MI-2012 President: 45% Obama (D), 44% Romney (R) (FMWB 9/12)

Foster McCollum White Baydoun (D)/Fox 2 News Detroit
9/12/12; 1156 likely voters; 2.88% margin of error
Mode: Automated phone
FMWB Release

Michigan

2012 President
45% Obama (D), 44% Romney (R) (chart)

2012 Senate
47% Stabenow (D), 42% Hoekstra (R) (chart)

Comments

[rafaelkafka]

04:24 PM on 09/16/2012

A lib pollster giving O. just 1+ in MIchigan? You are toasted, O.!

[axt113]

10:00 PM on 09/16/2012

Its an outlier, clearly the problem arose from the massive weighting they had to do for the age demographics, the other polls show a larger Obama lead

[TJampel]

11:21 PM on 09/16/2012

Actually, YOU gave O 1+ in MI; the pollster gave O 1.84+ in MI.

(President Barack Obama): 45.49%
(Governor Mitt Romney): 43.65%

Naturally, you'd need to read the report. I also wouldn't expect you to have any opinion regarding the fact that 82.68% of the respondents to this poll were over the age of 50. Maybe you think that that reflects our electorate.

Let's see... this same pollster gave Romney a 14 point lead in FL on 8/17. Check the FL chart to see it.
So you can now tell everyone that a lib pollster has Romney up 14 in FL in their most recent poll so Romney must be up even more.

[Rightbrainedleftwinged]

02:44 PM on 09/16/2012

Again, I say, that this polling firm is probably hoping that there will be small turnout in 2012 in Michigan. Not even a chance.

Stabenow, is probably up by 10. Not surprised she'd outperform Obama, since Hoekstra is unknown and very weak as a candidate.

[dpearl]

03:39 PM on 09/16/2012

Rightbrainedleftwinged,

Actually, they make their turnout prediction for Michigan explicit in their release (scroll down to page 30 to see that). They are predicting a 70.3% turnout which is a little higher than 2008. They note that there has been an historical trend of increasing turnout per presidential election cycle going back to 1960.

[Rightbrainedleftwinged]

05:14 PM on 09/16/2012

6.5 percent African American? Mich has about 17 percent, and at least 12 percent will probably vote in the election. If it were Iowa or WI I'd be nervous, but not so much in MI.

[LtCmdrWalrus]

12:24 PM on 09/16/2012

Quite a contrast to the earlier poll that had Obama up by 10. It's also very odd that Stabenow is outperforming him. Still, the trend line on the chart is pretty clear-MI isn't going red anytime soon.

[TJampel]

10:57 AM on 09/16/2012

Wow....this firm really went "there"! For all of you who keep saying "push poll" all the time when you talk about a firm you don't like....here's a classic example of a Push Poll question right outta Politics 101. Wish I was teaching it (not even my field) so I could have students ready this:

Question #4 says it all:

"Question #4:
Since the Republican National Convention, do you believe that you have a clear
understanding of Mitt Romney and Paul Ryan plans to lead America and solve the
problems of our country (taxes, the economy, national security and political gridlock)?"

No similar Q re Obama/Biden---but of course. Push polls are done to change opinion, and now it makes sense why they're polling old folks.

I'm. just. speechless ----that Pollster didn't notice this "push poll" tactic and axe the poll.

[dpearl]

11:49 AM on 09/16/2012

I agree that there seems to be an imbalance there (and have e-mailed Eric Foster for comment). But your theory that Fox News, who sponsored this poll, specifically asked that question #4 be included to make the Republican ticket look bad doesn't seem plausible to me at all.

[StatsNo1fan]

12:13 PM on 09/16/2012

Just to be clear - Fox 2 News Detroit probably is not an affiliate of the Fox News Channel, but of Fox's broadcast network, which, considering the kind of shows it airs, certainly does not have a conservative ideological lean.

