Mitchell Research/Rosetta Stone Michigan 2012 GOP Primary Poll
- Mitt Romney 36% {32%} [25%] (31%)
- Rick Santorum 33% {30%} [34%] (15%)
- Ron Paul 12% {7%} [11%] (15%)
- Newt Gingrich 9% {9%} [5%] (16%)
- Undecided 11% {22%} [25%]
Survey of 430 likely Michigan Republican primary voters was conducted February 23, 2012. The margin of error is +/- 4.7 percentage points. Party ID: 77% Republican; 13% Independent; 7% Democrat. Results from the poll conducted February 20, 2012 are in curly brackets. Results from the poll conducted February 14, 2012 are in square brackets. Results from the poll conducted February 2, 2012 are in parentheses.
–Data compilation and analysis courtesy of The Argo Journal
February 24th, 2012 at 12:06 am
Santorum leads 36-3 among Dems. Really?
Operation chaotic might flip this to Santorum. Will Santorum remind people he won because Dems supported him? He always points out Romney only won NH due to Indys (which wasn’t true).
February 24th, 2012 at 12:06 am
Same poll among registered Republicans in Michigan:
Romney 42
Santorum 33
Paul 8
Gingrich 8
Santorum is only still close to Romney because Democrats in Michigan are going for Santorum 36% versus Romney 3%
February 24th, 2012 at 12:07 am
In afterwords, this poll was done after the debates….so it may reflect the impact.
February 24th, 2012 at 12:08 am
7% of likely voters on Tuesday are going to be Democrat.
Romney needs to blast that out to make sure he gets Republican turnout.
February 24th, 2012 at 12:13 am
1. Dems know a vote for Santorum is a vote for Obama.
February 24th, 2012 at 12:16 am
This will be a 10 point lead within a week. Time for Rick to read the tea leaves and shift resources. Michigan is Mitt country. There won’t be enough Democrats to save Rick’s principle-violating behind.
February 24th, 2012 at 12:17 am
Oh no, Santy is becoming a tool of the liberals…How ironic….
February 24th, 2012 at 12:17 am
Honestly. After Florida I said to myself, this thing is over.
February 24th, 2012 at 12:19 am
5.
So true. The day the Republicans nominate Santorum is the day the GOP commits suicide.
February 24th, 2012 at 12:19 am
Apparently only 4 1/2 Million people saw the debate….which is not surprising given that it was the 20th debate they’ve had. Not only is that about 10 too many, but they’ve all been done by media people, which means that there have been very few questions about economic agendas or Obama’s many failures. You’d never know the country is in the midst of an economic crisis.
Most questions have been about the horse race, insults, SoCon issues, and other irrelevant trivia.
February 24th, 2012 at 12:22 am
10 Dave
The CNN ratings were 4.7 + 1 (west coast rerun) = 5.7
Remember Arizona is Mountain time zone so quite a few opted for the rerun. This is pretty typical for CNN debates (which floats around 5 million viewers). The only times debates have had more than 6 million are when it’s on ABC/NBC networks for those without cable.
February 24th, 2012 at 12:23 am
Mitt should run an ad in MI that highlights the fact that the dems are trying to spoil the election.
February 24th, 2012 at 12:25 am
http://cnnpressroom.blogs.cnn.com/2012/02/23/cnns-gop-arizona-debate-was-1-last-night-with-4-7-million-total-viewers-and-1-4-million-in-key-demo-25-54/
Yeah, CNN South Carolina debate was 5 million, CNN Florida debate was 5.4 million, considering that people from states already voted are probably less eager to watch the whole debate as the time goes on, it isn’t that bad.
February 24th, 2012 at 12:28 am
Husky, Santorum will win this because I just heard several union members from my dad in the Auto workers union that they are planning a huge vote turnout in the G.o.p race and this will, if true,s make the estimate of the Democrats this poll and even PPP’s and Rasmussen significantly undersample the number of Democrats that may end up composing the actual primary electorate.
I hope this is not true because if it is Romney is going to lose and I really do not want that to happen. I fear this may be over if my dad is right. He is a business agent for U.A.W. local 2209. This really pisses me off.
