Sarah Palin leads a hypothetical Republican primary in Alaska. The media vs. campaign pollster methodological divide is especially pronounced in New York City. And the Onion reminds us that almost none of it actually matters. This is HuffPollster for Tuesday, July 30, 2013.
PALIN LEADS AMONG ALASKA GOP, BUT TRAILS AGAINST BEGICH - Luke Johnson: “Former Alaska Gov. Sarah Palin (R) is Alaska Republicans' top choice to challenge Sen. Mark Begich (D-Alaska) in 2014, even though she would perform poorly against him in a general election, a Public Policy Polling survey released Tuesday found. Palin was the choice of 36 percent of the Republican primary voters surveyed, but Begich would beat her in a hypothetical match-up, 52--40 percent. While Palin is well-liked among the GOP base, she is unpopular with voters overall, with 58 percent viewing her unfavorably and 39 percent viewing her favorably. Only 47 percent of voters still consider her to be an Alaskan, while 46 percent do not.” [HuffPost, PPP release]
SAN DIEGO MAYOR BOB FILNER ON SHAKY GROUND - SurveyUSA: “San Diego residents have two things to say to Mayor Bob Filner. One: Therapy won't work. Two: While in therapy, he should temporarily step aside.” Filner is checking into a behavior counseling clinic for two weeks of intensive therapy, but 70 percent of San Diegans don’t think he’ll change his behavior. Seventy-one percent said he should temporarily step down while in treatment, while just 24 percent thought he should remain mayor. [SurveyUSA]
A separate, earlier poll also found support for a potential recall - 10News: “60 percent of San Diegans surveyed believe Mayor Bob Filner should be recalled if he does not resign....67 percent of people thought he should resign, 22 percent thought he should remain in office and 11 percent were not sure. The poll also asked, ‘If Filner does not resign, do you think he should be recalled?’ 60 percent said yes, 28 percent said no and 12 percent were not sure.” [10News]
DID WEINER LISTEN TO HIS (FIRST) POLLSTER? - In a look back at the NY Times Magazine article that recounted Anthony Weiner's "fall from Democratic prominence, his sort-of retreat into domestic life and his initial steps back into politics," the Washington Post's Erik Wemple remembers a passage about a survey conducted for Weiner by Democratic pollster David Binder: "The focus of the poll, Binder says, was the question 'Are voters willing to give him a second chance or not, regardless of what race or what contest?' And the answer? 'There was this sense of 'Yeah, he made a mistake. Let’s give him a second chance. But there are conditions on that, and there are a couple of things we’re going to want to know: What have you been doing since this incident occurred? Did you learn anything from this mistake? How did you deal with it?’ They want to know that they’ve put it behind them.'" Candidate Weiner later replaced Binder with Paul Maslin, another Democratic pollster, earlier this year. [WaPost, Capital New York]
NYC: TWO DIFFERENT POLLING APPROACHES AT WORK - Any effort to survey voters in New York City presents a greater degree of difficulty than other races. Survey response rates are typically lower in urban areas, and New York has long had a reputation among pollsters as being one of the toughest placed in the nation to get respondents to stay on phone. Add to that the typically low turnout of the Democratic mayoral primary, and the challenge facing the various pollsters is significant. Faced with that challenge, however, news media and campaign pollsters have chosen very distinct approaches to the way they have sampled the race.
Media pollsters sample random digits - All but one of the 20 surveys logged into the Pollster chart on the mayor's race since April 2012 have been conducted by three organizations: the Quinnipiac University Polling Institute, the Marist College Institute for Public Opinion (sponsored most recently by NBC--4 and the Wall Street Journal) and Siena College (sponsored by the New York Times). All three have used a similar methodology for sampling voters: They begin with separate samples of landline and mobile telephone numbers in which the last digits of the numbers are randomly generated (the term of art is "random digit dialing" or RDD). They then interview samples of all adults, weight the complete sample to match census demographics of the full population, and then select registered Democrats and (more recently) "likely voters" to use to measure voter preferences and other attitudes on the mayor's race. [HuffPost chart & NYC polls]
Why RDD? - The advantage of the random digit sampling method used by media pollsters is that they can reach all working phones, listed and unlisted, landline and mobile. But that wide coverage also comes with some downsides: One is that they must rely on "self-reports" -- i.e. questions answered by respondents -- to identify registered and likely voters. As demonstrated in last year's presidential election, that process is imperfect. Another is that the turnout of Democrats is likely to be a relatively small fraction of all adults. There are roughly 6.5 million adult New Yorkers that pollsters might reach with their initial calls, of whom only roughly 3 million are registered Democrats characterized as "active" by election officials. And only about 25 to 30 percent of those (or roughly 12--14 percent of all adults) are likely to vote in September. That means a lot of screening -- and cost -- for pollsters using an RDD methodology, and it explains why the two most recent polls sampled fewer than 500 "likely voters." [For turnout details see NY BoE, Census, HuffPost]
Campaign pollsters sample from voters lists - The Huffington Post has confirmed that the pollsters for the leading Democratic candidates -- Quinn, de Blasio, Thompson and Weiner -- have all drawn their samples from the lists of registered voters provided by voting officials. The advantage of this approach is their ability to narrow their samples to truly registered voters with a past history of voting in Democratic primaries. The prime disadvantage is the need to look up phone numbers for most of those sampled, so their surveys will miss voters with non-listed numbers and will have a much tougher time reaching voters in cell-phone only households.
How much does it matter? That question is hard to answer in New York City because, so far at least, almost none of the internal data has been publicly released. But in tomorrow's HuffPollster, we'll take a closer look at whether the internal polls may be yielding different results than the public polls and, if so, how.
HUFFPOLLSTER VIA EMAIL! - You can receive this daily update every weekday via email! Just enter your email address in the box on the upper right corner of this page, and click "sign up." That's all there is to it (and you can unsubscribe anytime).
TUESDAY'S 'OUTLIERS' - Links to more news at the intersection of polling, politics and political data:
--80 percent of 5 U.S. adults struggle with joblessness, near-poverty or reliance on welfare for at least parts of their lives. [AP]
--54 percent of Americans are unsure about whether the budget sequester is a good or bad thing. [Gallup]
-Americans' confidence in the economy has dropped to its lowest level since late April. [Gallup]
-The RNC hires Chuck DeFeo as its chief digital officer. [Politico]
-Jonathan Stray presents an equation-free talk on how not to get fooled when you go looking for stories in your data. [Source]
-Quartz releases an open source chart building tool. [Nieman Lab]
-Only .00003% of things that happen actually matter, per an exhaustive fake Pew Research study. [The Onion]