race.
@13 Upton
Over a year ago I outlined the strategy ofr BHO. He has followed it closely so maybe I am correct?
I think his next target is the religious right. There is, I believe, a lot more lability in this group of Rs then most folks understand.
If I am right, BHO will:
a. choose a running mate who is acceptable to the believers. Thgis rules out Hillary but maybe not Pellosi.
b. Give a speech , perhaps THE nomonation speech, on the relationship of religion to the state. He will shock some libs by committing to funding FBOs and proposing ways religious private schools can work with the public schools.
He will also rail against religious suppression in Muslim and communist countries. BUT he may also havea muslim speak at the convention … a “save my faith” talk?
c. He will make a showy visit, /c ministers, to Black and White chiurches across the country. NOT including the radical religious.
d. he will identify the protestant work ethhic, Jewish commitment to education, Catholic mcommitment to Mary’s comapassion etc (/o openly stereotyping anyone) to national service.
e. He will rpos0pse a MOrmon like period of national servece for all Americans,
This also released today by Rasmussen:
On Friday, the Rasmussen Reports Balance of Power Calculator shows Democrats leading in states with 190 Electoral College Votes while Republicans are favored in states with 189 Votes. State with 121 Votes are “leaners,” and states with 38 Votes are Toss-Ups.
Next Rasmussen on Washington Governor’s Race comes out around 5/1–
Here is the 4/2 Rasmussen:
Wednesday, April 02, 2008
The latest Rasmussen Reports telephone survey of Washington voters found a virtual tie in the race for the state’s next governor. Democratic Incumbent Christine Gregoire holds a statistically insignificant lead over Republican Dino Rossi, 47% to 46%.
Darryl–
If you are going to be a “poll-jockey”..or “pollock” or whatever, why not use a proven reliable poll?
The Rasmussen had Cantwell up by only 4 points over McGavick about this time in their campaign.
Polling at this time is pretty useless, unless you’re a campaign manager.
It’s closer than a race pitting a seasoned, successful professional against a cartoon politician ought to be. This state has too many stupid voters. Can we box them up and ship them to one of the atolls used by the U.S. government for atomic testing? They can safely glow in the dark there.
@10 No — rhp has her history right.
@11 You’re absolutely right and I couldn’t agree more. Overconfidence based on over-reliance on polls nearly cost Gregoire and the leaders of her campaign team the ’04 election. I hope they won’t repeat that mistake.
@16 “My take is few who voted for Rossi last time will vote for Gregoire this time.”
And my take is none who voted for Gregoire last time will vote for Gregoire this time, and a goodly number who stayed home or voted for Bennett will also be in her camp in ’08. Your pooch loses, Cynical. But thanks for playing.
@18 “BUT he may also havea muslim speak at the convention … a “save my faith” talk?”
WTF? Are you buying into the rightwing Obama-is-a-Muslim crap, SJ? If so, why?
Cynical: You are dreaming if you think those who voted for Gregoire in 2004 are going to flock to Rossi in 2008. I’m friends with lots of conservatives, and they all seem to hold the same opinion you do, all agreeing with one another about what a “lousy governor” Gregoire has been, and how Rossi will win it convincingly this time around.
The problem is they are a fairly insular group, who tend to talk mostly with those who already agree with them. And their numbers are becoming increasingly limited. Everyone else I meet, from a wide variety of backgrounds and occupations, won’t vote for a Republican this year under any circumstances. There is a distinct anger abiding in the land, and it is being directed at the Republicans.
The fact that the last election was so close will tend to bring a lot more voters out of the woodwork this time. I think you will see record turnout in 2008, and Democrats will be the beneficiaries.
So 2004 is no basis for making judgements upon the 2008 election. Looking at what happened in 2006 is likely to be more accurate. The only question I still have is whether the results will be similar to those in 2006, or if Democratic gains will be significantly bigger in 2008.
RR @ 23 says: “RHP has her history right”.
Thanks, but I’m not a “her”.
Are you looking at a gravitar of a woman by my screen name? If so, that isn’t a picture of me. The picture is of a woman closely related to me, but that’s as far as I will go in describing her on this board. I was using that picture as a sample while I tested the system.
My real gravitar is of a guy with a beard (me, in a photo taken some 15 years ago, before my hair & beard started turning grey). If you are still seeing the picture of the woman, clear your cache on your computer, and the new gravitar should show up.
“There is a distinct anger abiding in the land, and it is being directed at the Republicans.”
Fuckin’ A.
Politically Incorrect @ 9
“Did you run a Monte Carlo simulation on whether Hillary or Obama wins the nomination for the Demcorats?”
No.
I do almost daily analyses of current general election match-ups, however.
SeattleJew @ 15,
“Hmmm, still, couldn’t you use a fudge factor for the minimal estimate pf those absolu7tely committeed?”
I could, but I won’t. I do these analyses with an absolute minimum number of assumptions.
“Perhaps some fraction of those who calim party affiliation?”
The poll reported that 91% of Republicans went for Rossi and 7% for Gregoire, and 81% of Democrats went for Gregoire and 16% went for Rossi. So those were the numbers I (inherently) used.
“I think this is important because the model you use assumes an equal probaability of all demo and all rep votes being labile. I doubt that is true but really have no idea what then the factor would be for each party.”
I don’t have any idea, either. Hence, I don’t impose numbers I might pull out of my ass on the analysis.
