Just a nudge in the dark
In a world where academics are assessed on their impact as well as the number and quality of their peer-reviewed papers, findings must answer one of two questions in the affirmative – “Is it important?”, or “Is it cool?”
There is no doubt that Behavioural Economics is cool – this has been assured by books like Dan Ariely’s Predictably Irrational, which brought the subject to widespread attention, as well as Nobel Prize winner Daniel Kahneman’s Thinking Fast and Slow.
From a public policy perspective, Thaler and Sunstein’s book, Nudge, is probably the most important of these, as it is geared most heavily towards applying insights from behavioural economics and psychology to real world problems with which governments are concerned.
The attention of policymakers, as well as the nascent replication crisis in psychology changes the question that must be asked of findings – it is not sufficient simply to be cool. The majority of the famous studies in this field are conducted in laboratories, primarily on university students, and primarily on female economics and psychology undergraduate students at that. This is the way science is conducted, but it is important to know whether findings generalise to other contexts.
At the same time, large scale field experiments can be costly, time consuming, and are subject to a lot of noise. These are the very reasons that lab experiments are so attractive for researchers.
There is a halfway house – conducting lab-like experiments, where we take the structure of the lab into the field, where things are noisy and people might be distracted by other things. In a study published today, we do just that.
Participants were members of the public, and they were asked to play a trust game. In this game, one player starts off with a small amount of money, for example £1, and is asked if they’d like to trust the other player. If they do so, the other player receives the £1, plus an additional £3, giving them £4 in total. They then choose whether or not to split the money with the first player. If they do, both players keep £2 – if they don’t, the second player gets £4 and the first player walks away empty handed. The question is – will you trust, and are you trustworthy?
Previous experiments, carried out in the lab, found that when you frame the game as a “Betrayal game”, people were significantly less trusting, and significantly less trustworthy as well, than when the game is described as a “Partnership game”. In our experiment, we showed graphs for one condition or the other to different participants. So, some people saw the results for the “Betrayal game”, while other people saw the results for the “Partnership game”. According to theories on framing, the first group should be less trusting and less trustworthy, as they have been “primed” by Betrayal.
What do we find? Interesting, despite the noisy environment, and the fact that we were using members of the public rather than students, we still saw some effects. People were significantly less likely to trust each-other if they were primed with the untrusting game, but, they were no less likely to split the money. This suggests that outside of the lab, people can still be primed in their beliefs about other people, but that their own preferences don’t change – in the real world, people are more difficult to make less nice, but it might still be possible to make them trust other people less. We should never read too far into an individual study, but if it holds up in repeated and larger studies, the implications are potentially interesting. Not trusting other people is sensible if other people aren’t trustworthy, but if they are, we can lose many beneficial opportunities if we trust too little. If priming can lead people to hold incorrect beliefs about how trustworthy other people are, it could actually make us worse off as a society.
 Ariely (2008): “Predictably Irrational”, Harper Collins
 Kahneman (2012): “Thinking, Fast and Slow” Penguin.
 Thaler and Sunstein (2008): “Nudge: Improving decisions about health, wealth and happiness” Penguin
 Burnham, McCabe and Smith (2000) “Friend-or-foe intentionality priming in an extensive form trust game” Journal of Economic Behavior and Organisation Vol 43 No.1 pp57-73