“Trump has received about the most disproportionate media coverage ever for a primary candidate. The risk to Trump and candidates like him is that polling built on a foundation of media coverage can be subject to a correction when the news environment changes.” – Data gura Nate Silver on the polls and Trump.
Never in the course of American politics have so many paid so much attention to so many polls with so little relevance to what is really going on. Polls drive media coverage. Polls determine who gets to debate in prime time. Polls have become the oxygen of American politics. If you are up in the polls you are “surging.” Drop a few points – calling Ben Carson! – and you are “slumping.”
We are obsessed with polls, or at least political editors, reporters, campaign operatives and politicians are obsessed with polls. OK, let’s admit it, we are all obsessed with polls. I have been drafting survey questions and trying to analyze results for most of my adult life. I love the “cross tabs” and the idea of insight into the population, but we need to admit the business of polling is an art and not a science. Surveying a nation as big and diverse as ours often means channeling Monet and creating an impression rather than proclaiming a survey as scientific fact. Additionally, the rapid attention to all of the polling holds the real potential to skew the democratic process itself.
Time for a Deep Breath…
It’s human nature to want to know, as Donald J. Trump might say, “just what the hell is going on.” But political polls have become a little like really good Belgian chocolate. A little taste of really quality chocolate is satisfying and may even be good for you, but indiscriminately gobble too much and you’ll get a sugar high and put on a few pounds.
As The New Yorker’s Jill Lepore noted in a widely discussed piece last month: “From the late nineteen-nineties to 2012, twelve hundred polling organizations conducted nearly thirty-seven thousand polls by making more than three billion phone calls. Most Americans refused to speak to them. This skewed results. Mitt Romney’s pollsters believed, even on the morning of the election, that Romney would win. A 2013 study—a poll—found that three out of four Americans suspect polls of bias. Presumably, there was far greater distrust among the people who refused to take the survey.”
The Pew Center, which conducts widely respected surveys, estimates that the participation rate for its surveys is now just eight percent. Twenty years ago pollsters considered an 80 percent participation rate acceptable. With lower and lower participation rates, not to mention the challenge of reaching potential voters on a cell phone, some polling outfits have turned to “online” surveys, but the online methodology and sample quality have not kept pace with the frantic nature of polling.
Rutgers University professor Cliff Zukin, a past president of the American Association of Public Opinion Research, wrote a while back in the New York Times that “Internet use correlates inversely with age and voting habits, making this a more severe problem in predicting elections. While all but 3 percent of those ages 18 to 29 use the Internet, they made up just 13 percent of the 2014 electorate, according to the exit poll conducted by Edison Research. Some 40 percent of those 65 and older do not use the Internet, but they made up 22 percent of voters.”
With polls and chocolate it is a case of all things in moderation. So, while you nibble on a little pre-Christmas chocolate consider at least two principal things that are wrong with the overriding obsession with polling in our political process.
Two Big Problems with Too Many Polls…
First, too many polls these days are the political equivalent of a sleazy used car salesman who washes and polishes the old clunker in order to peddle it to some unsuspecting rube who doesn’t take time to look under the hood. Methodology matters, as does the professionalism and integrity of the polling organization. When assessing the latest polls its essential to “look under the hood” and understand how the survey was conducted and for whom it was conducted.
The data-crunching guru Nate Silver, he’s the guy who has nailed the prediction in the last two presidential elections, has a nifty analysis of the vast assortment of polls that make their way into the national news machine. Silver has ranked the polls according to their accuracy and methodology over time. It’s become standard practice for me – a certifiable “news junkie” – to check Silver’s ranking against the latest poll that assumes to convey heaven sent wisdom.
Second, it’s an old cliché in the polling business, but its still true: a political survey is a snap shot of a moment in time, specifically the time when the survey was conducted, as well as the slice of the electorate surveyed. At a given moment in time a well-constructed, well-researched survey utilizing a well-conceived sample can provide real insight into broad themes and valuable information about how voters might make specific choices among well-articulated alternatives. What polls are not particularly good at doing, at least at the current stage of the presidential campaign – remember not a single vote has been cast yet – is to serve as a predictor.
A skilled Republican pollster told me recently that one of the hardest things to master in the survey business is the “quality of the sample” – what portion of the electorate is actually going to make the effort to get to the polls and vote. “You need to be very careful,” this pollster said, “to pull a sample that accurately reflects age, party, geography, race and gender. That takes time and again costs more money.”
