Memo to MSM Bloviators: Get a Clue About Polling
The Iowa caucuses are about five months away and the 2016 presidential election is almost a year and two months away and yet not a day goes by without another poll being released – NBC/Marist, Survey USA, Florida Times Union, Fox/Morris News, Monmouth, PPP, Loras College, Quinnipiac, ABC, CBS, CNN, NYT, WashPo. We’re expecting one soon from Wossamotta U.
And every day, TV motor-mouths, so-called analysts and some huckster candidates breathlessly explain the importance of these findings, virtually none of them understanding (or caring) who was polled; how they were surveyed; whether the margin of error is two, five or 10 percent or if there even is a margin of error (which internet polls don’t have), and most importantly, whether the pollsters have an agenda or even a clue.
Always Trust Calbuzz Given their desperate 24-hour scramble for fresh “content,” we don’t expect the mainstream media or the online alternate outlets to adopt serious standards about polls or intellectual honesty in reporting them. But as loyal readers know, you can rely on Calbuzz to shed some light from time to time.
Consider the so-called poll that Sir Donald of Trump continues to cite, “proving” that he’s doing just fine among Hispanics.
Even before he kicked Univision anchor Jorge Ramos, the Walter Cronkite of Spanish-language TV, out of his press conference – “Go back to Univision” – Latinos had no use for Trump.
“The Hispanics love me,” Trump continues to insist. But this is a grandiose function of his narcissistic personality disorder, not an actual grasp of the facts.
A recent Washington Post-ABC News poll (a genuine, professionally-run survey) found that 81 percent of Hispanic voters had an unfavorable opinion of Trump, compared to just 13 percent who viewed him favorably. (Attention MSM: “Hispanics” includes more than the handful of Hispanic Republicans Trump imagines constitute the Latino universe.)
“Pay no attention to those losers,” Trump would argue. Besides the thousands of Hispanics who have worked for him over the years – who love him so dearly – Exhibit A in Trump’s contention that he’s in great shape with Latinos is a poll of Nevada voters by Gravis Marketing, done for the One America News Network, which is dedicated in its own words to “providing a platform for independent and conservative ideas.”
We Probe So You Don’t Have To: Seeking to get more detail about this poll, at least half of us had a pleasant conversation with Charles Herring, the boss at OANN, about the Nevada poll. We were encouraged to submit our questions in writing so that Chuck could get with his pollsters at Gravis and get back to us.
We sent a number of questions to Herring via the email address he provided and never heard back. Which is not surprising because, based on the scant information OANN has released about the poll, Trump’s claim to have support of Hispanics in Nevada is utter nonsense.
According to OANN, the automated telephone poll (alarm bells here!) included 1,276 respondents, of whom 1,039 self-identified as primary participants (which would be an astonishing, i.e. unbelievable, 81% participation rate) of whom 623 said they were Republicans and 416 said they were Democrats (an unrealistic 60-40% split in favor of Republicans, but who’s quibbling?).
Elsewhere, OANN said 16% of the respondents were Hispanic. But whether that’s 16% of the total sample, of the primary participants or the Republican primary participants is not made clear. Our Department of Abacus Operators and Public Opinion Pulse Feelers would be shocked if 16% of the Republican primary participants were Hispanic, but who knows?
Even if we give Gravis and OANN the benefit of our towering doubt, and 16% of the 623 Republican participants were Hispanics, that would be about 100. And if Trump had 31% among Hispanics – as the survey alleges – that would be 31 voters. (But the margin of error on that 31% would be about 9 percentage points, plus or minus.)
In other words, given the information at hand, the best we can say is that Trump is, as usual, blowing thick clouds of smelly gas when he says he’s leading among Hispanics in Nevada. He MIGHT be leading among the tiny number of Hispanic Republicans who plan to participate in the primary. But we don’t even know that for sure.
Again, Calbuzz tried to get OANN to more fully explain their results. Here’s a few of the questions for which we got no response:
Among the 623 GOP primary participants, what was the racial breakdown? % white, % Latino, % black, % other? Also, the same for the 416 Democratic primary participants?
What was the breakdown by race of the GOP primary contest?
What was the breakdown by race in the general election match-up between Clinton and Trump among all voters?
Bottom Line: The next time you hear Trump say he’s popular among Hispanics, the proper response is a Bronx cheer. And when you hear TV blow-drys tell you that Joe Biden runs better than Hillary Clinton against Trump or Jeb Bush or Marco Rubio, ask yourself: does the person telling me this have any idea what he or she is talking about?
Polling has become incredibly accurate. Pollsters apply terabytes of data to predicting election outcomes but that data is useful toward the end when they can measure “likely voter”. Each pollster has his/her proprietary method of defining “likely voter”.
It seems that these Nevada pollsters applied kilobytes of data to determine Trump’s popularity among Latinos.
Polls are accurate, just not this early because pollsters have not even begun to try to predict likely voter. Additionally, when I look at what has happened in the Republican Party this summer, I think that everything I know about politics is wrong. Maybe pollsters are thinking the same thing.
1. isn’t the “one america network” for people who think Fox News is too leftist or something?
2. (and this is meant as a constructive critique, not trying to be rude here) but have you considered updating the “share this” options on your site? NO ONE uses technorati, or half the ones up there, and reddit is likely to die off soon, and del.icio.us was bought out by yahoo ages ago and killed off the way yahoo does….
In the 2011 summer, the polls had a new Republican President every 10 days. Hard to remember their names now.