LBS logo London experience. World impact.

Donald Trump's election: a behavioural science perspective

The Trump campaign influenced voters’ attitudes to risk, uncertainty, loss and controversy.

By David Faro . 09 November 2016

The Trump campaign influenced voters’ attitudes to risk, uncertainty, loss and controversy.


What can behavioural science tell us about the US election result? Leaving aside the candidates’ messages, Donald Trump’s victory, a surprise to many, crisply illustrates some key behavioural concepts. The content and language Trump employed triggered a host of emotions such as fear and anger. His campaign also emphasised one of the great behavioural triggers – uncertainty. Several consequences on voters' behaviour were predictable.

Quantifying and conveying risks

Consider the infamous graphic by the Trump campaign showing a bowl of Skittles sweets with the following message: “If I had a bowl of Skittles and I told you just three would kill you – would you take a handful? That’s our Syrian refugee problem”. This message can be seen as a manifestation of two behavioural concepts identified by psychologists Daniel Kahneman and Amos Tversky.

The first relates to anchoring and insufficient adjustment: When people need to make an estimate, they can be influenced by irrelevant initial “anchors” provided to them and not adjust their final estimates sufficiently. While the audience may know the chance of a terrorist attack is far lower than picking three “bad” skittles, their adjustment based on the message’s implied probability would tend to be insufficient. So the Trump message can lead people to overestimate the likelihood of a terrorist attack linked to refugees or other groups.

The second shows people overweight small probabilities, especially when they are presented with vivid imagery. This means that the probability of a terror attack has a disproportionate impact – compared to less vivid risks such as accidents at work – on our decisions.

How do we reduce the impact of these biases? One remedy for anchoring is to research the numbers on websites such as Factchecker and compare threats of various kinds. Insufficient adjustment from the anchor is in part due to a lack of effort or motivation; giving people enough time and mental resources would help.

To reduce people’s tendency to overweight small probabilities Yuval Rottenstreich, from UC San Diego, and I used the following approach: we asked one group of participants how much they would pay to protect a large amount of cash if there was a small risk of losing it all. We asked a second group to predict how a random MBA student would respond to this dilemma. We then asked this second group to make their own decision.

We found that the members of the second group were less likely to overweight small probabilities, with most predicting that the MBA student would spend little to eliminate the low risk of losing the money. More importantly, the second group anchored on the predictions they made for the random MBA student and subsequently applied less overweighting to their own choice. When facing vivid risks to which we may overreact, it might be useful to first take the perspective of another person, perhaps a friend or a colleague; to consider how he or she would react and then make a decision.

Framing and risk-taking

Another notable feature of the Trump campaign was the emphasis on losses. He talked about how America is economically losing out to China, Mexico and many others. He asked African-Americans what they had to lose (implying they were already in a “loss state”). He talked about people who have been hit by the financial crisis.

Whether it is a conscious decision or not, Trump relied here on a fundamental behavioural theory of decision-making under risk – Prospect Theory (Kahneman and Tversky 1979). The theory predicts that when people adopt a “loss frame” they tend to be more open to risk-taking. This so-called “break-even effect” has been observed in the TV show Deal-or-No-Deal. Players are more likely to reject the banker’s offer and take significant financial risks following a bad round (Post et al 2008). This is also reflected in investors’ behaviour – they tend to hold on to losing stocks, hoping they will recoup their losses (Odean 1998).

In these elections Hillary Clinton represented the status-quo, the safe option with political experience. By emphasising losses, the Trump campaign made voters more likely to choose a risky alternative, the anti-establishment candidate (Surowiecki, 2016).

What could the Clinton campaign have done in response to the loss-framed messaging that favoured Trump? First, emphasising that Trump is a risky candidate would not have helped, since taking the risky option is precisely what loss-frame messaging tends to accomplish. Clinton could instead have focus people’s attention on what voters have – thereby creating a gain-frame – to have attempted to make people more cautious and risk-averse.

Reference points and adaptation

Gain and loss frames are relative to reference points – or expectations – which played several roles in the elections. The one we just noted was about how voters can be made to feel worse off relative to where they used to be, or where they believe they should be.

The public also had expectations about the candidates. One media trend was the narrative that Trump was setting low expectations before the debates. It was reported that he barely prepared for the debates and therefore underestimated the demands of staying focused. True or false, this takes into account the tendency to evaluate events in a narrow frame and relative to a reference. Whether a candidate wins or loses is determined in part by how they perform relative to an expectation. Trump was expected to do poorly and could be seen to “win” a debate in some people’s eyes simply for not being as bad as everyone anticipated.

