Quick Links:

MPR Poll: bad methodology, slim chances for Emmer

by: Joe Bodell

Tue Aug 31, 2010 at 07:55:10 AM CDT


MPR is out with a gubernatorial poll with a rather questionable methodology, showing a dead heat between DFLer Mark Dayton and Republican Tom Emmer at 34% apiece.

If it takes an 8-point oversample in Tom Emmer's favor to get him up to a tie, I feel pretty great about Mark Dayton's chances in a real electorate in which younger, cell-phone-only voters show up.

But aside from the weird methodology, check out the published crosstabs:

1. Independent voters:
Undecided: 38%
Horner: 26%
Dayton: 23%
Emmer: 13%

There's a lot of room for movement there, but there is virtually no way Emmer picks up significant enough ground among independent voters to make a dent in the overall results. Keep in mind that this is a mid-term election, and the non-partisan vote is generally going to be a lot lower than it is in presidential years, so given a normal partisan breakdown, or even a slightly GOP-leaning one, Emmer has a LOT of ground to make up.

2. The gender gap: MPR's writeup indicates that there's no significant gender gap -- that women are currently favoring Mark Dayton by a similar margin to men favoring Tom Emmer. However, what they fail to mention directly is that the sample includes 52% women (about normal for Minnesota) which is yet another built-in advantage for Dayton. Again, given a more reasonable partisan sample, this will go straight through to the final results of this election.

3. Age gap? MPR doesn't appear to have published the support breakdowns by age, only the sample sizes -- which look weird in and of themselves, since it's a decent bet the senior vote will be bigger than this poll indicates. If it is, it's another good bet that those voters will go with Dayton in big numbers, especially outside the city -- as we found in the DFL primary, these voters are more likely than not to go with the name they know and trust, and that is Mark Dayton.

Again, if it takes a huge GOP over-sample to get Tom Emmer up to a bare tie, I think Mark Dayton is in pretty darned good shape right now.

Joe Bodell :: MPR Poll: bad methodology, slim chances for Emmer
Tags: , , , , , , , (All Tags)
Print Friendly View Send As Email
Tweet This!
Related Posts Plugin for WordPress, Blogger...
8 point oversample (0.00 / 0)

Help me out Joe--where is that?  At Math not so good am I.    

Normal distribution in Minnesota (0.00 / 0)
is about 38% DFL, 35% GOP, with the rest identifying as independent or unaffiliated. So for this poll to indicate an electorate that is 46% GOP is somewhere around 8 points (actually closer to 11, but that's what margins of error are for) too heavy in the GOP's direction.

[ Parent ]
The way the demographics broke out (0.00 / 0)
their sample consisted of 46% republicans, 41% democrats and 13% independent. This is very different from the sample compositions used by other pollsters in this race so far, and a very different composition of voters than usually show up at the polls in Minnesota.

SurveyUSA, for example, in their last poll of the state used 32% GOP, 36% Dem and 28% indy.

This sample composition issue seems to be a running theme in Humphrey institute polls and is perhaps a result of their methodology, which while being random, does not appear to correct for oversampling.


[ Parent ]
GOP % too high (0.00 / 0)
The only way to make this race close is to skew the sample pull to make it more favorable for the Republicans.  Self-identifying Republicans are more likely around 25-30% with Democrats in the 37-43% range.  

At least this was the case in the '08 elections and I don't think things have changed much since then.


[ Parent ]
I won't use (0.00 / 0)

'08 as a plum line; that was a very different year in terms of party ID and turnout.  Anybody know what the turnout in terms of party ID tends to be for MN gubernatorial races?

[ Parent ]
Most pollsters (0.00 / 0)
seem to be using a tight/very tight likely voter screen on self-identified Dems, from what I've seen.

The 'dispirited Dems' narrative is certainly suffocatingly pervasive in corporate media.


Why don't the press "oversample"" (0.00 / 0)
to make Tarryl Clark seem more competitive with Michele Bachmann in the Sixth?

When they release something like this--whether it's true or not, whether the supporting info nonsense or not--it helps the candidate who's behind.

And they know that. So if they report this: why don't they explain why the poll results make sense in light of what is already known about the political biases of those surveyed?

They don't do that? They just print the "result" whether it makes sense or not? You'd almost think there was a conservative bia in the media, or something.


To be fair, (0.00 / 0)
I don't think MPR/Humphrey oversampled on purpose, it's more likely an artifact of their methodology, although like I said before, this is hardly the first Humphrey poll with a weird partisan composition.

