Martha Coakley sees her initial lead slip away in a Massachusetts race...again. Surveys disagree on candidates' popularity in North Carolina and elsewhere. And few people feel personally affected by the campaign against ISIS. This is HuffPollster for Tuesday, September 30, 2014.
GOVERNOR'S RACE CLOSES IN MASSACHUSETTS - Massachusetts is blessed with more active local pollsters than most other states, and they have produced four new polls since last week showing a very close race between Democratic Attorney General Martha Coakley and Republican Charlie Baker. Three of the four new surveys -- Boston Globe/Social Sphere, UMass Amherst/WBZ/YouGov and Western New England University -- gave Baker the edge by a percentage point or two, while a Boston Herald/Suffolk University poll gave Coakley a 1 point advantage. Coakley had enjoyed wide leads in polls conducted earlier this year. The HuffPost Pollster poll tracking model, based on all of the public survey data, still gives Coakley a roughly three percentage point advantage as of this writing (46.4 to 43.4 percent), partly because it also gives weight to the previous polling. However, the trend lines show Coakley's lead contracting significantly, and the model currently estimates her probability of winning in November as just 52 percent. [Pollster Massachusetts chart, Boston Globe, UMass Amherst, WNEU, Suffolk]
'A sense of déjà vu'? - Harry Enten: "You probably remember Coakley as the Democrat Scott Brown beat in the 2010 Massachusetts special Senate election. A month before that election, Coakley led every poll by at least 25 percentage points....As MSNBC’s Steve Kornacki correctly pointed out, three of Massachusetts’s past four elected governors have been Republicans. So, it’s not entirely clear that Coakley is at fault for her vanishing lead this time around. Still, you can’t fault Democrats for feeling a sense of déjà vu." 
One of many close governor's races - The tightening race in Massachusetts brings to nine the number of races for Governor where the top two candidates are separated by 3 percentage points or less according the HuffPost Pollster poll tracking model. [HuffPost Governor's Forecasts]
POLLS TELL DIFFERENT STORIES ON NC CANDIDATES' IMAGES - Nate Cohn: "The conventional view holds that Ms. Hagan has a lead because Mr. Tillis is unpopular….There’s one weakness in this narrative: the amount of evidence supporting it. A USA Today/Suffolk poll in August found Mr. Tillis’s favorability rating at minus -14, but there isn’t much other evidence to back it up….Enter the CNN/ORC poll on Sunday. It found that 47 percent had a favorable impression of Mr. Tillis, and 40 percent had an unfavorable view. It didn’t find Ms. Hagan faring much better than she has in other polls: 46 percent favorable, and 47 percent unfavorable. If there were more surveys showing Mr. Tillis as unpopular as the conventional view, then perhaps we could discount the CNN/ORC poll as an outlier. But in the absence of more evidence to the contrary, one should at least be open to the possibility that Mr. Tillis is more popular than was thought. That opens the door to a different view of the race: that Mr. Tillis is poised to narrow or eliminate Ms. Hagan’s lead behind undecided voters who view him favorably and who view Ms. Hagan and President Obama unfavorably." [NYT]
Why favorability ratings can vary - Comparisons across polls can be tricky with candidate favorability ratings, since differences in question wording and order can produce highly inconsistent results. Such questions typically ask whether voters have a "favorable" or "unfavorable" impression, but they vary widely in the degree to which they prompt respondents to say they are unfamiliar with the individual or have a middle or "mixed" opinion. Pollsters who make it easier to express uncertainty or neutrality typically show lower numbers in the positive ("favorable") category. Also, many pollsters opt to ask the question just after asking respondents who they'll vote for, which effectively informs respondent of a candidate's party affiliation and the office they're seeking -- information they may have lacked if the pollster had asked the favorable ratings first. [See Huffpollster on favorable ratings]
Favorable rating house effects in NC? - These issues of order and question format may play a role in the measurements of Thom Tillis' popularity. Two earlier polls conducted by Suffolk University and the Democratic firm PPP both asked the favorable ratings first, while the new CNN/ORC poll asked it later, just after the vote questions. While both Suffolk and CNN gave Tillis similar unfavorable ratings, CNN found a much higher number of voters with a positive opinion of him. Republicans, especially were far more likely to give Tillis a positive rating in the CNN poll (88 percent, compared to 43 percent in Suffolk, although that comparison too may be complicated by differences in the way pollsters asked about party leaning). Tillis is likely gaining recognition and favorability with Republicans, but it is difficult to be certain how much given the differences in how these pollsters measure candidate popularity. [See CNN, Suffolk and PPP crosstabs]
Not the only race where favorability is unclear - Recent polling in other states also shows a substantial variation in candidates' ratings. Take Colorado, where five non-partisan polls released in August and September tracked favorability ratings among likely voters, and showed little consensus as to which candidate's image is better. Quinnipiac shows a massive favorability gap between Sen. Mark Udall (D) and Cory Gardner (R), with Gardner's net rating 18 points higher than Udall's, while SurveyUSA and Suffolk/USA Today show smaller gaps of 7 and 5 points. Rasmussen and NBC/Marist's polling, meanwhile, find Udall better-liked, by a 2 and 5 point margin, respectively.