February 24th, 2012 at 12:28 am
I think timezone is definitely a factor for this debate, since the debate started at 5pm in Arizona. Most people aren’t home for work at 5pm. Florida debates naturally had a much larger state tuned in.
But all in all, the viewership was in line with past debates, which again is bad news for Santorum.
February 24th, 2012 at 12:28 am
A couple things…
This pollster already had Mitt up by 2 on Monday. I’m not sure we can conclude a debate bump.
Second, look at their likely voter screen…
In order to be surveyed, voters had to pass the following screen:
In the upcoming February 28 Presidential Primary you can either vote for a field of
Republicans or you can vote in the Democratic Primary for Barack Obama. Whichever
primary you vote in, you have to declare yourself to be a member of that party. The fact
that you declare yourself to be a Republican or Democrat will be public information that
anyone can get. The list of those who vote in the Democratic or Republican Primary will
be made available to list vendors. Knowing this, are you definitely voting, probably
voting, not sure yet, or definitely not voting. Those who said they were definitely not
voting were not surveyed.
Why did they survey those who weren’t sure whether they would vote for a poll of “likely voters?”
February 24th, 2012 at 12:29 am
10. Yep. Sad. Hopefully Mitt and the entire GOP effort can turn the focus back on Obama. Our only chance is to make this a referendum on Obama’s terrible record. But we will need to fight this on multiple fronts as we are already hearing from the MSM how much better the economy is…even though there are 4-5 million less people working than when Obama became president, house prices are still falling, housing is still a mess, a record number of people on food stamps, it goes on and on but you would think contraception is the most important issue facing this country listening to the MSM (including some of the ‘conservative’ talking heads).
February 24th, 2012 at 12:29 am
It looks like Romney needs to suppress voter turnout in Michigan among democrats. I think that if Romney can get his lead up by 7-8 points in the Michigan polls just prior to primary it might discourage dems from wasting their time voting for Santorum.
Still, it’s a smart move by the Obama supporters to come out strong for Santorum.
February 24th, 2012 at 12:32 am
16
If you look at only registered Republican voters, Romney had a 4 point bump since the debate. But that was offset by the Dems going hard for Santorum, so overall he had a 1 point bump.
February 24th, 2012 at 12:36 am
If Romney can get a double digit lead against Santorum before Tuesday, it will suppress the Dem vote. I might be possible. If he got a 4 point bump among Republicans after one day, it could easily grow to 10-12 points after a few more days. It takes a while for the meme to hold. And if news grows that Santorum is only holding on because of the Dem vote, it would I believe galvanize more Republicans to vote against Santorum.
February 24th, 2012 at 12:36 am
I did hear there were very good numbers in early voting in the places where Mitt is strongest. Hopefully, this areas are more motivated to vote FOR Mitt, than the haters are to vote AGAINST Mitt. Clearly none of the others is worth coming out for.
February 24th, 2012 at 12:37 am
19, The margin of error for the poll is +/- 4.7, and probably closer to +/- 6 for the Republican subset. Nothing can be concluded from this when comparing it to Monday’s poll.
February 24th, 2012 at 12:37 am
So, the Obama supporters are coming out for Santorum. interesting
February 24th, 2012 at 12:39 am
The only reason this isnt’ over is the artifical propping up that Santroum and Gingrich are getting from Fox and talk radio… real Americans would never pick those two with out massive free media talking them into it 24/7.
This it the most false election we have ever had… WE THE PEOPLE are having their voice silenced by Fox News and the hate radio.
February 24th, 2012 at 12:40 am
This poll’s political ideology distribution
Moderates apparently had to choose between conservative and liberal, just having those three option might also pushed also few people who would not categorize themselves as “very conservative” in 5 option poll, but I think there is still oversampling of very conservative.
If we don’t look at percentage points, but relative lead to Romney, Santorum now leads more among liberals (29/19 = 1.5263 almost 53% lead) than very conservatives (44/31 = 1.4194 almost 42% lead)
2008 Michigan Republican primary exit poll political ideology distribution.
With this kind of ideology distribution and calculating from cross tabs:
25% conservative (very)
50% conservative
25% liberal
You get almost 7 point lead for Romney.
February 24th, 2012 at 12:43 am
As for the Democrats, it looks like they polled a total of 30. Santorum got 10, Gingrich 6, Paul 5, Mitt 1, 8 Undecided.
