McClatchy/ Marist finds positive reactions to a potential budget deal. USA Today and Pew Research find little knowledge of the Iran nuclear deal. And a National Journal article says "political pollsters still don't know what they're doing." This is HuffPollster for Monday, December 9, 2013.
AMERICANS SUPPORT POSSIBLE BUDGET DEAL - David Lightman: "A Congressional budget deal expected this week might not be a grand bargain to solve the country’s long term fiscal woes, but it is largely what Americans want, according to a new McClatchy-Marist Poll….By 52--41, voters support replacing some cuts with increases to taxes and fees over going ahead with the cuts as scheduled. Democrats are most in favor of replacing the cuts, 64--29. Independents favor replacing them by 52--40. Republicans, however, would prefer the schedule cuts to raising fees and taxes, by 55--38. The across-the-board cuts were enacted in 2011 as a doomsday threat to force Congress and the White House to come up with a more deliberate alternative to curb budget deficits. The Congress and White House could not agree on other spending cuts or tax increases, and the sequester started taking effect March 1, with more scheduled every year for 10 years….Just 15 percent of voters thought the sequester had a positive effect on the economy, while 37 percent saw its impact as negative and 42 percent saw no effect." [McClatchy, Marist toplines]
'POLLSTERS STILL DON'T KNOW WHAT THEY'RE DOING' - Under that headline, National Journal's Scott Bland uses recent Quinnipiac surveys of Colorado to identify what he describes as a "recurring and increasingly disruptive problem," namely that pollsters are struggling "to reach the right mix of people." Bland: "This could be the case in Colorado, where the surveys have a very different makeup than what exit polls suggest is the typical Colorado electorate. In Quinnipiac’s Colorado surveys, whites without a college degree outnumber college-educated whites as a share of the electorate—43 percent to 35 percent in the most recent poll. But the exit polls that measured who actually voted in recent state elections have consistently found more degree-holders than not among white voters. Those exits, which have varied, put the gap at 40 percent with degrees to 37 percent without in 2012 and as high as 56 percent with and 25 percent without in 2010. What this means is that the polls in Colorado appear to be putting too much weight on the views of Republican-leaning voters, and thus exaggerating Democrats’ struggles in the state." [National Journal]
An old debate - Although uncertainty about the composition of the likely electorate is certainly a source of much confusion and perhaps even "disruption," the underlying problem isn't new. It's about the challenge of predicting the demographics of likely voters as much as a year before an election. Quinnipiac opts against identifying "likely voters" this early in the election cycle, instead choosing to report the results for all of the adults they interview that say they are registered to vote. They do weight their larger samples of adults by education (and also by age, race, geographic region) to match the Census estimates of the demographics in each state. "I don’t come at it from the vantage point of, do we have too many noncollege-educated whites," Quinnipiac's Doug Schwartz explained to National Journal. "I come from the vantage point that we’re just weighting to what the census is telling us should be there."
Electorates differ...a lot - Bland's analysis implies that knowing the "correct" demographic composition of the likely electorate is simply a matter of looking at exit polls conducted in previous elections. "The number of college-degree holders is, if not a fixed point, at least a demographic measurement that should be slower to evolve," He writes. "Colorado’s voters probably won’t look much different on that score in 2014 than they did collectively in 2012, 2010, or even 2008." Except that according to his own report, the percentage of voters classified as white and college educated looked quite a bit different from year to year. It shifted by 16 percentage points (from 56 to 40 percent) between 2010 and 2012. What share will they be in November 2014? The answer isn't obvious.