I agree that it would be nice for the media to also report what the partisan breakdown of the poll was, I've heard MPR plug this poll twice today and they talked about many aspects of it but not the partisan ID breakdown.


[ Parent ]
Horserace haiku (0.00 / 0)
Sun rises at dawn!
Who will listen breathlessly
to the obvious?

As Iris [DeMent] says, 'Let the mystery be...'

Yeah, I hate horserace journalism (4.00 / 1)
too. I've always thought there was way too much horserace stuff, as a substitute for substantive journalism. I've always thought that political journalists should focus on the investigation of the factual record, on the policy debate, on the performance of the politicians on behalf of the voter--all that substantive stuff, instead the "who's ahead, who's behind" horserace stuff.

But it's September in an election year, and I want to stay on top of this horserace stuff as we get closer to November...

Because if the guys on the other side win this particular horserace--the only thing we'll be able to do on substantive stuff after the election is talk about that substantive stuff.

Just talk, nothing else. No action on the substantive stuff, no policymaking on the substantive stuff. We'll be limited to "just talk," if they win back control.


[ Parent ]
Shiny objects (0.00 / 0)
Producers wonder
what can we put between ads?
Free circus would do.

As Iris [DeMent] says, 'Let the mystery be...'

[ Parent ]
Gatorade (0.00 / 0)
The headlines, anchor desk banter and bullet points prompted by any new poll distort reality and in terms of voter perception favor the candidate who had seemed to be trailing.   In this instance, Emmer is being served Gatorade by MPR, the Strib, KSTP and the others.



About That Oversampling (0.00 / 0)
"If it takes an 8-point oversample in Tom Emmer's favor to get him up to a tie, I feel pretty great about Mark Dayton's chances in a real electorate in which younger, cell-phone-only voters show up."

While that's statistically significant in terms of registration, it isn't unreasonable in terms of turnout models.

Anyone that thinks Dayton is attracting large numbers of young voters needs a serious shot of reality.

Finally, I wouldn't get feeling too confident yet. First, there's an ABM backlash happening. Second, ABM has hit Emmer with everything including the kitchen sink, some IEDs and RPGs. Despite all that, Dayton is tied.


Hi Gary (0.00 / 0)
Dayton doesn't have to "attract" large numbers of young voters -- their tendencies are their tendencies, and if one leaves them out of a poll like this (which one does when excluding cell-phone-only households), one builds in a bias in Emmer's favor.

As for the alleged "ABM backlash" I'd challenge you to prove it. I know you're not a fan of ABM as a whole, and you have your strategic reasons, but other than the feigned outrage over ABM talking about Tom Emmer's DUI history in relation to his previous efforts to soften DUI penalties on behalf of a couple of DUI attorneys, where's the beef? It's a pretty tenuous charge at best, and simply isn't borne out in the numbers in any substantive way.


[ Parent ]
Oh, and... (0.00 / 0)
...I feel pretty confident that the folks at ABM won't appreciate being compared to the terrorists who actually do use IEDs and RPGs. Just a thought, Gary. They're your neighbors and care just as much as you do about Minnesota's future. One might try to leave out the over-the-top rhetoric and comparisons in the future.

[ Parent ]
Still, This Is A Wake-Up Call... (0.00 / 0)
Greetings from Brookings, SD on the Buffalo Ridge. Even when you adjust for Repug oversampling, the race is in single digits- that's too close for comfort.

[ Parent ]
No, Gary and his ilk don't agree... (0.00 / 0)
LFRGary and his ilk constantly use inflammatory and violent rhetoric to describe us.  They are serious, just as Tom Emmer was when he said:

"I don't think you can call yourself a freedom-loving American and be a Democrat."

There is no room for discussion.  You either agree with them or they label you as a terrorist sympathizer, un-American or some other such bull**** -- it's easier than thinking.  The reason they simply CANNOT defend what they believe using logical argumentation is that it's often indefensible (climate change denial or torture) or they refuse to face the failure of their ideology (Bush's Depression/Recession).  

In summary, when their arguments fail, they call us names.  Since their arguments always fail, all they can do is call us names.

LFRGary or one of his friends will most likely respond by calling me names.

Watch...


[ Parent ]
Sutton would agree, this poll is bunk (0.00 / 0)
For comparison, let's look at the late July Minnesota poll, which was based on a sample of 32 percent Democrats, 27 percent Republicans and 30 percent independents.

According to the accompanying Star Tribune article, both GOP chair Tony Sutton and DFL chair Brian Melendez said the party breakdown from the poll sounded "about right."

