Media & persuasion

Far more happens in a day than any paper or bulletin can carry, so somebody has to choose. For a study published in 1950, the wire editor of an American newspaper kept every story he turned down in one week, with his reason. He had rejected nine-tenths of what came in. A person who decides what passes through to the audience like this is a gatekeeper, and the news you see is always a selection somebody made.

The selection has an effect of its own. A study published in 1972 compared what undecided voters in one American town named as the big issues of an election with what their papers and broadcasts had covered most. The two lists matched closely. The media may not tell people what to think, the saying goes, but they are very good at telling people what to think about. This power is called agenda-setting.

A message can change a mind in two ways, which the psychologists Richard Petty and John Cacioppo described in the 1980s. On the central route the listener weighs the argument itself: is the evidence good, does the reasoning hold? It takes effort, and a mind changed this way tends to stay changed. On the peripheral route the listener hardly considers the argument and is moved by surface cues instead: an attractive face, a confident voice, a catchy tune, the sense that everyone else is convinced.

Much advertising takes the peripheral route, because few people will study the case for a soft drink. A company pays a famous footballer to be seen wearing its watch, and hopes that his glow rubs off on the product. That paid vouching is called an endorsement. It tells you nothing about the watch.

Advertising wants your money. Propaganda wants your loyalty. It is one-sided information spread on purpose to win people to a cause, and it selects and slants where honest reporting would try to be fair. In 1937 a group of American scholars and journalists, alarmed at how much of it the public now met, published a list of its seven commonest devices so that people could learn to spot them.

Two are still everywhere. Name-calling hangs an ugly label on an opponent, such as traitor, extremist or enemy of the people, so that the audience rejects him without weighing what he said. The bandwagon tells you that everyone is already on board, and that you had better join before you are left behind. Neither offers a reason. One works on dislike and the other on the fear of standing alone, and both do their work before thinking starts.

The simplest device of all needs no cleverness, only stamina. Say a thing again and again, on every channel, and people start to believe it. This is repetition, and psychologists have measured its effect. In a 1977 experiment volunteers rated how true various statements seemed, in three sessions two weeks apart. Statements they had met in an earlier session were rated truer than new ones, whether or not they were in fact true.

The usual explanation is a shortcut the mind takes. A statement we have met before is easier to take in, and that ease feels like recognising something known. So familiarity is mistaken for truth. Slogans, jingles and talking points are built on this, and so is the liar’s rule that a big claim repeated often enough will stick. Having heard a thing many times is evidence only that someone wanted it heard.

False information comes in two kinds, and the difference lies in the person passing it on. Suppose your aunt forwards a health tip that is wrong. She believes it and means well. False information shared by honest mistake is misinformation. Now suppose someone invented that tip, knowing it was false, to sell a product or to stir up anger. False information spread deliberately to deceive is disinformation.

The claim can be word for word the same. What differs is the senders, and whether they meant to deceive. That is why disinformation works so well: its makers need only start it off, and honest people, passing it on as misinformation, do the rest of the work for them. It is also why being sincere is no defence against spreading a falsehood. The aunt did the harm without telling a lie.

In 2018 researchers at MIT reported on about 126,000 stories that had spread on Twitter over eleven years, each checked and labelled true or false by independent fact-checkers. False stories won on every measure. A false story was 70 per cent more likely to be passed on than a true one, and true stories took about six times as long to reach 1,500 people.

The researchers looked for the reason in the stories themselves. The false ones stood out in one way: they were more novel and more surprising, and people share what startles them. A lie can be shaped to startle, while the truth is often something people half knew already. An unverified story passed from person to person like this, with no checkable source, is a rumour. Machines were not to blame. Automated accounts spread true and false stories at the same rate, so people made the difference.

A report can mislead without one false word, by its choice of wording. In 1981 the psychologists Amos Tversky and Daniel Kahneman asked people to choose a plan against a disease expected to kill 600. Of one group, told that a plan meant ’200 people will be saved’, 72 per cent chose it. Of another, told that the very same plan meant ’400 people will die’, only 22 per cent did. Presenting the same facts in a wording that steers the audience’s response is called framing.