And they weren't really "exit" polls in Colorado - As Bland notes, "more than three quarters of the voters in 2012 cast their ballots by mail in 2012," but misses the reason why this issue presents a challenge for his comparison to "the after-action political gold mines of exit polls." Since 2008, because of the high early voting percentage, the National Election Pool consortium has conducted its exit polls in Colorado entirely by phone, calling over the week just before each election day. As such, except for timing, the exit polls have no special methodological advantage over Quinnipiac's surveys. Both were conducted by telephone using "random digit" samples of landline and mobile phone numbers. As AP polling director Jennifer Agiesta noted via Twitter, the "early voter" exit polls face the same kinds of challenges in identifying true voters as the Quinnipiac poll. The tendency of respondents to overstate voting is a problem for both, and a growing issue for exit polling as more and more voters cast their ballots early. [@JennAgiesta]
There may be a better way, but it's not about weighting to old exit polls - Bland only hints at what may be the larger issue: "One way that some pollsters try to combat these issues is by contacting their respondents using lists of known voters, instead of dialing phone numbers randomly." That is true, but up until very recently, voter list samples and the demographic estimates gleaned from them were used mostly for the internal polls conducted by campaign pollsters (see the HuffPost links below). But even Edison Research's Joe Lenski, who has conducted the exit polls in Colorado and elsewhere since 2004, warns against weighting poll demographics to match old exit polls: "You know that from election to election they are not going to be the same, so to assume the past demographics from an exit poll are a prediction of future demographics is just not borne out by the data...When we're looking at pre-election data, we'll compare to past exit polls just to see who's up, who's down, but that's more to inform us, who looks like they're more likely to turn out this year compared to past elections, not to trash the validity of the poll itself." [See also HuffPollster on voter list sampling here, here and here]
NEW POLL FINDS LIMITED SUPPORT FOR IRAN DEAL - Susan Page: "The White House and Iran face an uphill selling job to convince Americans to embrace the interim nuclear pact negotiated with Tehran last month, a USA TODAY/Pew Research Center Poll finds. In the survey, taken Tuesday through Sunday, 32% approve of the agreement and 43% disapprove. One in four either refuse to answer or say they didn't know enough to have an opinion. By more than 2--1, 62%--29%, those who have heard something about the accord say Iranian leaders aren't serious about addressing international concerns about their country's nuclear program." [USA Today, Pew Research]
With little knowledge about program, wording matters - The poll also found that just 24 percent of Americans had heard "a lot" about the deal, and 28 percent had heard "nothing at all." The USA Today/Pew Research poll gave few context clues to respondents, asking simply, "From what you know, do you approve or disapprove of the agreement between the United States and Iran on Iran’s nuclear program?" In contrast, a HuffPost/YouGov poll described the deal as one that "sets limits on the country’s nuclear program in exchange for easing of international sanctions and unfreezing of Iranian assets," and found Americans evenly split. An Ipsos/Reuters poll which said the deal would "freeze [Iran's] nuclear program in exchange for lifting some sanctions on the country" found 2-to--1 support. [HuffPost, Ipsos]
STAN GREENBERG RECALLS POLLING FOR MANDELA - Sam Stein: "Looking back nearly 20 years later, Mandela’s pollster from that campaign said the civil rights icon faced real political challenges. What made Mandela successful was that he didn't shy away from them. 'He was so self-conscious that the role he was about to play was different than the role he had played and also for the ANC, that they had to make a transition from being a military liberation organization and a social movement into a political party,' said Stan Greenberg in an interview with The Huffington Post….The campaign faced other hurdles as well. Though there was little quality polling in South Africa, the first numbers they were able to pull weren't overwhelmingly encouraging. As Greenberg wrote in his book Dispatches From the War Room, 'there was no white support for the party and the Coloured and Asian voters had defected to the National Party, despite it being the party of apartheid. Among Africans, who comprised 76 percent of the potential vote, the ANC had stronger support, but it still ranged unimpressively between 60 and 70 percent.'...….'What we discovered was the voters were pretty dispirited because of the long negotiations but also because the election wasn't offering them anything, other than power,' said Greenberg. 'He knew he couldn't win many whites. But for him, winning Coloured and Asian voters was critical so it wasn't just black nationalism. ... He wanted to be inclusive. And it was in that process that we came to better realize -- it became apparent that it was about using power and not just seizing power.'" [HuffPost]
-More on Greenberg's work for Mandela. [Bloomberg]
QUOTE OF THE DAY - Logan Dobson (R): "Saying poll aggregates are better than individual polls is like saying cookies are better than eating spoonfuls of flour & raw eggs." [@LoganDobson]
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).
MONDAY'S 'OUTLIERS' - Links to more news at the intersection of polling, politics and political data:
--62 percent of Americans say they prefer saving over spending. [Gallup]
-AP profiles the "rising rich" in an era when "poverty is at a record high." [AP]
-The AP's interactive chart showing exit poll presidential vote-by-demographics from 1972 to 2012 includes "inflation adjusted income." [AP via @jennagiesta]
-An analysis of Census data proves that the rent is too damn high. [HuffPost]
-YouGov surveys political insiders on the likelihood of a 2014 Senate or House flip. [YouGov]
-SurveyMonkey hires Jon Cohen as VP of survey research. [PR Newswire]
-David Leip explains how his U.S. Election Atlas collects county level vote results. [OpenElections]
-A study on twins suggests political beliefs may be "hard wired." [Pew Research's Fact Tank]
-Data science produces a "foolproof strategy" to find Waldo. [Slate via @AlexLundry]
-TresQuintos brings poll aggregation to Chile. [WaPost's Monkey Cage]
CORRECTION: The National Election Pool (NEP) exit poll consortium conducts telephone surveys of early voters for seven days leading up to the election, not just the weekend before.