They all fall within the enormous margin of error with a subset of 30.
February 24th, 2012 at 12:45 am
24
I think that’s why the debates this year have had such an outsized impact. It allows Americans to hear the candidates themselves, and when they did they realized all the rightwing fawning was BS. Case in point was Perry: “Heartless.” Santorum pretty much pulled a Perry last night. People realized, wait a minute, he doesn’t sound like a true conservative outsider at all, he sounds like a DC insider.
February 24th, 2012 at 12:46 am
PPP will release new numbers according o their twitter account on Monday. They will poll Sunday and Monday for their final poll and Rasmussen will release their final poll on Monday Morning.
February 24th, 2012 at 12:48 am
It appears that the Dems do really intend to try to interfere on the race
February 24th, 2012 at 12:53 am
In their previous poll that gave Romney only 2 point lead and showed the categories “definately”,”probably”,”not sure”,”absentee”, Romney tied Santorum both in definitely and probably category and had huge lead among already those voted absentee, so this isn’t probably that huge factor.
February 24th, 2012 at 12:57 am
Santorum did do better in their previous polls also among Democrats than Romney did.
February 24th, 2012 at 1:17 am
The Arizona Republic endorses Mitt: http://www.azcentral.com/arizonarepublic/opinions/articles/2012/02/23/20120223editoriarl0224-romney-right-these-times.html
February 24th, 2012 at 1:26 am
Mendy Says:
February 24th, 2012 at 12:46 am
PPP will release new numbers according o their twitter account on Monday. They will poll Sunday and Monday for their final poll and Rasmussen will release their final poll on Monday Morning.
In twitter, has good news for Mitt
February 24th, 2012 at 3:08 am
In addition the only evangelical/non-evangelical distribution that gives correct results for every candidate is 50-50, in 2008 evangelicals were 39% of Michigan Republican primary. Now it might be slightly smaller or larger because of rounding within crosstabs, but despite of “very conservative”,evangelical sample Romney does pretty well.
February 24th, 2012 at 3:54 am
I know this my sound too obvious a question but why do folks think that the Democrats who support Santorum are only voting to cause chaos, whereas they may be simply union members who are disenchanted with Obama and be Reagan Democrats?
February 24th, 2012 at 4:23 am
#35…Union members disenchanted with Obama?
They’d have to be going against their union.
It would be nice it it’s true…
February 24th, 2012 at 5:01 am
If anyone is ever interested in computing margins of errors for subsets, the standard 95% confidence margin of error is available as a calculator from ARG. The page is at http://americanresearchgroup.com/moe.html . I’ve used it for five years, and it is extremely useful, when pollsters are late with their margins of error.
In this specific case, if you want to calculate the margin of error for a subset (for example, Right Wing Nut mentioned the Republican subset in posting 22 above), here’s how you would do it.
First plug in the expected vote total in the Republican primary in the Population Size box. I use the Almanac of American Politics as my source for that information, but it’s likely available online in many places: that number was 790,540. Then, in the Sample Size box plug in the number for the Republican subset for this poll, which was 77% of 430 respondents, which translates to 331 respondents. So plug in that number into the Sample Size box.
Now click the Calculate Error button. When you do, you will get the margin of error for that subset. In this case, that margin of error is 5.39. I’ll confess I’m lazy and I usually round it off to the nearest tenth (which many pollsters do, actually).
So the error for the Republican subset is +/-5.4, which translates to the following PSR picture for the Republican subset in this poll:
Romney 36.6-47.4
Santorum 27.6-38.4
Gingrich 2.6-13.4
Paul 2.6-13.4
Undecided 2.6-13.4
Without having looked at the Republican subset for the other nine surveys currently available which agree with this survey (WZZM 2/11-13, Rasmussen 2/13, MRG 2/13-14, ARG 2/15-16, PPP 2/17-19, NBC/Marist 2/19-20, Rasmussen 2/20, EPIC 2/18-21, ARG 2/21-22) and then comparing those numbers with the three surveys taken in the previous universe which disagree with this universe (EPIC 11/13-16, Strategic National 12/8, EPIC 1/21-25), I can’t tell you who’s moved a statistically significant amount among Republicans only. But I’d say, based at least on these current numbers from Mitchell, Romney and Santorum are tied among Michigan Republicans only.
