Deja blew it: How pollsters misled the media – and the media misled voters

Voters of America! Pundits of journalism! My fellow reporters! If it seems we’ve been here before, it’s because we have. And we have. And we have again.

First, the current events. With Election day come and gone, it’s a good time to take a look at how the pollsters fared in calling one of the important elections in American history — where election deniers, fans of authoritarian rule and doomsday prophets threatened to hijack our democracy and point us square in the direction of the United States of Viktor Orbán.

I’ll concentrate on the 800-pound polling gorilla, FiveThirtyEight. Here’s how they called the balance of power in Congress  based on 40,000 simulations and as of Election Day itself:

Oh well. Rune sticks and ouija boards might’ve worked better. No wonder FiveThirtyEight’s fox mascot appears to be badly nearsighted, what with those square specs. Is that a jar of Powerball ping-pong balls they’re holding?

Barking up the wrong poll

Now let’s turn the clock two years to another barnburner of an election that held critical implications for not only our national leadership but also the future of democracy itself. As one analyst concluded, there was “a failure across the political universe” as the polling intelligentcia failed to voting tallies in the house, Senate and even the presidential race, where Joe Biden’s lead was not nearly so assailable as they asserted.  

Thus I wrote a fairly scathing op-ed on how FiveThirtyEight and its (occasionally foul-mouthed) czar Nate Silver royally dropped the political football in forecasting how U.S. elections would turn out. As I wrote then:

Following Election Day 2020, Silver got defensive. Fox reported that on his political podcast, he fired back at his critics: “If they’re coming after FiveThirtyEight, then the answer is, f–k you, we did a good job.”

Hmm. Sounds fairly Trumpian to me.

In fact, here’s The Donald himself talking about the slate of Republican candidates in last months midterms, as told to NewsNation’s Markie Martin:

“Well, I think if they win, I should get all the credit, and if they lose, I should not be blamed at all.”

And yet, another constituency deserves as much blame for the 2022 clusterfudge forecasts as the pollsters. And that’s the media itself.

When reporters rely on polls full of holes

My question is this: Given the remarkable string of polling fails, why do we in the media give even a second’s attention to pollsters anymore? Reporters are rigorously trained (or at least were once upon a time) to know that context is everything. If the Bozo Institute tells us that 7 of 10 doctors recommend aspirin to, say, cure cancer, then the obvious next question should be: How many times did those bozos miss the target entirely in their last ten proclamations? Really, how reliable are they?

What’s more, numbers can be fudged, deep faked and cooked in a way to make Enron accountants blush — more than ever, in fact.

But let’s assume the numbers are right, and that FiveThirtyEight used accurate data in its forecasts. Then other foundational questions come into play:

• How relevant were those data points?
• How static as opposed to fluid?
• How much did they account for the milieu (the overturning of Roe V. Wade) as opposed to the candidate?
• And: How much did they take into account the fact that everyone at some point “abandons the shopping cart” at the last minute and reverses their stated intention?

That final phenomenon, arguably the most relevant, is well documented in the bestseller The Human Element: Overcoming the Resistance That Awaits New Ideas. Authors Loran Nordgren and David Schonthal are professors at Northwestern University’s Kellogg School of Management.

Why some journos can’t stop acting like caveman

Here’s another point of human nature: Ever since the first Neanderthal wagered the first slab or raw mean on the first fist fight between knuckle draggers, humans have loved a good wager and the tension of an unknown outcome.

Yet we also hate uncertainty. And since turning to shamans and clairvoyants isn’t exactly en vogue this digital age, we embrace pollsters as soothsayers of algorithmic truth.

But if it’s numbers we’re going to worship, who’s analyzing the percentage of times polls fall well outside its so-called margin of error? Or even on a coin-toss basis, how often they can’t pick accurately between a loss and a win?

I’ll repeat what I said in 2020, because it’s just as important to grasp today:

Polls don’t just report information. … They influence outcomes. It creates a feedback loop. Maybe that’s not the intention. But it sure as hell is a consequence.

And as in 2020, reaction to the 2022 FiveThirtyEight fail was swift and spot-on, per its own comment section:

The stories coming out of 538 seem to be ignoring the elephant in the room. Their forcast and polling in general failed us miserably once again. Where do we go from here?

Just how wrong are you lot over there? I don’t even know who you guys talked to in your “polls,” but maybe find another line of work.

I should say. How come nobody polled these people?

A new line of work?

Look, I could do this polling job. Really, I could. A comrade in financial journalism once suggested — cynically, for sure — how I could elevate my standing in this profession as a prognosticator. He compared it to stock pundits such as Jim Cramer, who make a zillion calls, but only have to get one bullseye — which they promote the living daylights out of once it materializes.

Never mind that Cramer made perhaps the worst stock call in history in 2008 when he urged his followers to hold onto Bear Stearns stock just before the company executed a crash-and-burn for the ages. He had to apologize just recently when he trumpeted Meta stock prior to a massive drop in price.

But if you’re Cramer or Silver, why worry? Just wait for it to blow over and return in a few with your chest-puffing best. Or take a page from TV weathermen and their more degreed colleagues, the meteorologists. Not a single one ever got canned for being accurate on an occasional basis. Plumbers, surgeons, architects, accountants and yes, journalists: All historically have faced tremendous pressure to nail it near 100 percent of the time or else find another line of work.

The same can’t be said of political polling, which is doing a pretty good job of impersonating pseudoscience these days.

The big trouble is this:

When journalists rely on the pollsters, we get it wrong, too. In an age when the best of us are trying to kick the fake news label to the curb, it doesn’t help our credibility one bit.

This call, I would submit, needs no margin of error — not in a profession, a calling, where our margin of error must be slim to none.    

Lou Carlozo is the Editor and Publisher of Talking Biz News, and Qwoted’s Editor in Chief. All views expressed are a tongue stuck to a frozen poll. Email lou@qwoted.com or connect on LinkedIn.