Further, the question about "clear understanding of plans to lead America" can as easily, if not easier, be construed as a negative one, since most observers complained that neither convention offered clear or specific plans. So omitting a similar question about the Dem candidates can be viewed as favoring them, since the answer to the question about either is likely to be a "no".

Further still, this is a Dem polling firm, as noted on this page, on RCP, and as described by Silver.

[TJampel]

01:40 PM on 09/16/2012

That's not my theory. However, what explanation do you have for this question being used? It's a question no polling firm would ever inject on their own, yet, there it is. And what's your take on this type of question in terms of whether it constitutes a "Push Poll" type of question designed to move the respondents to a different position?

[dpearl]

01:05 PM on 09/16/2012

Eric Foster just got back to me about Question #4. He said that their team did debate that question before they went into the field. They decided to put it as question #4 so as not to bias questions 1-3. The remaining questions in the survey were about state propositions and they did not think that #4 would bias those.

That is a reasonable point to make here - but that still leaves a problem with the interpretation of question #4 itself. It's sort of like having an experiment with a treatment group but no control group. When 51% say that the Republican convention gave them a "clear understanding" of Republican plans - is that good or bad? Without a comparison with how people would respond to the same question about the Democrat's convention you can't tell.

[TJampel]

01:43 PM on 09/16/2012

Sorry....didn't see this response; you answered my question quite well. And I agree with your conclusion. I think Eric Foster's a jerk, and I think that there's no way to read this question except as one which promotes a particular agenda, due to its wording. It's not asking "what did you think of how the Rpeublicans ran their convention" or....do you think Republicans or Dems are better qualified to lead America...etc..

[dopper0189]

07:04 AM on 09/16/2012

This is the election cycle where the media is finally going to have to confront the really bad poll syndrome. You have a number of polling firms that will now just publish polls in order to try and set narratives. National media and polling aggregators may have to be more discerning in the future about what firms they site. Polling firms methodologies, techniques, and track records are going to have to be taken into account. This cycle most of the "crazy" polls have been on the right. But if these techniques aren't called out, I could see a future world where super PACs on both sides sponsor garbage polls every week and flood out legitimate polls in order to try and get momentum for their candidates and to set the narrative.

It's something to think about after the election. What are the minimum qualifications to be considered a legitimate poll?

[ashenthorn]

07:53 AM on 09/16/2012

I get your point about the narrative-setters (an example of this was Scotty Ras spotting Obama a 3 point lead on the eve of the Republican convention in order to produce a fake "bounce" for Romney). I'm not sure these guys are part of that set. These guys, in their notes, simply throw up their hands and admit that they don't (or can't) get in touch with people who tend to be Obama voters. They explain that they only call landlines because calling cell phones is too "difficult" for them. And 83 percent of their respondents are over 50, and mostly all white. They give a higher weight to very few younger and minority responses they do get to try to reflect some semblance of electoral reality, but that they don't even actually try to include a representative number of these voters in their sample makes their conclusions about how they might vote extremely suspect.

[dpearl]

12:13 PM on 09/16/2012

One "qualification" I care about is transparency and this pollster is exemplary in that regard. All pollsters are getting a terrible response rate these days and this is the only pollster that regularly owns up to the exact response rate they are getting (here a bit less than 3%). Similarly, other pollsters don't tell you anything about the relative response rates in different demographic groups (resulting in high weights and hence more volatility for those groups) - but this one does.

As I have posted before, there are some overall structural problems with automated phone surveys and these organizations will have to address them moving forward. However, accusing them of being purposefully deceptive is completely unfair. Further, accusing the most transparent companies of being the most deceptive is completely illogical.

[Rightbrainedleftwinged]

11:17 PM on 09/15/2012

Gimme A F'in break.

[gbshabo]

12:19 AM on 09/16/2012

Their last poll had Romney up 3 if its any comfort.

[Williame123]

12:54 AM on 09/16/2012

They had him up by 4 in MI and up by 15 in FL. Apparently they believe only white men over age 50 will vote.