By the way, as Smack1968 will also probably note and as I’ve noted above, once one takes into account the margin of error for this poll, it is not an outlier; it matches comfortably with the other nine current surveys that agree with each other, and therefore reinforces the likelihood that they are all correct. Therefore, once one narrows each candidate’s CPSR to a range on which all surveys agree, it looks like Santorum is still narrowly ahead, since Santorum’s CPSR is now 37.5-37.7% once one plugs in this survey while Romney’s CPSR remains 33.3-34.5%.
Also I wouldn’t be too quick to presume that the results in this latest Mitchell poll is the result of Santorum’s lackluster debate performance. A lot of people did not watch the debate. Their first exposure to that debate won’t be until the Thursday night evening newscasts. So their reaction to it won’t show up in a lot of polling until the surveying that is done today. And that won’t be released until Saturday or late Friday night at the earliest. So I would say we really don’t know yet the result of the debate. This poll certainly does not give us that.
February 24th, 2012 at 5:07 am
New York Times posted an article, while not favorable to Romney, said many things which rang completely true.
Most of the more dedicated race42012er’s would appreciate the insights.
“This has been Romney’s race to lose all along, which is why it is about time for him to stop losing it to the likes of Santorum, who doesn’t even rise to the level of empty suit — call him an empty sweater vest and leave it at that.
Read more: http://www.nydailynews.com/news/election-2012/mitt-romney-biggest-obstacle-race-republican-nomination-mitt-romney-article-1.1027781#ixzz1nIP7e3yH
February 24th, 2012 at 6:26 am
Santorum and Mitt will both be in Michigan today.
Both will be in Michigan tomorrow morning.
Santorum is going to Tennessee tomorrow afternoon, while Mitt will be in Michigan.
Mitt is (someone explain this) going to Daytona International Speedway in Florida on Sunday (wtf?).
February 24th, 2012 at 6:39 am
I don’t think dem turnout will make a huge difference, after Wednesdays debate Newt is going to take a lot of Republican votes from Santorum, this will offset any dem votes Rick gets. Newts good debate performance actually helped Romney. By the way, Mitt won the debate so I think he takes votes from Rick as well just not as many as Newt will take.
February 24th, 2012 at 6:41 am
39. Because everyone will be watching Daytona…great op.
February 24th, 2012 at 6:45 am
Every poll being able to “touch” each other’s results though doesn’t in itself mean that there has not been a change or that one poll isn’t skewed.
The confidence interval of margin of errors is 95%, 5% change that the result is outside the margin of error range, 2.5% change it is below it, 2.5% change that is above it.
So let’s say we have two polls that are able to barely touch each other’s results within the margin of errors range, one of them having need to be on the upper limit of their margin of error, one of them at the lower level of their margin, for results to have been from same sample, no change, no skewing in either of the poll:
0.025^2 = 0.00063
0.063%.
February 24th, 2012 at 7:10 am
39, Daytoa, cars…Michigan.
Duh. Great car-love media op.
February 24th, 2012 at 7:12 am
lol
Santorum surged 3 points since the debate, even in this garbage outlier poll, and Gingrich got NOTHING from the debate. Santorum is going to win Michigan.
February 24th, 2012 at 7:17 am
#44:
On what basis do you believe that?
a) Turnout
b) Reagan Democrats
c) Bleeding of Gingrich support to Santorum
February 24th, 2012 at 7:18 am
Nope, that MoE is too high.
February 24th, 2012 at 7:20 am
44:
If undecided drops from 22% to 11%, it isn’t really surprising if 3 of 4 candidates increase their percentages…
February 24th, 2012 at 7:23 am
this pollster was proven to be an outlier last time and you can count on it being an outlier again in the same direction as last week. What matters is the relationship between Romney and Santorum is basically unchanged, but more importantly that Gingrich got nothing from winning that debate.
February 24th, 2012 at 7:28 am
The #1 question that nobody wants to tackle who believes that Gingrich’s campaign is going to kick into high gear again based on Santorum bleeding massive support:
Other than Georgia where Newt is in the lead, why would any conservative, evangelical or Tea party voter move their vote into Newt’s column knowing he has no chance of winning but as a result could hand the state to Romney?
