Yesterday I noticed some interesting results in the Marquette University Wisconsin poll. As in some other polls, many more voters express unfavorable than favorable impressions of Mitt Romney. In Wisconsin his favorable/unfavorable are 32%-47%, a negative showing indeed. Barack Obama’s favorable/unfavorable numbers are much better: 52%-43%. Yet when respondents are asked which candidate they would vote for, Obama leads Romney by only 48%-43%. In other words, a lot of voters who have unfavorable feelings toward Romney or who say they don’t know enough about him to say are nevertheless ready voting for him. And at least a sliver of those who have favorable feelings toward Obama aren’t voting for him when paired against Romney.
So I checked the national polls in the realclearpolitics.com summary of Obama-Romney pairings for the favorable/unfavorable numbers for each. The following table shows the Obama-Romney pairing, with Obama’s percentage listed first, and the favorable/unfavorable numbers for Obama and Romney. The polls are listed in the order in which interviewing was conducted, with the most recent poll at the top.
Poll O vs. R Obama Romney
Rasmussen 44-46 n.a. n.a.
CNN-ORC 54-43 56-42 37-49
McClat-Marist 46-44 n.a. 45-45
PPP 48-44 n.a. 33-58
Fox News 46-42 50-47 39-49
Reuters-Ipsos 52-41 n.a. n.a.
Bloomberg 47-47 52-45 42-48
CBS/NYT 47-44 n.a. 37-30 “favorable/not favorable”
Pew Res. 54-42 56-41 29-51
In the seven polls which have both a pairing and a Romney favorable/unfavorable question, the average result of the pairing is 49% for Obama and 44% and the average favorable/unfavorable rating for Romney is 37%/47%. The question wording and position in the poll of the favorable/unfavorable questions differ from poll to poll (as suggested in the note above, CBS/NYT offers the options of “favorable,” “not favorable,” “undecided” and “haven’t heard enough to have an opinion yet”) results may not be commensurate. Nevertheless it appears that Romney is getting the votes of at least 7% of those who do not have a favorable impression of him.
Several possible explanations come to mind. (I would list them with numbers, but Microsoft Word annoyingly insists of indenting each paragraph when I do so.)
First, the election is a referendum on the incumbent president. Note that Obama’s numbers against when paired against other Republicans are not much different from his numbers when paired against Romney.
Second, a variant of the above: Some very conservative voters don’t have a favorable impression of Romney but are determined to vote for just about anyone against Obama.
Third, Republicans are different from Democrats. Democrats like to think of their presidential nominees as philosopher-kings, clearly smarter than their opponents, more charismatic and sophisticated. Republicans tend to regard politicians as tools to achieve certain ends, and don’t care much about charisma or likeability. They are the party that nominated Richard Nixon for national office five times.
Whatever the explanation, Romney’s high unfavorables seem to be hurting him in pairings against Obama, and perhaps enough to make the difference between victory and defeat, but they aren’t disqualifying him for many voters. He outpolls his favorable, sometimes significantly, in six of the seven polls where we can compare them. In contrast, in the four polls where we can make comparison, Obama tends to underpoll his favorable. Some small number of voters like him but aren’t willing to vote for him, even against a candidate many of them don’t find likeable.
