The neuromarketing of politics and voting

Thomas Zoëga Ramsøy

October 2, 2020

Are voters rational?

In this important year of voting in the US, there is a global interest in understanding what is happening during voting. Along political scientists, psychologists and neuroscientists have chimed in to better understand what happens during voting. Is it likely that, given the past decades of behavioral economics showing our profound irrationality, we are rational when it comes to choosing the right candidate for a presidency or other major political position of power? Unsurprisingly, scholars around the world agree that our political opinions are indeed skewed.

We act more like tribes members than members of a larger national or global community. At best, scholars such as Prof. Justin Wolfers suggest that "voters make systematic attribution errors and are best characterized as quasi-rational." Quasi-rational. That's a word we'll have to think a bit about. Does this mean that some of the most important decisions we make as members of society are at the whim of our emotions? Indeed, there is much to suggest this. And to make matters worse - or more interesting - several studies suggest that we make voting decisions based on non-political reasons. A few examples include:

  • Party cues are processed by brain regions that are typically more involved in emotional responses than controlled thinking (Westen et al, 2004)
  • During experiments that show that people's choices are affected by party cues, the participants still hold that party cues are the least influential factor in their choices (Cohen, 2003)
  • Observing one's party losing leads to a change in testosterone levels that is comparable to a personal status loss (Stanton et al. 2009)
  • Showing poll results affects many voters, as poll winners are perceived to be stronger than if the polls were not shown (Cukierman, 1991)

Based on this, it is safe to say that humans are not making political choices purely based on rational, cognitive processes. Emotional and social -- even tribal -- processes are at stake.

Predicting outcomes from independent measures

At the core of our political choices is therefore that our votes are cast through our emotions. This also suggests that we can look even deeper into how emotions drive voting choices. Here, an even more interesting research trend can be seen. In a study by Armstrong and colleagues (2010) showed just how direct and emotional our voting behavior can be. In the study, performed on young adults in Australia and New Zealand, the participants were shown the faces of people for a brief time. Here, they had to make snap judgments about how competent each person was.

Unbeknownst to the participants, the faces included people who were potential candidates for the Democratic Party nomination and for the Republican Party nomination for the 2008 U.S. Presidential election. The results surprisingly showed that face ratings matched the political outcomes! Interestingly, this was shown in a sample of people who did not know about political candidates in the US.

Politics and captains at face value

An even more staggering result was found in a study by Antonakis and Dalgas, of children who were asked who they would like to be the captain of their play ship. The faces they chose as captain(s) were also the ones that were more likely to be chosen to the political post.

Two photos of two men side by side.
In the study, both children and adults rated the faces. In the final election, Jean-Jacques Denis (left) lost the vote to Laurent Hénart (right). In the study, 77% of children and 67% of adults in experiment 1 chose Hénart to be the captain of their ship.

Fast herd or slow ideology?

Sometimes politicians make claims that may seem at odds with the ideological background they represent. Would you agree with your party if they presented a statement that went against their ideological foundation? Put this way, you would probably not, right? But in fact, you’d very likely do so. It’s been known for a while that in the face of conflicting information about political ideology and group belonging, people tend to follow their group and dispose of their ideology (at least for a while). Groupthink comes first. But how does this happen? Does knowing the party provenance of a statement leads to faster or slower responses? Does knowing that statement X comes from party B make it less or more easy to make up your mind?

In a prior study, in my lab at the Copenhagen Business School, we approached this problem, in which the literature has suggested two opposing proposals. On the one hand, knowing the political party behind a statement could trigger some heuristic that makes decisions easier. On the other hand, knowing the sender could trigger a more complex weighing of the opportunities, leading to an overall more difficult decision process (even though it would only take an additional few milliseconds). In two related studies, we used subjects response time to assess the level of conflict, adhering to prior studies (e.g. this one, PDF).

A graph showing political party policies.

In the experiment (top) we tested how people responded to left- and right-wing political statements that were either associated with a left-wing party (A), a right-wing party (B) or a neutral outlet (C). The results (bottom) showed that during a conflict between one's ideology and party loyalty, people typically towed the party line at the cost of the true ideology. This level of conflict was also related to longer response times, depending on the strength of one's party sympathy and whether the statement was consistent or inconsistent with the party ideology.