I take that to mean that even Tony Sutton would disagree with the party sample in this Humphrey Poll.  


the best response (0.00 / 0)
We can go over methodology all we want, but no one should be shocked if the race is close. We know Republicans are enthused about voting this year, and we know that despite the national trend of depressed Democratic turnout, DFLers are pretty fired up about the governors race. I've sensed a let down among those who supported Margaret, but that's no surprise since she relied on a ground game, and that meant her supporters put in a bunch of time and got emotionally invested in her campaign. It's time to get past that, but I don't think they've all done so yet. That's just my sense, not hard data.

The best response is what the DFL is already doing. The phone banks and doorknocks that nearly overcame that big poling gap before the primary is going again. It's time to find out where Dayton's supporters are and get them out to vote. It's time to one-on-one ask undecideds to vote for our candidates. The DFL could use more volunteers for the booth at the state fair too. Some conservatives come in for I'm not sure what reason, but there are large numbers of people who are  persuadable.


Senate Republicans decide to make contraception an issue (ericf)
Something to ease the pain (dan.burns)
Bad scene in Greece (dan.burns)
Sorry, but it's true (dan.burns)
The cancer in OWS (dan.burns)
Economic suffering as moral purging (dan.burns)
Dayton vetoes tort reform (ericf)
Must-read on why the 99% is about fraud, not envy (ericf)
Is big-ticket demand bound to improve? (dan.burns)
Conservative evangelical sects still losing young members (dan.burns)
Competent governance makes a difference (dan.burns)
Wall Street whiners say financial reform is working (ericf)
(view all recent)



Liberal Blog Network
Agonist
All Spin Zone
AlterNet
AMERICAblog
American Street
ArchPundit
BAGNewsnotes
BartCop
Blogging of the Pres
BlogACTIVE
Bluegrass Report
Bluegrass Roots
Blue Indiana
BlueJersey
Blue Mass. Group
BlueOregon
BlueNC
Bob Geiger
Booman
Brendan Calling
BRAD Blog
Buckeye State Blog
Burnt Orange Report
Capitol Annex
Carpetbagger Report
Chris Floyd
Clay Cane
Calitics
Cliff Schecter
Confined Space
Corrente
Crooks and Liars
culture kitchen
Cursor
Daily Kos
David Corn
Dem Bloggers
Democrats.com
Deride and Conquer
Democratic Underground
Digby
DovBear
Drudge Retort
Ed Cone
ePluribus Media
Eschaton
Ezra Klein
Feministe
Feministing
Firedoglake
Fired Up
First Draft
Frameshop
Green Mountain Daily
Greg Palast
Hoffmania
Horse's Ass
Hughes for America
In Search of Utopia
Is That Legal?
Jesus' General
Jon Swift
Juan Cole
Keystone Politics
Kick! Making Politics Fun
KnoxViews
Lawyers, Guns & Money
Left Coaster
Left in the West
Liberal Avenger
Liberal Oasis
Loaded Orygun
Mahablog
Majikthise
Make Them Accountable
Matthew Yglesias
MaxSpeak
Media Girl
Michigan Liberal
MN Campaign Report
Minnesota Monitor
MyDD
My Left Nutmeg
My Left Wing
My Two Sense
Nathan Newman
Needlenose
Nevada Today
News Dissector
Newshoggers
News Hounds
Nitpicker
Oliver Willis
onegoodmove
OpenLeft
PageOneQ
Pam's House Blend
Pandagon
People's Rep. of Seabrook
PinkDome
Politics1
Political Animal
Political Wire
Poor Man Institute
Prairie State Blue
Progressive Historians
Raw Story
Reno Discontent
Republic of T
Rhode Island's Future
Rochester Turning
Rocky Mountain Report
Rod 2.0
Rude Pundit
Sadly, No!
Saterical Political Report
Seeing The Forest
Shakesville
SirotaBlog
SistersTalk
Skippy
Slacktivist
Smirking Chimp
SquareState
Suburban Guerrilla
Swing State Project
Talking Points Memo
Talk Left
Tapped
Taylor Marsh
Tattered Coat
Texas Kaos
The Albany Project
The Blue State
The Democratic Daily
The Hollywood Liberal
The Reaction
The Talent Show
This Modern World
Town Called Dobson
Wampum
War and Piece
WashBlog
Watching the Watchers
West Virginia Blue
Young Philly Politics
Young Turks

 Joe Bodell




Hate ads? Make them go away -- Subscribe to MPP!

Change.org|Start Petition





MN-PLAN Ads

Blog Ads




Privacy Policy
Powered by: SoapBlox