For example, when many Republicans, including his own campaign strategists, thought Mitt Romney would pull out a win over President Obama in 2012, they had misread the electorate that actually bothered to vote. A good deal of the polling miscalculated, for instance, the level of voting by Hispanic Americans. In his last election George W. Bush captured 40+ percent of the Hispanic vote. In 2012 Romney got only 25 percent. That difference in the makeup of the electorate helps explain Obama’s second term. At the same time, it would be a mistake to automatically assume that any Democrat in the future will always gain that level of the vote by a specific demographic group. Every election is different. Understanding the shape of the electorate is critical to good survey work.
The data-crunching site FiveThirtyEight has teamed with the NPR program On the Media to create a new consumer’s handbook for deconstructing the polls. You might want to copy the dozen points and tape it to your television.
But What of Trump…
I’ve become convinced that the Donald J. Trump phenomenon is, at least in some significant measure, a function of Trump’s own narcissistic pre-occupation with his standing in the polls. Trump regularly touts his “YUGE!” lead in the latest poll. He Tweets his standing from early in the morning until late at night. In the circular logic that drives news coverage of political campaigns Trump is, in many ways, a creation of the polls he loves to cite.
It is probably not a surprise then that Trump comments almost exclusively on polls that show him doing well. As the website FiveThirtyEightPolitics noted recently: “Trump also likes to tweet or retweet about the same poll a lot. He tweeted 29 poll results that he approved of more than once. He also tweeted one poll (a September CNN national survey) favorably, before turning on it more than once to highlight a better poll result for him from NBC and SurveyMonkey.”
As Trump tweets to his 5.3 million Twitter followers a poll that he likes, perhaps more than once, he is simply building the buzz about how successful his candidacy seems to be. It’s little different, from a marketing standpoint, than what Disney is doing by branding everything it can touch with a Star Wars logo. Marketing works to create impressions, drive coverage and, temporarily at least, move polls. It also help encourage people to buy Star Wars themed merchandize.
Trump may be an idiot about policy, but he understands the psychology of perception and he is constantly using polling data, some of it decidedly specious, to continually reinforce his leadership of the Republican field.
It is also no surprise that Trump dismisses as fatally flawed any poll that shows him slipping. When the very well regarded Des Moines Register poll recently showed Trump trailing Texas Senator Ted Cruz in Iowa, the place where the first voting takes place right after the first of the year, Trump dismissed the poll as the biased work of a newspaper that has been critical of his candidacy. Should I remind him that he touted his lead in the same poll earlier in the year?
The other factor that has driven Trump’s poll numbers (in addition to him constantly talking about his poll numbers) is the unprecedented media coverage his campaign has generated. He has been making wild claims about Syrian refugees, Mexican immigrants, banning Muslims, even his “bromance” with Vladimir Putin, but it has served to give him a hugely disproportionate share of media coverage.
Vermont Senator Bernie Sanders, based only on his two million individual campaign contributors (not to mention the polls), arguably has a larger following than The Bloviator-in-Chief, but by one measure Trump has received 23 precent more coverage than Sanders.
As Nate Silver says: “Trump probably realizes, the media’s obsession with polls can become a self-perpetuating cycle: Trump’s being in the media spotlight tends to help him in the polls, which in turn keeps him in the spotlight, which in turn helps in the polls, and so forth.”
It’s worth pointing out once again that no one has voted yet and there is considerable historical evidence that voters in early state contests like Iowa and New Hampshire decide very late in the process as to who they will support. Trump’s national polling lead may yet translate into real votes in the Iowa caucus – remember that in the best poll conducted at the state level talking to people who have actually participated in previous caucus voting he is behind – but there is also an argument to be made that his lead is to a large degree a function of prospective (or possible) voters seeing and hearing him constantly. Name ID matters, particularly when voters are still weighing choices.
As Jill Lepore noted in her New Yorker piece: “Donald Trump is a creature of the polls. He is his numbers.” But there could well be a bigger and longer-term problem for our democracy than one self-obsessed, poll-centric billionaire.
Reporting incessantly on polls and allowing this week’s polling results to determine the shape of political coverage, Lepore and others argue, is more than just a sign of the times, it is a signal of the increasing disintegration of American political culture.
“Turning the press into pollsters has made American political culture Trumpian,” Lepore writes, “frantic, volatile, shortsighted, sales-driven, and anti-democratic.”
No political junkie – I’m certainly guilty – would encourage a complete disregard for political polling, but there is a very strong case to be made for backing it off a full turn. Let’s have a little real voting and then we can see who is really winning.