A new paper, published by our Department of Economics, shows that voters tend to retain strong attachment to their own opinions even when this is challenged by evidence. It also suggests that voters tend to distrust experts and place disproportionate weight on their own experiences and evidence.

For the study, which has interesting parallels with the Leave vote in the referendum on Britain’s membership of the EU or the election of President Trump in the United States in 2016, 188 participants were asked to vote between two options – vote red, or vote blue, with a third possibility to abstain.

Voters were told that one option would be better for them economically and would result in them receiving a real cash, monetary payment if sufficient voters ‘chose correctly’. But participants were only given limited information on which to base their political ‘bet’.

Asssessing the political bet

Researchers were not interested in which way voters swung, but rather how they used different types of information in coming to their conclusion. They presented participants in the experiment with two types of information which varied in quality:

  • ‘Public information’, seen by all and which they referred to as ‘expert information’ and;
  • ‘Private information’, a nugget of insight given to voters individually and referred to as ‘personal opinion’.

There were no certainties for voters. Both sets of information were presented with the caveat that there was only a probability of this being correct ranging from under 50%, to as high as 95%.

Rationally, the researchers predicted that voters would opt with the choice where the probability that a certain decision was correct was highest, or they would abstain where it was not clear. If expert advice said to vote red with 95% certainty of that being ‘correct’ and private information said to vote blue with only 55% certainty, then logically and rationally participants would vote red, the researchers believed.

But although some followed this logic, the majority did not vote on this basis. Instead they followed their personal opinion. In fact around 55% of participants voted based on personal information, against expert information, when only around 10% should have done so even when the evidence was described as ‘borderline useless’.

The team found the same behaviours every time they ran the experiment. As a result, all participants earned far less money in the experiment that they could have.

Lessons for future evidence-based campaigns

Lead researcher, Dr Javier Rivas explained: “We studied an election scenario where subjects had to choose between two options and where options were presented as either good or bad in terms of the economic outcome for participants. By presenting different types of information we gave participants insights about which was the better option, but surprisingly many rejected this ‘advice’ and went with their gut instinct instead.

“Whilst our experiment is in no way a replica for a real election or referendum, there may be interesting parallels with the Brexit vote in the UK or the election of President Trump. In both instances – as has been widely observed – the weight of evidence about the economic consequences of the different decisions was ignored.”

They suggest that for those wishing to run successful evidence-based campaigns in the future these findings reveal interesting insights about what carries weight with voters and how messages need to be communicated.

This paper did not study the importance of the messenger in delivering a message - for this study all messengers were neutral. Future research will look at this as well as how the topic of information had an impact of voters, for example comparing evidence about immigration or the NHS in relation to Brexit.

The researchers are keen to point out that the long-term economic outcomes of Brexit or President Trump are not yet known.

Read Javier Rivas in The Conversation 'Our experiment into how voters think shows that they go with their guts'.