Setting expectations low is a risky affair. If expectations are low and a candidate does badly, their performance simply confirms what people thought would happen. Setting expectations high and not delivering fully is also risky. Barack Obama was almost guaranteed to disappoint as expectations after his election were so high. His highest ratings in the second term apparently came towards the end of his presidency, perhaps because he was considered more favourable when compared with the “less desirable” Clinton and Trump.

Perhaps most importantly, Trump managed to shield himself from standard comparisons by talking and behaving so unlike any other politician. It was difficult to compare him to Clinton – the dimensions on which he is judged seem very different. Trump was seen as the outsider and was evaluated on his own, in a separate rather than a joint evaluation mode, where one candidate is compared and contrasted to another. This typical joint / comparative process of assessing the nominees rarely featured in this campaign.

Research by Chris Hsee from the University of Chicago shows that whether we evaluate things in a separate or a joint mode affects which dimensions we take into account to make a decision. Dimensions which may be relevant for the presidency, such as years of experience in public service, may be completely ignored when we engage in a separate evaluation of the candidates.

Of course, the thing with reference points is that people adapt to them over time. Your reference point with respect to wealth, for instance, might be your current income. When your income changes, you may feel richer and more satisfied for a while, but then you adapt to this new reference point – hence people refer to increasing income and stable life satisfaction as the hedonic treadmill. A similar thing happens with other stimuli, such as music or irritating noises – people adapt to them.

Campaign messages may have a similar effect. Trump made increasingly controversial statements, perhaps realising he needed to be more and more outrageous as people adapted to his views. People adapt to political talk and jargon and, at some point, may stop processing messages which start to sound scripted. 


The US presidential campaign is a long and difficult affair. It requires months of travelling, talking to audiences and debating, requiring knowledge, ability and stamina (which became an issue in the campaign, to Clinton’s detriment). But it also requires metacognition – assessing one’s knowledge, ability and stamina – from the candidates and their teams. Metacognition is important because it can affect how much we prepare for a task, the level of help we seek and the risk-management options we choose.

Much behavioural research shows a pervasive tendency for people to be overconfident – especially in high-ego situations. We can’t precisely assess whether and the extent to which Clinton or Trump are overconfident – that would require an evidence-based comparison of perceived skill with actual skill. But we observe some signs of this tendency in both candidates. Clinton ignored her doctor’s advice to rest when suffering from pneumonia and nearly collapsed at the 9/11 memorial event. Trump reportedly paid little attention to briefing materials and found preparing for debates boring.

Overconfidence poses great risks for society – it has been linked to failed mergers and acquisitions, wars and economic crises. When exposed, overconfidence also comes with potential reputational costs to the individual. Former Labour Party leader Ed Miliband chose to make one of his key speeches during the UK elections without a teleprompter and forgot to mention key issues concerning the economy and Britain’s deficit. The story persisted throughout his campaign to become the next prime minister, which he lost.

My students often ask me whether overconfidence can be a good thing, especially in competitive settings such as the job market or elections. It’s certainly the case that overconfidence may lead you to take risks that you’re rewarded for, sometimes. Confidence is also a signal for competence (although an imperfect one); overconfident people may be seen as competent, especially when the assessment is made under uncertainty and without feedback.

The choice and looking back

Both Clinton and Trump had strong supporters, but it seems that for many American voters neither candidate was very appealing. When presented with unappealing options, or difficult decisions, people tend to avoid making a choice altogether. Indeed, low voter turnout was an ongoing issue in the campaign.

When people do eventually make a difficult choice, they often feel dissatisfaction or regret. As with other contentious events (Brexit being the obvious example), they struggle to reach closure with their choice. Time can help to achieve a sense of finality with past decisions and events. But closure can also be triggered externally.

In recent work with my colleague Simona Botti and Yangjie Gu from HEC, we showed that engaging in acts that symbolise closure can help people achieve a sense of closure. This can in turn make them more satisfied with difficult decisions they have made. In one study, participants who had chosen from a large set of chocolates were more satisfied with their decision after they placed a lid over the set of options. In another study, participants choosing biscuits were happier with their choice after closing the menu that listed the numerous available alternatives.

Voting for the next US president obviously has more serious ramifications than choosing a chocolate or a biscuit. However, this research shows the importance of reaching closure after making a difficult choice. The candidates’ behaviour after the elections can both facilitate and inhibit closure. But our own cognition and behaviour also play a role.

Comments (3)

KMB1011 8 months and 5 days ago

Disappointing biased research, language, and writing. Didn't expect this from LBS.

tschakawaka 8 months and 7 days ago

Education, Education, Education!If you have nothing to lose (anymore), what can you lose with your next/final decision (anyway)??

urban spaceman 8 months and 9 days ago

Did you think that Hilary would win ?