Paul Romer has talked about two types of discourse, one political and one scientific. He uses that distinction to critique aspects of current practice among economists. I want to do the same for journalism.
Political discourse involves taking sides, and promoting things that your side favours. It is like a school debate: you consider only evidence that favours the point of view you want to promote. Scientific discourse involves considering each piece of evidence on its merits. You do not aim to promote, but assess and come to a conclusion based on the evidence. That does not prevent the scientist arguing a case, but their argument is based on considering all the relevant evidence. There are no sides that are always right or invariably wrong.
Of course, any scientist makes choices about what evidence is relevant, and this will be influenced by existing theories. Ideally the theory you prefer can be changed by new evidence, but scientists being only humans can sometimes be reluctant to accept evidence that contradicts long held theories. But there are always younger scientists looking for new ideas to make their name. The scientific method works in time, which is why we are where we are today.
My argument is that journalists should be like amateur scientists. Amateur because part of their work will involve seeking out expertise rather than starting from scratch, and they do not have the time or resources to investigate each story as a scientist might. A term frequently used is ‘investigative journalist’, but that normally means someone who has weeks to work on one story. Instead I’m talking about journalists who only have a day. The key point is that they should not search for evidence that fits the story they wanted to write before doing any research, but allow the evidence to shape the story.
For example, suppose the story is about EU immigrants and benefits. What a journalist should note is that unemployment among EU immigrants is lower than natives. What a journalist who wants to write a story that makes immigrants look bad might do is say that the number of EU immigrants without a job make up a city the size of Bristol. This combines selection of evidence (where is the equivalent figure for natives is not reported) with simple deception: most people conflate ‘without a job’ with ‘unemployed’, rather than being people happy looking after children, for example.
If this all strikes you as obvious, at least to journalists working in broadsheet newspapers, the example above is taken from the Telegraph, and the post in which I discuss it contains a tweet from a Times economics editor saying that all journalists (and yours truly) take a stance and select facts that supports this stance.
There is actually a third type of journalism, which you could call acrobatic discourse, because it is always looking for balance. It is sometimes called ‘shape of the earth: sides differ’ journalism. Its merit is that it appears not to take sides, but as this extended name is meant to demonstrate, it is certainly not scientific. It is the kind of journalism that says the claim that £350 million a week goes to Brussels and could be spent on the NHS is ‘contested’, rather than simply untrue. In that sense, it can be uninformative and misleading, whereas scientific reporting is informative and is not misleading. Here is a twitter thread from Eric Umansky on a particularly bad example from the New York Times. Of course acrobatic journalism is easier and keeps the journalist out of trouble.
One of the side effects of acrobatic journalism is that it typically defines the two sides it wishes to balance. It therefore tends to be consensus journalism, where the consensus is defined by the politicians on either side. To see why this is problematic you just need to look at how Brexit is discussed and reported by the BBC since the referendum.
I began writing this post during the debate surrounding Nick Robinson’s Steve Hewlett Memorial Lecture. It is certainly strange for that debate to focus on outfits like The Canary, rather than the elephants in the room that produce political journalism to millions every day, who also tend to criticise the BBC whenever they get the opportunity. Yet the copy from these newspapers, and not The Canary, is regularly discussed by the broadcast media. The emergence of left social media journalism is a result of the consensus defining by-product of acrobatic journalism, which for a year or more defined the other side as the PLP rather than the Labour leadership.
I suspect many journalists would say that my idea of them being an amateur scientist is just impractical in this day and age, when they have so little time and resources. But what I have in mind (journalism as amateur scientists) is not very different from what journalists on the Financial Times do day in and day out. Chris Cook is an example of a journalist working in the broadcast media who does the same. But it is wrong to blame individual journalists for being more acrobatic than scientific, because the institutions they work for often demand it.
Nick Robinson’s lecture is much more nuanced and interesting that the subsequent media discussion would suggest. For example he identifies the problem with the way Facebook selects news that is discussed in more detail by Zeynep Tufekci in this TED talk. But there are two elephants in the room that he fails to discuss: the role of the increasingly politicised right wing press I have already mentioned, and the conflict between scientific and acrobatic journalism, both of which he praises without addressing the conflicts between them. 
 There is a clear example of this in the comments he recalls making on the Brexit debate just before the vote. He proudly says he called the £350 million claim untrue, but he then adds