A new Senate poll by Rasmussen released on Wednesday gave a boost of confidence to Norm Coleman’s re-election campaign, with the Senior Senator from Minnesota leading Al Franken by a 43 to 39 percent margin. Dean Barkley received the support of 14 percent of the 500 likely voters surveyed on October 28th.
Franken supporters would be wrong to dismiss this poll as an outlier. In fact, as Barack Obama has surged into a consistent double-digit advantage over John McCain in recent polls of likely voters in the Gopher State, there is clear evidence Franken has failed to capitalize on Obama’s momentum in kind.
Smart Politics studied thirteen polls of the presidential and U.S. Senate races in Minnesota that were released in the month of October and measured the relative spread between Obama vs. McCain and Franken vs. Coleman. The week-by-week trends in October find support for Franken is not keeping pace with Obama’s surge: in fact, Franken seems to have lost his grip on Obama’s coattails.
In the first week of October, Obama’s advantage over McCain was just 7.3-points greater than Franken’s position against Norm Coleman (Pollsters: SurveyUSA, Star Tribune, Minnesota Public Radio, Rasmussen). Obama averaged a 9.5-point lead over McCain while Franken averaged a 2.2-point lead over Coleman. (With a strong third party candidate in the Senate race, Dean Barkley, it is to be expected that Obama would have a stronger lead than Franken from the onset, as the Senate race is more ‘compressed’).
In both the second and third weeks of October, Obama’s spread over McCain was 9.0 points higher than Franken’s position against Coleman (Pollsters: Quinnipiac, Research 2000, Star Tribune, SurveyUSA, National Journal). Obama led by an average of 10.2 points over McCain while Franken led Coleman by just 1.2 points.
As the fourth week of October closed, the difference between Obama and Franken ballooned to 13.5 points – nearly double the difference from the first week of the month (Pollsters: St. Cloud State University, Big 10 Battleground, and two polls by Rasmussen). Obama now leads McCain by an average of 12.75 points, while Franken trails Coleman by an average of 0.75 points.
While Minnesotans seem to have rallied around Barack Obama in the presidential race, it appears they may only have so much appetite for 'change'; there simply may not be enough of it to go around to send both Obama to the White House and Franken to D.C.
I would seriously like to see your own polls on the senate race, rather than commentary about other polls.
When you did the poll that you released today, did you also find out about the senate race?
Please share.
The HHH / MPR Senate poll results will be released on Friday.
Don't you expect Barkley's support to fade on election day?
I thought it was one of those beloved principles of political polling (right up there with incumbents below 50% and the Bradley effect) that third party candidates always poll much higher than they actually perform on election day.
In that case, the important question is which way will those voters go on Tuesday?
Hi JKruse- the fade effect you are talking about is mentioned in Third Parties in America. The problem is that the conclusion in that book is drawn on seven presidential races across eighty years, lopping in Lemke (36?) who was only at 3% (can't tell movement with that percent), George Wallace, a regional candidate which is bizarre by itself, with Anderson, and Perot, (there is a candidate from before Lemke I believe).
I saw from a conference in Toronto that a group of Harvard political scientists said that Third parties in America was not robust in its treatment (it is a start...we can give it that).
My analysis of IP candidates, using the same methodology as Rosenstone's (not one that I am keen on but it is his), shows that it is actually random chance that a candidate goes up or down. My group of candidates is more than twice as large as Rosenstone's, shares modern polling methodology and similar ballot access and funding (for state candidates).
A more robust analysis of strategic voting, which is what you are talking about, can be found in political scientists such as Michael Alverez, Fred Beohmke and a couple others. They have studies the effect over and over again. I have tried to convince researchers at the U of M to do the same, but nothing so far.
The fade effect can happen, it doesn't always, and it is tied in with many variables and nuances of strategic voting. Interestingly, Professor Jacobs noted in this last poll that Barkley was still gaining support, something unusual for an IP candidate at this state. I say it is random chance.
One of the interesting things I find fascinating in polling methodology is the estimate of the electorate size. Different polls can use different assumptions. I would caution you from looking across polls this year...that could be dangerous.
Peter Tharaldson
IPMN
Maybe the citizens of Minnesota decided that electing Ventura was enough of an embarrassment, and that Franken would make the state a laughing stock.