A tale of two polls: Santorum may win Michigan Primary

Two different polls paint very different pictures for Tuesday’s primary in Michigan. The PPP Poll released February 26ths puts Romney ahead of Santorum and makes a very solid argument that Romney is ahead and that it will be difficult for Santorum to move enough voters into his camp to take the lead. The Mitchell Research poll, released on February 27th, makes a good argument that although Romney was ahead as of last Thursday, Santorum has in fact moved enough voters into his camp to be numerically ahead of Romney by 2% points in a poll with a 3.34% margin of error.

Let’s have a look at the details.

PPP makes the point that 16% of Michiganonians have already voted in the primary, and among these, Romney has a 62-29% advantage over Santorum, making Tuesday’s vote an uphill battle for the latter. Among those who are likely to vote, according to this poll, Santorum leads Romney 39-34%. This indicates an upward move for Santorum, but the pollsters claim that it is not enough to put Santorum in the lead. The PPP poll predicts that Romney will yield 39% and Santorum 37%, and the margin of error for that poll is 4.8%.

PPP points out that during the days leading up to last weekend, Santorum’s favorability declined from 44% to 15%, while Santorum’s negative attacks on Romney has not affected Romney’s favorability rating very much.

The Mitchell Research poll is more recent. Apparently, Santorum has made significant inroads with Michigan voters. In Mitchel’s Thursday evning poll, Romney had reduced Santorum’s 16% lead among Tea Party voters, but over the weekend, the numbers have shifted to give Santorum a 6% lead in that group. Romney had reduced Santorum’s lead among Evangelic ls for 16 to 7%, but over the weekend Santorum’s lead has improved to 19%. A similar pattern was seen in other demographic groups. Polling finished yesterday puts Santorum at 37%, up from 30% last Monday, with Romney at 35%, having registered 32% on Monday and 36% on Thursday.

During this time, since last Monday, the percentage of undecideds according to the Mitchell Poll has gone from 22% to 10%.

So, according to the Mitchell Poll, Santorum now leads Romney by 2% in a poll with a margin of error of 3.34%, with 10% undecided, and with a disproportionate member of those undecideds having broken towards Santorum over the last few days, and a general shift towards Santorum and away from Romney in most demographic groups.

While the Mitchell poll is inconclusive and technically shows the contest as a two-candidate horse race, it is not inconceivable that even as we speak Santorum support has gained a couple of percentage points to move beyond statistical uncertainty. As 12% of undecided voters melted away, Santorum moved ahead 7% while Romney moved ahead 3%. Assuming a 2:1 ratio, the remaining 10% of undecided voters could give Santorum (conservatively) a 5% or more increase and Romney a 3% increase. This would predict that the final outcome will be Santorum at 42%, Romney at 38%, with Gingrich and Ron Paul in a virtual tie for third.

I should also mention that while Gingrich’s support remained steady at 9% in the Mitchell Poll since Monday, Paul’s numbers jumped form 7 to 12 to 8% during this time. It is possible that some of Romney’s loss was from Paul supporters shifting towards Santorum as well.

Going into Tuesday’s poll, I’m going to have to predict a Santorum victory in Michigan. Since Michigan is Romney’s “home state” this will be extra-significant.

This all assumes an equivalence between PPP and Mitchel polling, and accuracy of polling that is reasonable to expect much, but not all, of the time.

The polling data is available via Real Clear Politics (download the PDFs there)

Share and Enjoy:
  • Twitter
  • StumbleUpon
  • Facebook
  • Digg
  • del.icio.us
  • Yahoo! Buzz
  • Google Bookmarks
  • LinkedIn
Tagged , , , , . Bookmark the permalink.

5 Responses to A tale of two polls: Santorum may win Michigan Primary

  1. Phillip IV says:

    This all assumes an equivalence between PPP and Mitchel polling, and accuracy of polling that is reasonable to expect much, but not all, of the time.

    It seems the Mitchel poll had about twice the sample size of the PPP one. And I’d also grant them something of a local advantage – in local contests the smaller, local polling outfits often have a better track record that bigger national ones, because they have a much more precise grip on the demographics of the area and the prediction of turnout.

    So, yeah, it looks like Rick just might pull it off after all.

  2. rork says:

    But what do the Minnesotaonians think?

    Also, it depends on how many Democrats show up. Now that we can’t try to nominate Bachman for them, it’s taken some of the fun out of it.

  3. rork says:

    I was wrong. Bachmann is on the ballot in MI. Some Democrates may be opting for Santorum instead, out of anti-Romney enthusiasm. They aren’t organized though.

  4. Pingback: Michigan Republican Primary Results Prediction

  5. rgcheek says:

    Good effort, but I am not sure that the GOP is led by honst brokers any more, if they ever were.

    If Romney wins the noomination, I will vote for Obama, and I will look for a third party that represents reflective conservative thought based on our old Christian Western Civilization.

    If conservatism is not about doing that, then it isnt worth a bucket of spit.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>