HUFFPOLLSTER: Why Reaching Latinos Is A Challenge For Pollsters

Two pollsters weigh in on the difficulty of surveying Latinos. Recent polling misses at home and abroad give reason for caution. And Americans aren't buying Ben Carson's theory about the ancient pyramids. This is HuffPollster for Friday, November 13, 2015.
JOE KLAMAR via Getty Images

THE CHALLENGES OF POLLING LATINOS - According to a report from Pew Research, Latinos make up nearly 18 percent of the country’s population, meaning they are a critical group for pollsters despite the challenges in polling the population. Anna Brown: “Surveying Hispanics is complicated for many reasons – language barriers, sampling issues and cultural differences – that are the subject of a growing field of inquiry.… studies have shown that Hispanics are more likely to refuse to participate in surveys, or having agreed to take a survey, more likely to refuse to answer individual questions under some circumstances.… One other key factor to consider when planning a survey of U.S. Hispanics, 73% of whom report speaking either only Spanish or both Spanish and English at home, is the translation of the survey instrument. Conducting a survey in multiple languages is complicated, but necessary to capture a sample that is representative of the U.S. population.... Hispanics are the racial/ethnic group most likely to live in a cellphone-only household. And those who live in cellphone-only households have a different demographic profile from those who don’t.” [Pew]

A firm focusing on Latinos tried one new approach - Latino Decisions: “[We]interviewed 424 Latino registered voters who live in 14 presidential battleground states, using their new randomized email sampling method. Registered voters were randomly selected from the voter file and contacted by email, and asked to complete a short survey. The email invite was provided in English and Spanish, and respondents could choose to take the survey in either English or Spanish. The random selection approach is the exact same as in a telephone survey, with the only difference being the respondent completed the survey online. This has a significant advantage over most online panel samples which are “opt-in” or convenience samples and not randomly selected.” Latino Decisions is currently polling for the Clinton campaign, but conducted this survey for impreMedia, a separate, nonpartisan client.[Latino Decisions]

The results were unsurprising - The poll shows that most of the Republican candidates are not viewed particularly favorably by the group. The only Republicans with positive net favorable ratings -- the percentage favorable minus the percentage unfavorable -- are former Florida Gov. Jeb Bushand retired neurosurgeon Ben Carson, although 58 percent either had no opinion or had never heard of Carson. The least favored candidate of the ones they tested was businessman Donald Trump, with a -56 net favorable rating. [Latino Decisions]

More on the sampling methodology - Latino Decisions got this email-matched voter list sample from a company called L2. The company's president and CEO, Bruce Willsie, told HuffPollster that about 25 percent of voters can typically be matched to an email address based on the names and physical addresses in the voter file. Sylvia Manzano, a principal at Latino Decisions, said that about 20 percent of the records for Latinos in the 14 battleground states were matched to email addresses. “It’s lower than we had hoped for, but still a better baseline sample than an opt-in [web sample] since we know they are registered voters, and we contacted them,” she said.

How well did the sampling work? - Manzano: “The one thing we found is that this sample is a slightly more educated and higher income group. That’s consistent with what you see in online surveys. As far as the rest of the demography and the profile of the respondents the sample pairs up nicely with what we’d expect for Latinos in these states. We’re happy we were able to execute it. It’s not a secret that polling is getting more difficult; we have to find more comprehensive methods to reach voters and find them where they are."

MORE FALLOUT FROM RECENT POLLING ABROAD - Giovanni Russonello: “Pre-election polls in numerous countries this year have widely missed their marks, often by underestimating support for candidates on the ideological fringes. The polling failures in countries like Britain, Poland and Israel point to technical issues that could well foreshadow polling problems in the United States, many analysts believe. ‘The industry has a collective failure problem,’ said John Curtice, the president of the British Polling Council and a professor of politics at the University of Strathclyde in Glasgow.…General election polling in the United States typically doesn’t fall victim to factional conflicts [that affect polling in other countries], because in most cases only two major candidates and parties are represented on the ballot. But in the Republican nominating contest, where turnout is uncertain and a dozen candidates are vying for conservatives’ hearts, strong parallels abound.” [NYT]

...AND AT HOME - Jill Lepore: "Election pollsters sample only a minuscule portion of the electorate, not uncommonly something on the order of a couple of thousand people...Meanwhile, polls are wielding greater influence over American elections than ever….Even if more people could be persuaded to answer the phone, polling would still be teetering on the edge of disaster….Pollsters rose to prominence by claiming that measuring public opinion is good for democracy. But what if it’s bad?" [New Yorker]

Another view encourages more caution among poll consumers - Michael Barone: "I think polling will endure in some form, but consumers need to be more cautious in assessing polls. … Poll numbers, despite their seeming precision, are not hard data. They are clues to the mysteries that lie in human hearts....Technological change may be making polls less scientifically reliable, but reading polls has never been entirely a science; it has also been an art and seems to be getting more so. [WSJ]