“What happens to the model if you just assume that 30% of the lectorate is D and 30% is R?”
Nothing, since I made no use of party identification. BTW, the actual party identification (that was inherently used) is: 28% self-identified as a Republican, 41% self-identified as a Democrat, and 24% self-identified as an independent.
“Obviously there would be no effect in the trend, but your estimate of certainty would, I think, be much lower and therefore more realistic.”
Nope…my estimate of certainty would be biased by the new (and quite possibly, unjustifiable) assumptions.
“One reaso0n I think this is a critical issue jow is that I think we are going to se a decrease in firm R this year.”
Then that will be reflected in polls later this year. The analysis I did tells us what to expect if the election were held now, not sometime down the road when things change.
Mr. Cynical,
“WTF?! How do you find the time to spend on meaningless minuscia like this??”
I guess…the same way you find time to respond with meaningless “minuscia.”
“Did you do the same thing with the Polls in 2004?”
Not for the Governor’s race, but I did some similar stuff for the presidential election. And I did some for races in 2006 and 2007. Why do you ask?
“How accurate were they?”
How do you define accurate? As I stated, the analysis gives a probabilistic estimate of the outcome of an election held NOW. Maybe you can talk to your BIAW buds and get them to fund a general election early next week so we can test it?
“My take is few who voted for Rossi last time will vote for Gregoire this time.”
So you’ve said. And your opinion will make a mighty fine addition to the shitpile of opinions.
“Many who voted for Gregoire will vote for Rossi this time.”
Your opinion does not currently seem to be supported by the polls.
“Gregoire MUST attract new, uneducated voters.”
It looks like Gregoire and Rossi nearly split the vote (48% to 49%) amongst those without a college degree. Furthermore, Gregoire leads Rossi, 50% to 46% among the youngest voters (those from 18 to 34 years). In fact, the only age group that prefers Rossi over Gregoire (50% to 47%) is those from 36 to 49 years.
Mr. Cynical,
“Here is the 4/2 Rasmussen:
Wednesday, April 02, 2008
The latest Rasmussen Reports telephone survey of Washington voters found a virtual tie in the race for the state’s next governor. Democratic Incumbent Christine Gregoire holds a statistically insignificant lead over Republican Dino Rossi, 47% to 46%.”
And…my analysis of the Rasmussen poll was posted here.
Several days after the Rasmussen poll, a SurveyUSA poll came out with almost identical results. I did the same analysis here. The two polls were taken so close together, that I even did a combined analysis.
Are you happy now?
2cents,
“Polling at this time is pretty useless, unless you’re a campaign manager.”
If your goal is to predict the outcome of an election, then the value of a poll decreases with increasing duration to the election.
On the other hand, so is the score in a basketball game. Fans wouldn’t accept having the score hidden during the game simply because an ongoing score does not perfectly predict the winner.
Likewise, election polls are how us political junkies “keep score” during the game.
16 Cyn
Mr. Cynical spews:
My take is few who voted for Rossi last time will vote for Gregoire this time.
Many who voted for Gregoire will vote for Rossi this time.
Of course it is. Your “take” would never consider that some people might actually be impressed with Gregoire’s record as governor. Your “take” would never consider that some peoiple weren’t completely happy with Dino Rossi’s high-handed attempted coup (shades of 2000!) in calling himself Governor-elect before the votes were certified. Your “take” would never consider that some people might look at Dino Rossi’s four years of wingnut welfare and decide they’d rather vote for someone who is employed. Your “take” would never consider that anyone who would hire Lou Guzzo might be ridiculed for same.
Daddy Love @ 35,
Not to mention, his incredibly ungracious and un-statesmanlike slam on the Supreme Court when he begrudgingly ended the election contest.
SJ
d. he will identify the protestant work ethhic, Jewish commitment to education, Catholic mcommitment to Mary’s comapassion etc (/o openly stereotyping anyone) to national service.
No, that was Mitt Romney ridiculousness.
What makes me laugh is the “pro business” conservative crowd HATES Govenor Gregiore, and yet refuse to believe that Forbes magazine rates the state of Washington as one of the BEST places to do business.
Yeah, you know that magazine run by the LIBERAL Steve Forbes.
Sheeeeez, voting Republican is voting against your own best interest.
The other thing that makes me laugh my ASS off, is the Republicans who seem to be pleased that McCain is close to either of the Democratic challengers.
Holy shit, McCain ought to be SKY HIGH in the polls against the Democrats right now. As soon as we pick Obama to be our nominee, he’s going to get a major bounce in the polls.
Then, just as the economy keeps sucking for Middle Class Americans, and Paris Hilton keeps getting richer, there will be another major lift in the polls and DONATIONS to Obama.
Say it with me: “President OH-BAH-MA”
Good.
Sorry, I meant to comment that polling is useless in determining the election at this point.
If this were a basketball game I don’t even think we’re to the 1st quarter. It doesn’t appear the Sonics mess hurt Gov. Gregoire at all.
The backside of the polling is interesting. I’m surprised how many Democrats, pro choice, and liberals poll for Rossi. I’m curious how his numbers go now that his transportation plan is out.
For the unaware – Rasmussen typically favors republicans.
@28 This doesn’t change the fact that God is a “She.”
Funny as hell, who did they poll? Martin
Luther King County Election employees?