Trump exploits the fake news of opinion pollsters

The far right and serious journalism's dirty sercret

​

Photo by Pau Casals on Unsplash

Imagine listening to commentary on a horse race: the Grand National, Kentucky Derby, or Prix de l'Arc de Triomphe. The commentator assures you that one horse is ahead. But then, at the moment of victory, they hastily correct themselves. Not only was that horse never ahead, they say, but another horse entirely won. 

You would start to wonder about the point of listening to your sports channel. If it kept happening, you might think that routine false predictions were wide open for exploitation by crooks looking to rig the betting.

So it has proved with the poll-fuelled commentary of  “horse-race journalism”. It is a truth that is nowhere near as universally acknowledged in the media as it needs to be that the findings of opinion polls are often flat-out wrong, and are hardly ever exactly right.

Now Donald Trump, a genuine crook and proto-fascist, is planning to exploit the failings of opinion polling, that the media have tolerated for far too long.

Leftish critics have long condemned mainstream journalists. Instead of attempting to provide supposedly objective commentary on the state of play, they say, journalists should be warning of the stakes for democracy if men like Donald Trump take power.

In the words of Jay Rosen, Professor of Journalism at New York University:  “We need a press that can figure out how to defend democracy as it is being attacked and eroded by a potent movement.”

I have far more sympathy for traditional journalists than Rosen and those who agree with him.

But in their treatment of the media’s use of opinion polls the critics are nowhere near critical enough. Indeed, they miss the real criticism.

They want reporters to abandon objectivity and go on the attack. In their fervour they fail to see that opinion polls give a fake aura of objectivity to statistics that might as well have been picked at random. Truly objective journalists would have nothing to do with them. Instead reproters give an aura of legitimacy to a pseudo science that is wide open to abuse.

When the “facts” of polling are routinely revealed to be false - nothing happens. Editors don’t fire pollsters and vow never to hire a replacement. There are no inquests into how we got the story so terribly wrong. Not just the media, but politics, business and academia continue to commission polls as before.

Most editors I know just assume they must publish polls – like the weather forecast or TV listings. The cynical among them know that modern people are no better than the plebs of ancient Rome, who needed a priestly caste to divine the future in a chicken’s entrails.  

The result is that we are now in an absurd society where people like Nate Silver can make a good living adjudicating why one poll may be less false than another, and the Economist runs entire features asking whether the US election polls could be wrong. Like so many others it never stops to wonder whether, if it thinks poll results could be wrong, the Economist is publishing them in the first place.

As with so many other corruptions, Donald Trump is ready to exploit the media’s failure to clean up “serious” journalism. If he needs to, Trump will use fake right-wing polls to dispute the results of the US election.

Frankly, Trump’s fakes will be no less obviously false than the supposedly genuine articles.

But I run ahead of myself.

Like a good, objective journalist I must first present the evidence as impartially as I can.

In the 2015 UK general election the pollsters failed to predict that the Conservatives would win. Indeed, they weren’t even close.

In the weeks leading up to the November 2016 US presidential election, polls predicted an easy victory for Democratic nominee Hillary Clinton. Trump won.

Before the 2017 British general election, polling underestimated the strength of Jeremy Corbyn’s Labour party.  “Once again, the polls, taken as a whole, were not a good guide to the election result,” admitted the BBC.

Moving on to the 2020 US presidential election, the pollsters predicted that Biden would win. And he did! So well done them!

But they overstated Biden’s lead over Trump so badly that sober commentators concluded that “polling is irrevocably broken,” that pollsters should be ignored, and – my favourite – that “the polling industry is a wreck, and should be blown up.”

To paraphrase the late J.K. Galbraith, “The only function of opinion polling is to make astrology look respectable.”

As if to prove the point, pollsters predicted a Republican landslide in the 2022 midterm elections in the US. Instead, the Democrats did far better than expected.

Last but not least, the 2024 British general election was a disaster for the polling companies – I know! Again!  They overestimated the Labour vote by 6 per cent.  As so often, polling companies scrambled for excuses, claiming there were late swings and surges in tactical voting. All the reasons offered had one point in common, they were not the industry’s fault.

I know no other area of serious journalism where editors routinely provide readers with false information, and then refuse to issue corrections and change their ways when their stories collapse.

And it is not just journalism: businesses and politicians use polling to decide what products and policies to sell.

Suppose politicians picked a damaging policy because the same pollsters who so badly misjudged the 2024 election assured them it was 6 percent more popular than the alternatives.

In other words, like so many other pseudo-sciences, polling can cause real harm.

Nowhere more so than in the US today. Everyone watching Trump with horror and disgust can see that he and his supporters are preparing the ground for a coup d’etat.

If Harris wins, Trump will claim he is a victim of fraud and do what he can to encourage his lackeys to overturn the legitimate result of a free election – just as he did in 2020.

As a part of the softening-up exercise, right-wing operatives are releasing polls showing Trump ahead. They will use them to claim that the election was stolen if the need arises.

Last week, the right-wing influencer Ian Miles Cheong shared a survey with his 1.1 million followers on X. The forecast from a new polling company suggested, without sharing its methodology, that Trump would take 74.3 percent of the national vote — a landslide unprecedented in American history.

“Trump is absolutely going to win,” Mr. Cheong wrote. “The data shows it.”

Meanwhile, an analysis by the New York Times showed that in the final stretch of the campaign in 2020, Republican-aligned pollsters released 15 polls of swing states. In the same period this year, they have released 37.  All but seven of which had Trump in the lead.

Doubtless respectable mainstream commentators will denounce the fake polls if Trump uses them to help make a power grab next week.

But if they have integrity, they should reflect that opinion polling was always the most consistently misleading feature of “serious” journalism. They ought to have binned it long before the far right realised it could exploit it to advance the career of a fascistic demagogue.

---

See also…..

---

And finally…

I love writing this newsletter but it is a great deal of work! Please consider becoming a paid subscriber if you can.

You will receive access to all articles, archives, podcasts and debates and you will also allow me to keep this project going as a journalistic concern without advertising or clickbait or any kind of proprietorial interference.

Annual subscriptions work out at £1.15 ($1.45) a week, which is cheaper than the cost of the worst cup of coffee you could buy in any cafe in the developed world!

Click on the pieces below to get a taste of what is on offer. (There’s a free trial for the fainthearted.)