Independent and third-party candidates face overwhelming challenges in getting voters’ support, even if they polling well. Donald Trump’s polling numbers seem okay in the primary, but abysmal in the general election. And a research company is experimenting with polling people who dial wrong numbers. This is HuffPollster for Friday, April 15, 2016.
VOTERS ARE UNLIKELY TO RALLY BEHIND A THIRD PARTY CANDIDATE - Matt Dabrowski (R): "Pollsters regularly ask about third-party support, often in one of two ways. One way is to ask voters if they’d support a third-party candidate in some kind of yes-no question….Another way is to add the third-party candidate to the general-election horse race question….Pollsters know that both these question styles aren’t very predictive….But there is a way to accurately gauge potential third-party support using survey research…. A potential third-party voter must reject both the Republicans and Democrats (1,2) and both the Republican and Democrat nominees (3,4). In effect, voters have to say ‘no’ four times before they’re truly open to stepping outside the two-party system, making this a four-part test….This potential pool of third-party support will wax and wane with the nominees’ favorables. And the odds are that disgruntled primary voters today will become loyal partisans in November." [HuffPost]
Post-election analysis confirms that polling non-major party candidates is tough - A comparison of polling and actual results in the 2014 midterms, presented last year by exit pollster Joe Lenksi, found that if pollsters included the third candidate in their election tossup questions, the numbers overestimated support for that candidate. If they didn’t ask about the third candidate, though, the polls underestimated the vote share. The bottom line: it's very difficult to poll on independent and third-party candidates. [HuffPost]
REPUBLICANS ARE LESS SURE ABOUT VOTING IF THEY DON’T LIKE THE NOMINEE - Mark Blumenthal: "Seven out of ten registered voters (70%) describe themselves as 'absolutely certain' to vote on a Trump-Clinton contest, with 80% at least 'very likely' to vote. Slightly fewer (67%) say they are 'absolutely certain' to vote if Clinton faces a Republican other than Trump, with 81 percent at least 'very likely' to vote. However, these overall results mask shifting preferences among about a quarter of registered voters, with potentially important consequences for the ultimate Republican nominee….[A] quarter of Donald Trump’s backers (25%) rate their chances of voting as 50–50 or less if he is not the nominee, while a similar share of Cruz and Kasich supporters (24 and 27% respectively) say they are 50–50 or less if Trump gets the nod. Under either scenario—Trump or another Republican as nominee—the results reverse the usual pattern by party. Typically, the most likely voters are slightly more Republican than other registered voters. In this case, however, the potential hesitance of either Trump or #NeverTrump voters gives Democrats a slight edge on self-reported vote likelihood under either hypothetical scenario." [SurveyMonkey]
TRUMP’S POLLING NUMBERS ARE SIMULTANEOUSLY AWFUL AND OKAY - Greg Sargent (D): “The basic case for nominating Ted Cruz rather than Donald Trump is that, while Cruz would arguably be the most right-wing nominee in modern American history, and would probably lose to Hillary Clinton, he would not unleash the sort of blood-dimmed tide of down-ticket destruction that Trump would. A new Washington Post/ABC News poll lends some support to this view: It finds that Trump’s numbers are unspeakably awful among all the voter groups that Republican strategists had hoped to improve the party’s performance among in 2016.” [WashPost]
... But on the other hand - Philip Bump: “[A] new CBS News poll suggests that Republicans are pretty comfortable with the idea of Trump as the nominee -- or, at least, they are compared to the other two candidates….Republicans are slightly more likely to say that they would support Trump's nomination 'enthusiastically' than they were last month. Overall, two-thirds of Republicans say they'd support him enthusiastically or with some reservations -- a higher percentage than says that about Cruz or Gov. John Kasich. More Republicans are likely to say they'd support Kasich simply because he would be the nominee than any other reason.” [WashPost]
So which is it -- do Trump’s numbers show “unspeakably awful” or “pretty comfortable”? - Both are true. Trump’s numbers are awful among the general public and the minority groups that are discussed in Sargent’s piece. Among just Republicans -- the group Bump is discussing -- his support has declined overall, but his favorability ratings are higher than either of his opponents. The two pieces and the polls they discuss highlight a substantial gap between how Republican voters and general election voters feel about Trump and his potential nomination
‘RECONNECT’ POLLING ROUTES WRONG NUMBERS TO POLLS - Eric Chemi and Nicholas Wells: "Reconnect Research works with technology that accepts inbound calls and turns them into research and polling opportunities. One problem with traditional polling is it depends on respondents answering calls from unknown numbers and using landlines. With cellphones and caller ID, most people won't answer their phone if a random person or unknown number is calling. People hardly pick up landlines at all. However, people still have to call important numbers like banks, insurance or credit card companies. When people call those numbers and misdial, Reconnect can send that call to a researcher….Reconnect Research has been continuing to prove out its thesis. One example of this — the results of the Florida primary. Reconnect's polling was the most accurate in measuring Democrat Hillary Clinton's performance...Similarly for John Kasich's number. Reconnect's poll was number one out of 10 polls analyzed." [CNBC]
It’s hard to assess claims of accuracy on polling released after the election - It doesn’t appear that the Reconnect Research polls were publicly released prior to the elections in Florida. As with any new way of doing polls, skepticism is warranted until there’s evidence that the method performs well over time.
HUFFPOLLSTER VIA EMAIL! - You can receive this daily update every weekday morning via email! Just click here, enter your email address, and click "sign up." That's all there is to it (and you can unsubscribe anytime).
FRIDAY'S 'OUTLIERS' - Links to the best of news at the intersection of polling, politics and political data:
-0ptimus Consulting (R) finds that Donald Trump could lose up two dozen delegates in the New York primary. [Politico]
-Frank Newport writes that Hillary Clinton's image among Democrats is at a new low. [Gallup]
-Kyle Kondik and Geoffrey Skelley explain how a downballot Democratic wave could happen. [Center for Politics]
-Drew DeSilver finds that candidates who don't win on the first ballot at conventions are likely to lose. [Pew]
-Jacob Smith finds that a North Carolina politician who put her picture in every elevator in the state may have gotten an electoral a lift. [WashPost]
-Americans would rather pay lower prices than purchase items made in the U.S. [AP]