The same dynamic applies in Iowa, where the six most recent non-partisan polls to release favorability results range from a 9-point net favorability edge for Republican Joni Ernst (Quinnipiac), to a 5-point edge for Bruce Braley (Suffolk/USA Today, CNN), with the remainder (Rasmussen, Loras College, and the Des Moines Register) finding the two about even.
LOUISIANA 'TILTING DEMOCRATIC' IN 2014 - Justin McCarthy: "In 2014, more Louisianans identify themselves as or lean Democratic (45%) than Republican (41%), a shift from the slight edge Republicans have held for past three years. The shift in party preferences is likely a welcome indicator for Democratic Sen. Mary Landrieu as she attempts to win her fourth term in one of this year's most highly watched U.S. Senate races. Still, the Democratic Party's current edge in Louisiana is smaller than its 10-point advantage in 2008, when Landrieu last won re-election...Although Democrats currently outnumber Republicans in the state, Louisianans are most likely to describe their political views as conservative (45%), rather than moderate (34%) or liberal (17%)." [Gallup]
FEW SEE PERSONAL STAKES IN CAMPAIGN AGAINST ISIS - HuffPollster, with Emily Swanson: "Americans, who largely support airstrikes against the Islamic State, don't think that the military campaign will have much of an impact on their lives, according to a HuffPost/YouGov poll released Monday. Just 24 percent of respondents predicted that they will be affected a great deal or a fair amount by the airstrikes in Iraq and Syria, while 60 percent said the strikes will have not much or no effect on their lives. It may not be surprising that Americans feel unconnected to a remote conflict that has, thus far, been fought only from the air. But this phenomenon is not limited to the current air campaign: In 2011, a Gallup/USA Today poll found that only 36 percent of respondents believed their lives had been affected by the 10 years of war in Iraq and Afghanistan, while 62 percent reported little or no impact. (By contrast, in 1993, Gallup found that 56 percent of Americans said they had been personally affected by the Vietnam War, while 44 percent had not.)" [HuffPost]
HUFFPOLLSTER VIA EMAIL! - You can receive this daily update every weekday morning via email! Just click here, enter your email address, and and click "sign up." That's all there is to it (and you can unsubscribe anytime).
TUESDAY'S 'OUTLIERS' - Links to the best of news at the intersection of polling, politics and political data:
-Gallup continues to find Americans divided about how active the federal government should be. [Gallup]
-Melody Crowder-Meyer and Benjamin Lauderdale look into why the GOP doesn't have more female candidates running. [WashPost]
-A significant gender gap continues to favor Democrats in congressional voting. [YouGov]
-Democrats are spending more on staff and voter contact in key Senate races. [NYTimes]
-Charlie Cook still sees Republicans as holding the advantage in November. [National Journal]
-Philip Bump breaks down which TV shows each party prefers to air ads during. [WashPost]
-Matt Barreto and Gary Segura say both political parties, but especially Republicans, are missing opportunities to the growing Latino electorate. [NBC]
-Republicans really don't like Eric Holder. [HuffPost]
-Brendan Nyhan reports on how on social media, false rumors spread faster than truth. [NYTimes]
-Here is a word cloud made entirely of words about clouds. [Twitter]