The article, published in the journal Political Behavior, bears the saying title “Motivated Reasoning and Political Parties: Evidence for Increased Processing in the Face of Party Cues“. The abstract reads as follows:

Extant research in political science has demonstrated that citizens’ opinions on policies are influenced by their attachment to the party sponsoring them. At the same time, little evidence exists illuminating the psychological processes through which such party cues are filtered. From the psychological literature on source cues, we derive two possible hypotheses: (1) party cues activate heuristic processing aimed at minimizing the processing effort during opinion formation, and (2) party cues activate group motivational processes that compel citizens to support 15 the position of their party. As part of the latter processes, the presence of party cues would make individuals engage in effortful motivated reasoning to produce arguments for the correctness of their party’s position. Following psychological research, we use response latency to measure processing effort and, in support of the motivated reasoning hypothesis, demonstrate that across student and nationally representative samples, the presence of party cues increases processing effort.

The article is available from this page.

Studying implicit biases of political attitudes

Based on these studies, it is very clear that voters hold implicit biases about political candidates. Moreover, it is clear that we are affected by non-political aspects in our voting. The question remains: is it possible to test for these implicit biases? The answer is straightforward: yes, this is absolutely possible. To qualify this, we can identify two main ways of testing for implicit biases:

Method 1 - Research testing

In this method, we use neuroscience methods such as eye-tracking and EEG brain scanning to measure specific responses to political candidates, slogans, messages, and more. To make it relatively simple, neuroscience measures of political candidates and elements can typically answer questions such as the following:

  • What is the emotional "gut reaction" to each of the candidates?
  • Which way is the best way to portray a political candidate?
  • How are voters responding to different political slogans?
  • How much of a political argument do voters follow, comprehend, and recall?

Method 2 - Explore panel testing

When testing with online methods, we do not have the luxury of tapping directly into voters' brains. However, we do have neuroscience-validated methods, which allows us to do the next best thing, and at a much faster and national or even global scale. The advancement of both online technologies and neuroscience validation methods, we now have a host of measures at our disposal. Let's first have a look at how some current ads for the US Presidential election score on attention and cognitive demand, using the Predict tool. First, we look at an ad from Democrat candidate Joe Biden:

Joe Biden ad with heatmaps over it.

In this Biden campaign ad, the strongest attention goes to the main message, and viewers are likely to get the main message, and that the message comes from candidate Biden. There is some attention to his face, suggesting that viewers will also connect the message to Biden. Cognitive Demand is low (16.75%) and Clarity is high (63.65%), showing that viewers will have a clear impression of the most salient features in this ad. However, little to no attention is brought to the Biden logo or the slogan, suggesting that viewers are not likely to understand or remember the slogan from this ad.

We can then look at a political ad from President Donald Trump:

Trump ad with heatmaps over it.

In the Trump 2020 campaign ad, visual attention does not go to critical elements such as Trump's face, name, or the signed document he is holding. These are, we assume, critical elements to be understood. Instead, attention is drawn to the call to action and that the ad is sponsored by Trump himself. Here, Cognitive Demand is still low (20.88%) but the Clarity score is suboptimal (54.04%), suggesting that viewers' attention will not be led through the ad in a coherent way. Taken together, a version of this ad with fewer elements would benefit the ad to get the main messages across.

When turning to online testing, using the Explore panel testing tool, we have more measures and metrics at our disposal. By using methods such as the Implicit Association Test, Fast Response Test, and other methods, we can ask questions such as:

  • What are the emotional "gut responses" to political candidates? (same as Research)
  • What implicit emotional and cognitive associations do voters have for political candidates?
  • What implicit associations do voters have towards political slogans, programs, and ads?
  • What do voters remember from political materials they are exposed to?
  • To what extent do political materials produce changes in voting behavior and voter confidence?

These are but a few of many possible elements that online testing allows with a quick turnaround, broad national or global reach, and low cost.,

Neuropolitics is the way forward

Through the past couple of decades, political scientists have begun to look at how voter sentiments, preferences, and choices are influenced by both our brain's wiring, and how the brain learns over time. This is by far not the first time that the combination of neuroscience and politics has been suggested -- going under headings such as "neuropolitics." However, this is the time in which science and technology allow us to both have scalable, reliable methods that can be used to measure multiple touchpoints of voters and political candidates. True, there have been some stumbles along the way, such as the way in which the New York Times and Maro Iacoboni overly simplified and over-interpreted brain activity to political candidates (see here for harsh criticism of the NY Times piece).

Today, we are ready for a new and fresh take at neuropolitics -- one that combines the depth of view from Research, the high scalability of Explore, and the predictive algorithms of Predict. Stay tuned for more, and if you're interested in hearing more, tune in by writing to us here.

The neuromarketing of politics and voting

Neurons Icon

Ready to drive revenue with creatives that work?

Get a free demo