If you've got lucky, it's easy to convince people you're a sage

"Expert" predictions only need a few lucky guesses, then everyone believes them.

LSE's Nattavudh Powdthavee and Nanyang University's Yohanes E. Riyanto presented a paper last month titled Why Do People Pay for Useless Advice? (pdf) The authors demonstrate that people are prepared to pay for investment advice which they know cannot possibly be based on any real evidence – provided the advisor has got things right in the past.

Participants in the experiment were told to bet on five consecutive rounds of coin flips (they could pick the stake, provided it was higher than a minimum level and wouldn't make them lose all their money before the last round). Since they were using a fair coin, the chance of heads or tails coming up each time was 50-50, and the chance of someone calling the correct flips in all five rounds is one in thirty two. More importantly, no amount of expertise can help one predict the outcome of a sequence of coin tosses.

So when the participants were offered the chance to buy "predictions", you would expect them to ignore them. And most of them did, at least for round one (although even then, around 15 per cent of them did pay). But after the first round of tosses, everyone got to see their predictions, even those who hadn't paid. And for half the participants, those predictions were correct. Still, it's obviously just chance, right?

Apparently not so obvious. Those who got a correct "prediction" in round one were three times more likely to pay for one in round two than those who'd got an incorrect one. And by the time a subset of the group had got four correct predictions in a row, the possibility that they would buy a prediction for the fifth round had risen from 15 to 40 per cent.

In other words, just a few lucky guesses are all it takes to get people to pay for what the authors call "transparently useless advice". As they conclude:

Our experiment accurately describes how the real "false" experts typically operate in the information market, and the findings of this paper help to highlight how easy it is for an average person to form a belief in an expert when none may actually exist. 

Something worth bearing in mind next time you are looking at horse tips, stock pickers, or hedge funds, maybe.

George Bush tosses a coin. Is he an expert? Photograph: Getty Images

Alex Hern is a technology reporter for the Guardian. He was formerly staff writer at the New Statesman. You should follow Alex on Twitter.

New Statesman
Show Hide image

Quiz: Can you identify fake news?

The furore around "fake" news shows no sign of abating. Can you spot what's real and what's not?

Hillary Clinton has spoken out today to warn about the fake news epidemic sweeping the world. Clinton went as far as to say that "lives are at risk" from fake news, the day after Pope Francis compared reading fake news to eating poop. (Side note: with real news like that, who needs the fake stuff?)

The sweeping distrust in fake news has caused some confusion, however, as many are unsure about how to actually tell the reals and the fakes apart. Short from seeing whether the logo will scratch off and asking the man from the market where he got it from, how can you really identify fake news? Take our test to see whether you have all the answers.

 

 

In all seriousness, many claim that identifying fake news is a simple matter of checking the source and disbelieving anything "too good to be true". Unfortunately, however, fake news outlets post real stories too, and real news outlets often slip up and publish the fakes. Use fact-checking websites like Snopes to really get to the bottom of a story, and always do a quick Google before you share anything. 

Amelia Tait is a technology and digital culture writer at the New Statesman.