NEWYou can now listen to Fox News articles!
For 15 years, American political analysis has been driven by data. Poll numbers, demographic breakdowns, a moneyball of power. It is a kind of algebraic alchemy that was supposed to reveal the actual voting preferences and trends of the people. The only problem? Time and again, including in the current midterm cycle the predictions have failed, often quite badly.
This style of political coverage came to the fore during the 2008 election when Nate Silver, who came out of baseball analytics, tried his hand at political prognostication with some impressive results. A new day had dawned, no longer would journalists rely on anecdotal conversations at diners and truck stops, now they had science!
Then the election of 2016 happened, and not only did the polls miss the dart board, they flew entirely out the window of the bar. Donald Trump and the American people confounded all of the data experts. But not everybody was so caught off guard.
Some journalists, most notably columnist Salena Zito, stuck to the old ways, the chats in byways, conversations at gas stations, and very early she said, wait a minute, Trump has a real chance here. She was roundly mocked, even accused of fabricating sources. That is, right up until she was proven right.
GRAND RAPIDS, MI: Michigan Republican gubernatorial candidate Tudor Dixon (L) celebrates her win on primary election night at the Amway Grand Plaza. (Bill Pugliano/Getty Images)
Once again, this year Zito traveled to Pennsylvania, watching Republican Mehmet Oz interact with voters in his run for Senate. At the time his campaign was flailing, especially in the polls, but she said, don’t sleep on this race. Once again, as the Johnny-come-lately polls show, Oz catching up on Democrat candidate John Fetterman, she was proven right.
The biggest flaw with data-driven political analysis is that voter responses to polls are limited to multiple choice answers. The only people in this equation saying anything substantive, which is to say framing the questions, are the pollsters themselves. The pollster can only see what they already are looking for, or expect, which leaves them blind to everything else.
After their humiliation at the hands of Trump in 2016, mainstream news had a brief moment of self reflection in which they promised to do a better job exploring and understanding “flyover country.” For the most, part this never happened, and when it did it often looked like a nature documentary in which some guy from Brooklyn goes on safari in Tennessee.
The fact is that most elites in journalism do not want to hear what the hoi polloi have to say and are shocked by what they hear when they do bother. This was on glorious display this week in an MSNBC focus group of Trump voters in Pittsburgh on the subject of the Capitol Riot. Reporter Elise Jordan, who ran the focus group seemed stunned when these regular Americans had very reasonable answers to all of her gotcha questions about the supposed coup. It is half painful and half hysterical to watch.
Politics is a story, it’s not an algorithm. Polls are useful tools, they can help political parties allocate resources just as market research helps product development. But the next big thing, be it a widget or a congressional candidate will not show up in that simple snapshot of the present.
To determine what is going to happen in the future requires something much more complicated than data, it requires imagination and actually listening to Americans. Listening to full sentences of their own creation, that is, not just yes and no answers.
Data-driven analysis can also stunt our political discourse by creating an artificial and often false Overton window of possibilities. For months nobody talked about Republican Lee Zeldin’s gubernatorial race in New York, or Tudor Dixon’s in Michigan, because the data showed us they obviously had no chance. Now both races are dead heats.
Voters deserved months of serious political coverage about these and myriad other races, not just the two weeks before the election. Trust in almighty data does a disservice to the voters who will tell you more than a survey can if you just bother to ask.
Facts and figures don’t care about feelings, and that’s the problem. In politics the narrative matters, the tone of voice, the look in the eye, the off-hand surprising remark that opens a door to a new way of seeing a race.
The columnist and reporter have their ear to the ground, the pollster has their finger on the pulse, the ear hears more and all across America, the polls are catching up to it.