Trying to persuade people with different political views can be fatiguing and frustrating, but it’s not as pointless as we might think. New research shows that even on divisive issues, far more persuasion attempts succeeded than backfired.
A new study led by Northwestern University and the University of North Carolina at Chapel Hill focuses on the persuasive power of ordinary citizens, flipping the script on how political persuasion is commonly viewed. Prior research has primarily focused on who is persuadable and how political elites can shape public opinion.
The researchers sought to find out to what extent citizen-to-citizen persuasion is possible and identify the strategies and attributes of the people most effective at changing opinions.
What matters is meeting people where they are.”
“Today, with social media, anyone with internet access can broadcast their ideas, potentially to a massive audience, and this shift in who gets a platform calls for a shift in how we study persuasion,” said Martin Naunov, the study’s lead author.
Naunov is an assistant professor of political science and a faculty associate of the Institute for Policy Research at Northwestern.
The study’s co-authors are Timothy J. Ryan, a professor of political science, and Carlos Rueda-Cañòn, a political science graduate student at the University of North Carolina.
Testing hot-button issues like immigration and climate
In the first stage of the study, more than 400 survey respondents who the researchers called “senders,” were incentivized to write an argument that persuades someone who disagrees with them on three hot-button issues: immigration, climate and transgender rights.
In the second stage, the researchers tested to see if any of the sender messages succeeded in changing the opinions of others. More than 3,000 respondents, which were called “receivers,” were first asked about their opinions on immigration, climate and transgender rights on a 0- to 10-point scale. Receivers were then presented with a randomly chosen argument from the senders’ survey that argued against the opinion of the receiver. After reading the arguments, receivers were again asked for their opinion on the same issue to see if their position shifted on the 11-point scale.
In total, the researchers collected more than 9,000 argument evaluations.
A segment of 500 receivers acted as the control group for the study. This group did not receive a persuasive message but were asked their opinion on the issues again to see if their position shifted.
The researchers found that approximately 30% of the persuasion attempts succeeded. Backlash was rare, around 11%, and not distinguishable from measurement error.
“This shows while there isn’t a cascade of attitude change, it’s also not the hopelessly calcified political landscape we often hear about,” Naunov said. “Yes, Americans are increasingly polarized, but at least among ordinary citizens, there’s still a fair amount of openness to competing political viewpoints.”
Who and what makes a good persuasive argument?
In addition to showing that it is possible to change the opinion of someone with an opposing view, the study highlights what kinds of people and arguments are most persuasive.
The results showed the most successful persuaders bridged identity divides, were able to view another person’s perspective and utilized personal narratives and the highlighting of common ground.
“What matters is meeting people where they are. Persuasion is most likely when the would-be persuader can see the issue through the other person’s eyes,” Naunov said. “The most effective arguments bridge identity divides through personal narratives and a respectful acknowledgment of the recipient’s perspective, which lowers defensiveness and increases openness to the persuader’s viewpoint.”
The least successful approach, the researchers found, was when would-be persuaders emphasized their political knowledge and engagement, traits that can be seen as off-putting and lead to negative results. Similarly, bombarding people with facts to prove them wrong was not effective.