GENERAL ELECTION POLLS RELEASED NOW AREN’T USEFUL FOR PREDICTING OUTCOME - Harry Enten: “If you look at polls that tested the eventual Democratic and Republican nominees in the last two months of the year before the election, the average absolute error of the polling average is 10.6 percentage points. That’s more than five times Ben Carson’s current lead over Hillary Clinton in the Huffington Post/ aggregate. … Most recently, Barack Obama and John McCain were tied a year before the 2008 election. At that point in the campaign, more people cared about foreign policy than domestic issues. That changed dramatically as the global financial system collapsed, and Obama went on to win by over 7 percentage points.” [538]

CONTEXT MATTERS IN POLITICAL STATISTICS - Felicia LeClere: “I ask you to consider the case of the humble denominator….The trouble begins when you choose the denominator in a way that privileges one group over the other. You can do this in two ways. First, by choosing a denominator that varies widely among groups, you can create the impression of a difference when there may be none. Second, you can choose a denominator that is not the right dance partner for the numerator and emphasize a difference that may not be real. It's pretty subtle, and that's probably why it is such a useful tool in shaping opinion….Political comparisons that rest on statistics that are not correctly constructed can contribute to a feeling that the whole enterprise of numbers and politics is untrustworthy. " [HuffPost]

AMERICANS DON’T BUY CARSON’S PYRAMID THEORY - HuffPollster: “Much of the public does share Carson's faith in some mainstream Biblical precepts. Forty-two percent of those polled said they believe the world was created in six days, with 37 percent saying it was not and the rest unsure. A 52 percent majority believes the end of the world will happen as predicted in the Book of Revelation, with 13 percent saying that this will occur in their lifetime. But that doesn't mean people are equally on board with the pyramid theory: Just 4 percent of respondents said they think the pyramids were grain silos, while 72 percent believe they were used as tombs.” [HuffPost]


-The more Americans hear of Ben Carson, Marco Rubio and Ted Cruz, the more they like them. [McClatchy]

-Hillary Clinton's favorable rating has jumped 14 percentage points among Democrats since her last presidential primary debate performance. [Gallup]

-Ben Carson is the most liked presidential candidate among Republicans, with a 71 percent favorable rating. [WashPost]

-Donald Trump and Ted Cruz are tied in Texas.[Texas Tribune]

-Democrats see Hillary Clinton as more effective than Bernie Sanders. [NYT]

-Democrats think Donald Trump will be the hardest candidate to beat in the election. [CBS]

-Senate races look close for now in four battleground states. [Democracy Corps (D)]

-Americans say they'd be least likely to vote for someone who has no government experience or believes in socialism. [Marist]

-Americans welcome Paul Ryan into his new position as majority leader by giving Congress the lowest rating of the year. [HuffPost, Gallup]

-Republican voters want an outsider candidate, while most Democrats would prefer an establishment figure. [HuffPost]

-Women, city dwellers, and those in the lowest-income households are most afraid of walking home alone at night in their neighborhood. [Gallup]

-Nearly two in three Americans say they're more afraid of gun violence than a terror attack. [HuffPost]

Charlie Cook moves the Louisiana governor’s race to “lean Democrat.” [Cook Political]

-A majority of Americans want the FDA to review drug ads before the air. [Kaiser]

HUFFPOLLSTER VIA EMAIL! - You can receive this weekly update every Friday 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).

THIS WEEK'S 'OUTLIERS' - Links to the best of news at the intersection of polling, politics and political data:

-HuffPollster argues that Chris Christie and Mike Huckabee should not have been excluded from the main Republican presidential debate this week. [HuffPost]

-Ilia Blinderman explains what he learned from analyzing missed connections data off of Craig's list. [Vox]

-Andy Guess, Jonathan Nagler and Joshua Tucker analyze what tweets can say about a candidate's debate performance.. [WashPost]

-Are Americans lying to pollsters on how they really feel about gay marriage? Researchers say no. [The Atlantic]

-Economic and social change provide contrasting strengths for Democrats and Republicans in elections. [National Journal]

-Stu Rothenberg doubts the two term jinx will be an insurmountable barrier for the Democrats. [Roll Call]

-Philip Bump finds Facebook data is a bad predictor of political outcomes. [WashPost]

-A Cruz-connected data miner is promising a new approach to identifying voters. [Bloomberg]

-A new study finds that implicit bias means women have to be more qualified than male opponents to win an election. [HuffPost]

-Eileen Patten explains why figuring out who's multiracial can be tricky. [Pew]

-Pollsters discuss the ways technology is changing polls and how we can adjust to be better poll consumers. [BizJournals]

-Politico panelists discuss the state of the American electorate. [Politico]

Go To Homepage

Popular in the Community