How (Not) to Change Someone’s Mind

But there are two tactics that really seem to work: encouraging people to engage in perspective-taking—putting themselves in someone else’s shoes—and asking people to come up with reasons to support something they oppose, or vice versa. Psychologists call this second strategy “counter-attitudinal argument generation.”

Research suggests that each of these persuasive tactics can be individually effective. So what would happen if you combined them by asking people to generate arguments from someone else’s point of view? Would it be a case of “more is better” or “too much of a good thing”?

That’s the question Kellogg marketing professor Derek Rucker and his coauthors, Rhia Catapano and Zakary Tormala of Stanford University, posed in a new paper—and it is one they puzzled over. “We all saw the value in each hypothesis,” he says. “That made it a very exciting project to pursue. Going in, it was not clear what the outcome would be.”

And the outcome was clear: combining perspective-taking and argument generation ultimately backfired.

“In this particular case, one plus one does not equal two,” Rucker says. “What this study shows is that some of our intuitions—as well as what we might infer from prior research—about what might be most effective at persuading others might be wrong.”

The researchers also pinpointed the reason why the two tactics do not work well together.

When you are asked to take someone else’s point of view and generate arguments from it, the ideas you come up with don’t align with your own values. In fact, the process only serves to highlight differences in values between you and the person whose perspective you are taking. Rather than being swayed by the arguments you have developed, the researchers found, you feel distant from them and thus you are less persuaded.

More Isn’t Necessarily Better When It Comes to Changing Someone’s Mind

To conduct the research, the team turned to the online discussion website Reddit, a popular forum to swap memes and analyze pop culture. Because of its large group of politically engaged users, it was also an ideal place to start a study of persuasion.

The researchers recruited 484 Reddit users who had been active in political discussions on the site. The participants were told they would discuss the controversial issue of universal health care (UHC) with another Reddit user.

First, the Reddit recruits were asked to provide demographic information and to rank their political views from extremely liberal to extremely conservative on a scale from one to seven. Then, they were asked to rate from zero to a hundred how much they agreed with implementing UHC.

Next, the researchers told the participants about the person they would debate: a 22-year-old white man from Ohio with the opposing political ideology. (The participants didn’t know that the debate would not actually take place, or that the debate partner was a fictitious creation of the researchers.)

To prepare for the debate, participants were asked to do one of two things: in the control (argument generation only) condition, pro-UHC participants were asked to think of reasons why someone might oppose UHC, while anti-UHC participants were asked to do the opposite.

In the experimental condition, designed to assess the combined effects of argument generation and perspective-taking, participants were instructed to put themselves in the shoes of their debate partner, “visualiz[ing] clearly and vividly what this individual’s life and experiences may be like, and how they may feel,” and to come up with reasons why that person might hold their position on UHC.

Finally, participants were asked once again how much they agreed with UHC on a scale of zero to one hundred. They also rated from one to seven how strongly the arguments they had generated aligned with their values and how receptive they felt to those arguments.

When the researchers looked at the results, several trends emerged. First, it was clear that combining perspective-taking with argument generation was significantly less powerful than argument generation alone. In the control (argument generation only) condition, participants’ attitudes shifted a little, an average of 1.52 points on the 100-point scale. In the experimental condition, the average shift was just .39.

The results of the experiment also provide insight into why perspective-taking did not work: a lack of what the researchers call “value congruence”—that is, alignment with one’s own worldview. The control group reported generating arguments that were somewhat consistent with their values, an average of 3.53 out of 7. That number dipped to 3.27 for participants in the experimental condition.

The Role of Shared Values in Persuading People

The researchers wanted to understand more clearly why combining argument generation and perspective-taking was not an effective persuasion tactic.

If value congruence was the key driver, the researchers hypothesized, the effect should disappear if people are asked to generate arguments from the perspective of someone with whom they disagree on a topic but who is otherwise generally similar to them.

So they set about recruiting a new group of 1,200 online participants, who did not come from Reddit. The researchers repeated the same experiment they had run before, with a few tweaks. This time, participants were asked to consider universal basic income (UBI). They also left out the premise of engaging in a debate.

Additionally, they created two different experimental conditions. In the first, participants were asked to generate arguments from the perspective of people who disagreed with them on UBI and had a different political ideology from them. In the second, participants were asked to generate arguments from the perspective of people with their same political ideology who happened to disagree with them on the specific issue of UBI.

Just as they had seen the first time around, the researchers discovered that the argument generation alone was more effective in shifting people’s attitudes (an average change of 5.51 on a 100-point scale) than generating arguments from the perspective of someone with a different political ideology (an average change of just 1.72).

But, consistently with their proposal that value congruence mattered, this effect did not hold when participants generated arguments from the perspective of someone with the same ideology as them.

It turned out that generating arguments from the perspective of someone similar was extremely effective in changing minds—much more so than the researchers expected. In fact, participants in the perspective-taking/same ideology group saw much greater levels of attitude change, receptiveness, and value congruence than did the control group.

Rucker notes that they have not pinned down the mechanism for people’s greater attitude change in this condition. However, he says he and his coauthors have a few ideas. For instance, it may be because we tend to dislike conflict within our ideological peer group. “When we have tensions in belief, we want to restore balance, and one way to do that is to change our own beliefs,” Rucker says. “So, rather than argue with those who are like me, I simply adjust my ideas towards the group.”

Altogether, the results of the two experiments told a persuasive story: “In the process of generating arguments from [a political opponent’s] perspective, you start realizing that they have different values from you,” Rucker explains, and those arguments thus start to seem a lot less appealing.

For Rucker, there is an important takeaway: “There’s a lot of research that looks at individual persuasion strategies. This paper says you can’t just think about individual persuasion strategies. You have to think about how they might interact with each other.”