I might see it in states that are proportional but in winner-take-all states, I can’t see it.
February 24th, 2012 at 7:30 am
44,
Every poll being able to “touch” each other’s results though doesn’t in itself mean that there has not been a change or that one poll isn’t skewed.
The confidence interval of margin of errors is 95%, 5% change that the result is outside the margin of error range, 2.5% change it is below it, 2.5% change that is above it.
That is incorrect. =/- 5% indicates that it could be 5 points higher, or 5 points lower. In this case, a 1 point swing is completely meaningless.
February 24th, 2012 at 7:30 am
#49 going up 3 points is bleeding massive support?
February 24th, 2012 at 7:32 am
re: 50. My comments were directed at 42 (Teemu)
February 24th, 2012 at 7:34 am
Here are the take aways from this poll :
1. Relationship between Romney and Santorum is unchanged (good news for Santorum)
2. Gingrich didn’t get a bumb from the debate (fantastic news for Santorum)
3. Paul got a big bump out of the debate (bad news for Romney)
4. undecideds are breaking 50/50 for Santorum and Romney (bad news for Romney)
Basically the debate changed nothing, and I expect many Gingrich voters when they see it’s going to be close, cross over and vote for Santorum to stick it to Romney.
February 24th, 2012 at 7:34 am
I suspect Teemu layed awake all night trying to figure out how to put as much positive spin on a one point bump as possible.
February 24th, 2012 at 7:36 am
53, If Gingrich were smart, he’d urge his supporters to vote for Santorum in Michigan.
February 24th, 2012 at 7:36 am
#51:
My question was a theoretical one and not based on current polling. The media is trying to set the narrative that Newt is going to rebound at the expense of Rick. My question is why, if it means cutting your own throat and allowing Romney to prevail.
February 24th, 2012 at 7:37 am
9 point lead with Republicans is all that matters.
February 24th, 2012 at 7:39 am
57, I thought Mitt was going to beat Obama by winning over Independents and conservative Democrats?
February 24th, 2012 at 7:53 am
That was not comment to one point swing, but in general to this thinking that every poll touches each other, so they are all within margin of error which is fine. Here is one somewhat theoretical example based on the sub sample of this poll, no rounding taken into account etc.
Let’s take the 311 Republican sub sample, which had +/-5.4 margin of error (the normal confidence interval of 95%)
Romney 42
Santorum 33
Now without rounding margins, set the margin of error at +/-4.5, so that we get the 42 – 4.5 = 33 + 4.5
Interpolate the new confidence interval for this margin of error.
http://www.raosoft.com/samplesize.html
You get 88.7%, 11.3% that it is within margin of error, 5.65% at the below the lower margin, 5.65% above the upper margin.
Now the probability for Santorum being at 37.5 or above it and Romney being at 37.5 or below it is, 0.056^2 = 0.0032, 0.32% percent.
February 24th, 2012 at 7:59 am
Romney stands a good chance of winning AZ and MI on Tuesday, but I wonder if he
may have lost the messaging war going forward. Santorum has succeeded in moving
Romney toward him on policy – on taxes Romney is now revising his plan to cut more,
on immigration Romney is now calling AZ a model, and on social issues Romney is
now talking more about religion and even contraceptives. Those are all areas
more favorable to Santorum.
February 24th, 2012 at 8:04 am
43
People in Michigan don’t watch NASCAR, do they? They certainly don’t over here, lemme tell ya.
February 24th, 2012 at 8:05 am
Either way it looks like MI will not be called as soon as the polls close.
Bring it Mitt, bring it .
February 24th, 2012 at 8:05 am
Rasmussen: Romney leading by 6 points in Michigan
This also from Feb 23
Romney 40
Santorum 34
http://www.rasmussenreports.com/public_content/politics/elections/election_2012/election_2012_presidential_election/michigan/2012_michigan_republican_primary
February 24th, 2012 at 8:12 am
63
WOOOOOO-HOOOOOOOO!!!!!! NICE!
February 24th, 2012 at 8:14 am
Killjoy, please spin 63.
Kthxbye.
