But that did not stop experts from sharing their opinions. Plenty of economists, business leaders and political scientists argued for remaining in the EU. So why didn’t they win the argument?
To try and find out, my team and I studied how people use different information when it comes to voting. We presented participants with a simple voting situation in which they had to choose between three simple options: vote red, vote blue, or abstain. For the purposes of our experiment, and unlike in an election or referendum, each time participants voted, there was a definite “correct” answer determined at random by a computer.
Their choices became a political bet. If they chose correctly, and others did too, they won a small cash prize. If enough of them chose incorrectly, they’d all lose out. There was an incentive for participants to come to the right decision.
We also presented the voters with two types of information. “Public information” was seen by everyone and referred to as “expert”. “Private information” was given to individuals and referred to “personal opinion”. Each type was also presented with a probability of it being correct, ranging from 50%, to as high as 95%.
For the experiment, we weren’t interested in which way participants voted – but rather how they used the information they were given in reaching a decision. Which would hold more sway? Expert, public information available for all to see, or the non-expert, privately held information?
We predicted that voters would opt for the choice where the probability of it being correct was highest – or that they’d abstain where it was not clear. If, for example, expert advice said to vote red with 95% certainty of that being “correct” and private information said to vote blue with only 55% certainty of being correct, then logically and rationally participants would vote red.
Or, if both options had a similar probability (let’s say expert advice points to red with 95% certainty, whereas private information points to blue with 85% certainty) then the best action would be to abstain, in the hope that others who had a clearer picture of the correct choice would determine the winner.
But people did not do as we expected.
Although some followed that logical approach, the majority did not vote efficiently. They followed their personal information when it made no sense to do so. Around 55% of participants voted on personal information, against expert information, when only around 10% should have done so.
We found this behaviour consistently every time we ran the experiment. Even when voters had private information that we deemed borderline useless – where the probability of it being correct was about 50% – they still followed this private advice, ignoring public, expert options.
Going with guts
As a result, all the participants earned far less money in the experiment that they could have done.
Our experiment was a standalone economic one. But it presents an interesting observation of what might have happened when it came to the EU referendum in 2016, or the election of President Donald Trump in the US.
For political reasons, in those instances the electorate went against what was expected, and against the weight of “expert evidence and advice”. Of course, for both outcomes the jury is still out on what was the better economic decision.
In our experiment, we set up a gamble where participants would only end up losing out by small amounts. When the stakes are much higher, ensuring that expert advice gets through to everyone might be a gamble none of us can afford to lose.
For political campaigns to be successful, a focus on objective facts will only get you so far – and often that’s not far enough to win over the electorate. Both the messenger, and the message, need to be right.
But whatever the rights and wrongs of Brexit and Trump, we found that private information carries more sway when it comes to informing individual choices. It does not matter how good experts are or what they say, voters will often favour what their guts tell them when it comes to choosing where to put their X in the ballot box.