Take a look at mainstream or social media coverage of the US presidential election and you’ll find a lot of head scratching about how Donald Trump has managed to garner so much support. (Some of it has come from yours truly.)
Many of the arguments being tossed around boil down to this: Clinton is clearly better qualified than Trump, and her policy platforms are so objectively superior, that it hardly makes sense that the two candidates are so close in the polls. But, based on what we know about human psychology, it makes more sense than many of us may think. Or hope.
A few years ago, University of Chicago psychologist Christopher Hsee conducted a neat experiment using hypothetical ice cream. He asked one group of people to imagine an underfilled cup of their favourite ice cream – 8 oz of ice cream in a 10 oz cup – and asked them to offer him fair price to purchase it. The average willingness-to-pay was $1.66. To a different group of people, he offered a similar deal, but with a twist. This time, he offered them less ice cream – 7 oz, to be exact; but in this case, the ice cream was in a 5 oz cup, which runneth over. Not surprisingly, the overfilled cup garnered a higher price: $2.26. When he presented the same ice cream options side-by-side to a third group of people, the average willingness-to-pay reversed. People were now willing to pay more – as one might hope! – for the larger amount of ice cream, regardless of the cup’s size, than they would for the smaller amount.
Hsee explained the results this way: The absolute amount of ice cream, 7 oz or 8 oz, in a cup is difficult to instinctively evaluate for all but the most seasoned ice cream experts. But the appearance of “fullness” is easy to evaluate instinctively for pretty much anybody; we all know when we’re getting a good deal or when we’re being ripped off. In the absence of expertise in a situation like this, the mind uses the information it can to make a judgment, using the easy-to-evaluate factor in this case.
The psychologist Paul Slovic took these results a step further, suggesting that what makes a factor easy or difficult to evaluate has to do with the emotional connection we can make with it. Easy-to-evaluate factors – such as the fullness of a cup of ice cream – are ones that map neatly into what Slovic and others call System 1 in the brain. In the most basic terms, System 1 encodes information by way of an instinctive emotional response, feelings such as “dread’’ and ‘‘optimism,” or qualities such as ‘‘goodness’’ and ‘‘badness.’’ It’s easy to feel, on an emotional level, that an overfilled ice cream is good. It’s what many of us in my field refer to as deciding with our hearts.
Factors such as the absolute amount of ice cream, on the other hand, require our brains to use System 2, which relies on technical or scientific expertise and number-crunching. In other words, System 2 requires us to make decisions with our minds. In the absence of this expertise – crunching the numbers is hard work – our “hearts” fill the void.
Bringing the two cups of ice cream together in a side-by-side comparison provides much-needed perspective. Consumers can now see that 8 oz is better than 7 oz, regardless of the cup size. Likewise, people can recode certain emotional factors (such as cup size) as being less important to them overall when compared to other, more relevant ones (such as the absolute amount of ice cream). In this sense, a side-by-side comparison is a simple but effective decision-support tool because it provides much needed context and helps people to set priorities among variables that are important to them and ones that are less so.
Elections, and the campaigns leading up to them, are all about side-by-side comparisons, where the conflict between heart and mind should be diminished.
Candidates and policy options square off against one another, and the airwaves are crammed with information about the candidates’ strengths and shortcomings. At their core, elections are all about setting priorities among long lists of concerns, and making decisions that reflect our judgments about the issues that warrant the most attention. So, all of these comparisons should help, right?
Not so fast.
In Hsee’s ice cream study, the flavour of the ice cream itself was not a factor in the experiment; it was always assumed to be respondents’ preferred flavour. Thus, the only variables of interest were the relative fullness of the cup and the absolute amount of ice cream. But, in elections, if fullness and amount are the candidate’s stances on the issues, then the candidates themselves represent different flavours of ice cream.
Does a side-by-side comparison help in a situation like this? In an attempt to find an answer, Robyn Wilson, a professor at the Ohio State University, and I conducted a study in which we compared two problems that people face in many small towns across the Midwest: property crime and deer overpopulation. Both represent risks to human health and property. We knew from a prior study that crime would lead to a strong emotional reaction, while deer overpopulation would, relatively speaking, pale in comparison. So, when we presented information to people about the risks posed to people and property, we made sure that the risks from deer overpopulation were higher than those from property crime.
What did we find? When these two problems were compared side-by-side, people routinely chose to devote management attention and resources to property crime, even though the damages from deer overpopulation were always more severe. To test our ideas further, we repeated the experiment several times, each time making the damages from deer overpopulation even bigger, while holding those from property crime constant. No matter how big the difference in damages, we could not get people to take action on deer overpopulation. It was as though people simply couldn’t get past the emotional tug of the problem – crime – no matter what empirical data we presented about it.
Bringing this back to elections, results like ours – especially when they are combined with other research on the importance of the cultural underpinnings of preference – strongly suggest that the usual evaluative benefits of a side-by-side comparison virtually disappear when the options themselves invoke strong emotions. Robyn Wilson and I called this “value neglect.” The emotional signals sent out by the alternatives themselves are so powerful that we neglect to consider the objective information that should help us to distinguish the pros and cons of the options in the first place.
This isn’t good news from the standpoint of voters who claim they are voting for one candidate over another because of where each candidate stands on the issues. In reality, these voters are probably making their selection based in large part on the emotional connections they forge with the candidates themselves.
In other words, voters prefer Donald Trump – or Hillary Clinton, for that matter – because of the emotional connection they (or the parties they represent) have forged with the electorate. So, in a democracy where people are almost evenly divided between Republicans and Democrats, it’s hardly a surprise that people don’t take the time to look more closely at the issues. Put another way, when it comes to political ice cream flavors, people are either partial to Donald or Hillary; how much or – in Trump’s case – how little they actually receive in their proverbial cup or cone barely registers. The best evidence for this is that, in spite of the key differences between Hillary Clinton and Donald Trump (not to mention between countless other candidates and initiatives that can be found further down the ballot), they find themselves uncomfortably close to one another in terms of the expected popular vote where
All of this raises the question: Is there anything we can do if we want to try to make evidence-based decisions about candidates and ballot initiatives? The answer is, sort of.
Ongoing research in my lab suggests that considering the pros and cons of anonymized options – that is, evaluations of options where you only see the details that differentiate them, and not the labels (like “Trump” or Clinton”) that define them – can lead people to establish preferences and ranked preference orders that are better calibrated to their own values and concerns.
Likewise, decision-support processes that start by asking people to set priorities before they evaluate alternatives similarly help, initially, to overcome the biases associated with the emotional appeal of identifiable options. In the domain of elections, websites such as Vote Compass and I Side With are designed to do exactly this.
Still, tools and approaches like these are far from perfect because, in the end, voters have to look at the candidates and their emotions will inevitably weigh in during decision-making. Indeed, in a recent study of decision support tools that pointed to certain alternatives based on an individual’s stated priorities, a majority of users indicated they would rather stick with their own unaided preferences.
So are a majority of voters on November 8 likely to choose the candidates and ballot initiatives that align most neatly with their values, concerns and priorities? The answer, sadly, is probably not.