February 24th, 2012 at 8:16 am
59,
I’m no statistician, but I believe that formula works fine if respondents are given 2 options, such as as “yes” or “no” question. Multiple options complicates it somewhat.
February 24th, 2012 at 8:21 am
Finally a Rassmusssen poll showing Mitt leading in MI.
February 24th, 2012 at 8:21 am
The ideological distribution is pretty conservative compared to 2008 and still clear victory
Romney – Santorum
36% very conservative (28 – 51)
38% somewhat conservative (49 – 28)
26% other (43 – 20 )
February 24th, 2012 at 8:23 am
I seriously doubt Santorum leads among Democrats, 36-3, but I’m sure he has a large lead, and it’s not (primarily) Operation Chaos.
Michigan is the symbolic capital of the Reagan Democrat, a constituency that several of us have said Rick should do well with.
February 24th, 2012 at 8:25 am
This particular poll doesn’t show much impact from the debate. Mitt was leading it by 2 last week, and 3 this week. A one point shift is noise.
February 24th, 2012 at 8:26 am
70
See #63
February 24th, 2012 at 8:30 am
Let the surge begin!
February 24th, 2012 at 8:30 am
MassCon,
Right. I said “this particular poll..”.
February 24th, 2012 at 8:32 am
From the cross tabs
Certain to vote: Romney leads 42 – 35
Certain of who they are going to vote: Romney leads 45 – 36
February 24th, 2012 at 8:37 am
73
And my point was, there actually is a trend.
February 24th, 2012 at 8:43 am
MassCon,
Okay.
February 24th, 2012 at 8:45 am
BOOM!!!!!!
MITTMENTUM BABY!!!!!
February 24th, 2012 at 8:46 am
One poll shows a trend so far, the other doesn’t. I’m not saying there is a trend or not. I merely pointed out that this poll doesn’t show a trend.
We need a reputable tie breaker poll, before we start declaring there is, or isn’t, a trend following the debate.
February 24th, 2012 at 8:54 am
What a fantastic media catch from Michigan if Dems vote for Santorum. A full week before Super Tuesday and the published names of all Democrats who vote for Santorum. wOW! Publish those names in a campaign spot. Santorum supported by Obama Democrats.
February 24th, 2012 at 8:56 am
Matt…..Rasmussen has Mitt up by about 7 in Michigan this AM
February 24th, 2012 at 9:50 am
Craigs,
Right. That’s the poll that created the “tie” with this one, that I was referring to.
February 24th, 2012 at 12:13 pm
In Michigan there is a known factor in the race, some Dem’s or going to cross over and vote for Santorum. We, then have to draw the conclusion, the Dem’s want to run against Santorum not Romney. Why would anybody that claimed to be a Republican and wanted to have a chance of beating Obama vote for a candidate the Dem’s want to run against and are going to beat easily(Santorum)? Does talk radio have that much influence over these sheep?
February 24th, 2012 at 12:29 pm
69
There aren’t any Reagan Democrats left. After 8 years of Reagan, most of them converted to the Republican Party like Rick Perry did.
2012 election is more about winning the suburban Obama Republicans/Independents than about winning the extinct Reagan Democrats.
February 24th, 2012 at 3:43 pm
@69
Santorum is down 55-29 in marist poll out last week against obama in michigan.
Romney is down 51-33.
Santorum isn’t getting any reagan democrats because they don’t exist anymore.
Reading the detriot news sites democrats are flooding the primary to vote for santorum to hurt the gop and saying they will vote for obama in november.
Romney went from down 10 among republicans to plus nine yet Romney only went from plus 2 to plus 3.
I think the seven percent for dems in this poll is understating their effect. Obama doesn’t have a contest and last primary even with the dems having a semi primary had a larger dem turnout.
With the number of dems still undecided in this poll and unions pushing them to vote for santorum and dem officials getting the word out this could put him over the top.
Santorum’s statements that have hurt him the last week actually help him in michigan because it gets dems even more enthused to vote for him.
This mischief isn’t a small matter as it went from mccain losing michigan to winning by seven and usually doesn’t show up until the vote.
This will be spun as some great santorum win when it will be because of mischief. What is worst the media will make this romney’s waterloo and say he lost the nomination. All this because of rats.