Single words do it too. One crowd is ‘protesters’ in one paper and ‘a mob’ in another, and one government is an ‘administration’ or a ‘regime’. Words like these, chosen for the feeling they carry, are called loaded language. They hand the reader a verdict before he has seen any evidence.

Numbers look like the safe part of a news story. They are where a careful reader should slow down. In 2015 headlines announced that a daily portion of processed meat raises the risk of bowel cancer by 18 per cent. It sounded alarming. What the headlines left out was the base rate, the risk a person starts from. About six people in a hundred get that cancer in a lifetime, and an 18 per cent rise takes it to about seven in a hundred. Both statements are true, and only one of them lets you judge.

A second trick also uses only true numbers. A company reports the three months in which its sales rose and says nothing of the nine in which they fell. Choosing just the data that suits your case is called cherry-picking, and it is hard to catch, because every figure you are shown is real.

A chart can exaggerate without lying. Start a bar chart’s upright scale at 95, and a rise from 96 to 97 looks like a doubling. The honest rule for bars is that the scale starts at zero, so that a bar twice as tall means a figure twice as large.

A survey has to be read with the same care. Nobody can ask a whole country, so pollsters ask a small group, the sample, and reason from it to everyone. That works only if the sample resembles the whole. In 1936 an American magazine collected replies from 2.4 million people and forecast a clear defeat for President Roosevelt. He won easily. The magazine had drawn its names largely from telephone books and car registrations, which in 1936 listed the better-off. A huge sample of the wrong people is worth less than a small sample of the right ones.

‘People who do this live longer.’ A headline like that reports two things that go together and invites you to conclude that one causes the other. Two things that rise and fall together show a correlation, and a correlation alone does not say why.

The standard example is ice cream. In months when more ice cream is sold, more people drown. Ice cream does not cause drowning. Hot weather sends people to the ice-cream stall and into the water, and so drives both numbers at once. A hidden third cause like this, which produces a link between two things that do not affect each other, is called a confounding factor. Before believing that one thing causes another, ask what else the people who do the first thing have in common. They may be richer, younger or healthier to begin with.

The newest gatekeeper is not a person. A social media feed is put in order by software that learns what you click on and shows you more of the same, because that keeps you scrolling. In 2011 the writer Eli Pariser warned where this leads. Each user ends up in a private world of agreeable news with the rest quietly left out, and he called it a filter bubble. Its walls are invisible, since you cannot see what you were not shown.

People build something similar for themselves. We choose friends, channels and groups that share our views, and before long every opinion we hear is our own coming back. This is an echo chamber. Inside one, a claim feels true because everyone repeats it, and a mistake goes uncorrected because nobody present disagrees. The two together explain how neighbours can come to live in different worlds.

What should a reader do with a striking claim from an unfamiliar source? Researchers at Stanford University watched historians, students and professional fact-checkers try to judge websites. The historians and students read each page closely, from top to bottom, and were often taken in by a smart design. The fact-checkers left the page almost at once and opened new tabs to see what other sources said about the site and who was behind it. They were faster and far more accurate. The habit is called lateral reading: reading across sources before reading down one.

Two questions do most of the work. Who is saying this, and what do they gain if I believe it? For claims that have already spread widely, independent teams test them against the evidence and publish their verdict along with their sources. That work is called fact-checking.

Long before feeds and forwards, Islam made checking reports a duty. The Qur’an tells the believers that if a wicked person brings them news they are to verify it, so that they do not harm people in ignorance and then regret what they have done. The command is in the chapter al-Ḥujurāt, and the act it asks for is called tabayyun: making sure before you act.

The duty covers passing things on as well as believing them. In a hadith in the introduction to Ṣaḥīḥ Muslim, the Prophet said that it is enough of a lie for a man to narrate everything he hears. Repeating everything he hears, he is bound to repeat something false. The 2018 study, separately, found people, not machines, behind falsehood’s lead. One pause before sharing, to ask where a thing came from and whether it is true, is what this reading comes down to.