Sometimes, the more the media think, the less they know

Elections used to bring out the best in journalism. Now, too often, it feels like they’re bringing out the worst.

The relentless drive to guess the future — replacing fourth estate accountability and “what just happened” reporting — seems to be the problem. Worse, journalists become prisoners of their own predictions, shaping their “news” around what they expect to happen come election night, to ensure their stories are properly positioned as “the first draft of history” in what they expect to be the post-election narrative.

Right now, it’s distorting the coverage of this week’s US midterms, opening the media up to be gamed, particularly by the Republican right, happy to use their own media voices to part-nudge, part-bully the news media to adopt their favoured interpretation of “what’s news”.

It’s the end point of horse-race journalism: the tipping panel.

As repeated failures have broken confidence in polling, US media have reverted to future guessing based on a mix of pattern recognition and the vibes out of the post-industrial Midwest, all hammered into a repetitive, hardened orthodoxy.

Independent polling reflects a more nuanced picture, but too many US polls have got too much of the past few elections materially wrong. American pollsters face a challenge that Australian pollsters don’t: factoring in predictions of turnout, of determining just what the voting electorate is going to look like. Election by election they’re finding that past patterns of turnout are no guarantee of future performance.

The US traditional media have increasingly reverted to what political analysts call “thermostatic” analysis — the idea that election results rhyme in a predictable pattern. In the US, it’s the idea that the president’s party “always” loses the midterms.

We see a touch of “thermostatic” pattern recognition here in Australia, too, with a widespread assumption across the political class that the governing party federally goes backwards in the states. The Victorian election result will be an early test of that theory for the Albanese government.

The trouble with the theory (as with much of political “science”) is the data set is small and, by definition, there’s no control sample to test for other factors. The result of that media certainty? If you “know” just one thing about this week’s US election, it’s that the Republicans are set for a landslide. And if you do “know” that, it’s because the US traditional media have been predicting it as a near certainty pretty much since Joe Biden was elected two years ago.

But the media are more than just the audience. They’re players. How they report the campaign shapes the narrative which, in turn, shapes the outcome.

In the US right now, we’re seeing how that plays out with reporting of crime. The media expect Republicans to win. The Republicans are talking about rising crime. People must be concerned about crime. Better write about it. In fact, as The Washington Post’s Philip Bump wrote last month, this has it back to front: heightened concerns about crime are the result — not the cause — of it becoming a go-to Republican talking point.

The key vector for smuggling these talking points to the centre of political discourse? Fox News.

About two weeks after Fox started beating the crime drum in September, the competing cable networks, CNN and MSNBC, took up the ball. Bump wrote:

In late September, though, mentions on Fox News began to soar. In the middle of October, mentions began to rise on CNN and MSNBC, too, in part as a reflection of the increased discussion of crime on the campaign trail.

Bizarrely, it all ended up empowering the Republicans to hijack the terrorist attack on the Pelosi family into a misinformation loop to blame the victim.

Australia has its own experiences of how the media’s certainty affects political coverage. In the 2019 election, the assuredness of a Labor win produced a laser-like media focus on the details of Labor policy while waiving past the government’s record. It made some sense, based on the historic accuracy of Australian polling: likely governments should get greater scrutiny.

The early days of the Victorian election campaign suggest the Melbourne media’s assessment of electoral odds (coupled with News Corp’s odd Dan Andrews’ obsession) is making the government the centre of the campaign.

Whatever the result in the US this week (and in Victoria at the end of the month), there’s still one thing we won’t know: how much of the outcome will be due to the reporting of the facts and how much due to the choice of what to report?