Huff wrote his famous warning in 1954. Sixty-five years later, political polling somehow still dominates the “mainstream” national discourse. Even after what I call The Great Polling Uberfail of 2016, we still find ourselves mired in the statistical mud; “facts” are as malleable as mood, and clarity has ceased to exist, if it ever existed in the first place.
“The Princeton Election Consortium’s Sam Wang gave Hillary Clinton a whopping 99 percent probability of winning,” wrote Nicole Narea for The New Republic in the immediate aftermath of that polling debacle. “The New York Times, meanwhile, put her at 85 percent. Of the ten major polls averaged by RealClearPolitics, eight predicted she would win by a margin of at least three points. Not even the respective campaigns’ internal polls foresaw a Donald Trump victory, which only became clear at around 8:30 p.m. on November 8.”
Stay in the loop
Never miss the news and analysis you care about.
As another presidential election looms before us, we need to get a few things straight about political polls. I hate the damn things, loathe them, detest them, except when I agree with them … and therein lies the rub. Polls are the sniper rifle in the arsenal of confirmation bias, a quick way for the media to grab your attention by eliciting either glee or outrage with numbers, which of course never lie, until they do. Mr. Huff, your table is ready.
A pair of polls came down the turnpike over the last several days that seemed — to my eyes, anyway — so freighted with nonsense that they threatened to bend the light. The first, from CBS News, reported on the national reaction to Trump’s now-infamous racist Twitter tirade against Representatives Ocasio-Cortez, Omar, Pressley and Tlaib, collectively known as “the Squad.”
The CBS/YouGov poll was compiled from answers given by 2,099 U.S. citizens whose inclusion was “weighted according to gender, age, race, and education based on the American Community Survey, conducted by the U.S. Bureau of the Census, as well as 2016 presidential vote and registration status.” The margin of error was 1.5 percent.
According to the CBS poll, 59 percent of respondents disagree with Trump’s “go back to their countries” rhetoric, with 44 percent of that group strongly disagreeing. By the numbers, that means a full 40 percent — nearly half the country — agrees with Trump’s incendiary racist spew.
The second poll, from USA Today, measured the public’s opinion of “the Squad” themselves. As with the CBS poll, YouGov queried 2,099 respondents and adhered to the same weighting based on gender, age and the rest. According to their results, “Ocasio-Cortez had the highest favorable rating at 23 percent, followed by Omar at 19 percent, Tlaib at 18 percent and Pressley at 17 percent.”
If you are even mildly progressive, the numbers within these two polls might make you want to climb a tree and learn to speak Squirrel, because nothing good is happening down on the ground. If, however, we heed the words of Darrell Huff and remember “there may be a good deal less” than meets the eye here, we find some startling data beneath the “data.”
Let’s start with the second poll first, because frankly, it cracks me up. USA Today presented it under the incredible headline, “Many Americans Have Never Heard of the ‘The Squad’ and Many Don’t See Them Favorably, Poll Finds.” I … but … who … what … where … when … why … how can “many” who don’t know “the Squad” also be the “many” who don’t like them? Which many is how many, or is it who many, and how do they know what they don’t know?
The first poll, by CBS News, is demonstrably more ominous. How, did the pollsters arrive at 40 percent national approval for Trump’s racist comments? One answer may very well be that almost half the country is racist enough to answer “Yes” to a “Yay racism!” CBS poll question. I do not discount this as a very real, very grim possibility, and am keeping my English-to-Squirrel dictionary handy in the event the data proves out beyond question.
Before anyone goes looking for a snug live oak to populate, there are some other data points that very much seem to indicate the polling industrial complex has not yet caught up with the technological complexities of the 21st century.
GfK MRI, a consumer and media research organization, published the results of what it called “The Survey of the American Consumer” in 2017. Before anyone says, “Great, another poll, why am I here,” a review of GfK MRI’s methodology is worthwhile. The organization did not pluck their data from a couple thousand phone calls, but surveyed 25,000 people in person in their homes. No survey is perfect, but as far as they go, this one is about as good as it gets.
The GfK MRI survey found that 52 percent of voting-age adults live in homes without a landline telephone, choosing instead to rely entirely on cell phones. Only 29 percent of millennial voters live in homes with a land line, while 77 percent of seniors over the age of 65 live in homes with land lines. Another massive in-person survey of 19,956 households taken that same year by the Centers for Disease Control and Prevention shows that landline usage is dwindling across all sectors of the country.
And then there’s this, from a Politico report earlier this year:
The Pew Research Center reported Wednesday that the response rate for its phone polls last year fell to just 6 percent — meaning pollsters could only complete interviews with 6 percent of the households in their samples. It continues the long-term decline in response rates, which had leveled off earlier this decade.
What’s to blame for the recent slippage? [Pew’s survey research director Courtney] Kennedy says it’s harder to get people to complete polls over cell phones because they are getting more calls they don’t want, which makes them less likely to talk to pollsters.
As this information comes from yet another poll, it is as worthy of doubt as any other poll-driven conclusion. Yet the available data, as Gene Shalit might have said, is Tony Stark: The return rate on phone-only polls is cratering while the preponderance of cell-only households is rising. The two circumstances seem to be umbilically connected.
Most importantly, the age breakdown between cell and landline households cracks wide open along age group fault lines, which means the results are weighted significantly in favor of older respondents while almost completely excluding the opinions of younger voters.
I am not throwing stones at older voters. However, younger voters have been found to be less racist, misogynistic and homophobic on average, and so any poll lacking their inclusion will necessarily be stacked in favor of the racists within that body of landline-using older voters populating the 6 percent that actually answer phone polls nowadays.
Polling organizations like Pew and Gallup are taking steps to try and contract the technology-created crater within their data, but problems persist. “Some pollsters now sign up pools of respondents in advance and offer cash or gift cards as incentives, which is believed to skew the sample and the quality of answers,” writes Max Berley for Bloomberg News. “Other firms have turned to cheaper web questionnaires, which have the obvious problem of restricting the sample to people who are online.”
Concern over accuracy was apparent at the American Association for Public Opinion Research’s annual meeting in May. “Pollsters at the conference presented research on sending voters invitations in the mail to participate in a web survey and texting voters with links to online polls,” reports Steven Shepard for Politico. “Another study even had live interviewers texting back and forth with respondents — a 21st century twist on the telephone conversations long underpinning survey research. But none of the methods seem ready for prime-time deployment — or the widespread trust of pollsters, campaigns and media organizations — between now and 2020.”
None of this is conclusive, to be sure, but it is deeply compelling. At a minimum, it should give you pause whenever you see polls claiming that Joe Biden is leading the Democratic presidential field by 30 points, or that nobody really likes AOC, or that almost half the country is unabashedly racist.
Polls are a way to fill the 24-hour news cycle with chatter about polls. They provide a reason for people to dress up nicely and appear on TV to flood the air with verbal packing peanuts that serve only to fill the portions of a broadcast not taken up by commercials.
In our increasingly cotton candy-fied media landscape — all puffy sugar and few calories — polls are as necessary as electricity to keep the show on the road. Absolutely nothing about that makes them right. It only makes them there.
Before a poll leads you toward elation or despair, remember all this. As they say in Squirrel, “Eep iki ik eep ik-eep chitter eep,” which, roughly translated, means, “Get out of my goddam tree, it’s not time yet.”