A Syrian woman gestures through a plastic cover of her tent window at an informal refugee camp in the eastern Lebanese town of Marj, near the border with Syria, on  Jan. 28. (Hussein Malla/AP)

Since taking office, President Trump has signed two executive orders to prevent Syrian refugees from entering the United States, in keeping with his campaign call for a “total and complete shutdown of Muslims entering the United States.” Meanwhile, public figures ranging from Pope Francis to the United Nations High Commissioner for Refugees (UNHCR) have asked countries around the world to open their doors to refugees, particularly the millions of Syrians who have fled one of the most significant humanitarian crises of our time.

Refugee advocates have used a variety of strategies to advocate for refugees — relying on apps, videos, lesson plans and op-eds that draw on both emotion and logic. So which of the appeals actually influence Americans’ response toward refugees?

In mid-2016, we set out to answer this question, testing two common persuasion approaches:

  1. Empathy, by asking citizens to imagine life as a refugee.
  2. Information about how many refugees the United States has agreed to accept relative to other countries.

In a new paper, we find that encouraging empathy — imagining oneself in the shoes of a refugee — prompts Americans to support bringing in refugees. By contrast, detailing the United States’ relative commitment to accept refugees, a commonly used approach, has no effect — and may even prompt a backlash among Republicans.

Here’s how we did our research

In the weeks leading up to the 2016 U.S. presidential election, we conducted a nationally representative survey of 5,400 American citizens, implemented by YouGov, a market research firm. We randomly assigned respondents to one of three groups: to take part in an empathy exercise; to receive information about how many refugees the United States committed to accept compared to Germany, France and Canada; or neither.

Our empathy message borrowed a lesson plan from the Pulitzer Center. Here, we asked respondents to provide a written response to the following questions:

Imagine that you are a refugee fleeing persecution in a war-torn country. What would you take with you? Where would you flee to? What do you feel would be the biggest challenge for you?

To measure attitudes toward refugees, we asked all respondents to rate refugee profiles, assessing whether the refugee should be admitted to the United States. To do this, we showed respondents pairs of hypothetical refugees, with each profile randomly assigned characteristics such as religion, gender, age, occupation and degree of fluency in English. We then asked respondents how willing they would be to allow each refugee into the United States, and which of the two they preferred. Each respondent saw and rated three pairs of refugee profiles.

At the end of the survey, we asked respondents if they would like to write an anonymous note in support of refugees to the next U.S. president, who had yet to be elected at the time of the survey. We consider letter writing in support of refugees a behavioral measure of refugee inclusion, since it required more time and cognitive effort than the typical survey question. We also told respondents, truthfully, that we would compile all anonymous comments and send them to the new president.

What did we learn?

Empathy works

First, we find that our empathy exercise increased the likelihood that a respondent would write a letter to the president in support of refugees, although it did not affect refugee ratings.

Overall, letter-writing rose from 17 to 21 percent between the control and empathy treatment group, a statistically significant jump of four percentage points. The effect was largest among Democrats, where letter-writing increased from 23 percent to 34 percent. There was a smaller but still positive effect among Republicans, from 5 to 7 percent. While independents’ likelihood of letter-writing did not change, they did rate refugees more positively after the empathy exercise.

We believe the empathy message nudged Democrats to act on their preexisting supportive attitudes toward Syrian refugees. Empathy had a smaller but still important effect among Republicans and independents, who had less favorable attitudes toward refugees to begin with.

By contrast, information about the United States’ relative commitment to accept refugees didn’t boost the letter-writing rate at all — and may have prompted a backlash among Republicans. Republican respondents in this group gave the refugees ratings that were lower than those from the control group. What’s more, when asked to describe our survey to a friend, Republicans in the information group were more likely to describe our survey negatively.

While we did not expect backlash, and did not design the survey to explicitly test for it, refugee advocates may wish to consider the possibility that this relatively common narrative may bring a negative reaction.

Americans prefer refugees who are high-skilled, Christian, male and fluent in English

When we analyzed the refugee profile ratings, we found that Americans clearly prefer a particular type of refugee. The highest-rated profiles described the refugee as Christian rather than Muslim, male rather than female, professional rather than a low-skilled worker, and a fluent English speaker rather than a non-English speaker. This consensus cuts across party lines. Although Democrats rated the refugees more highly on average than did Republicans, they also preferred high-skilled Christian men who speak fluent English.

These preferences are consistent with previous research on attitudes toward immigrants, which suggests that the American public does not meaningfully distinguish between immigrants and refugees, and that there is a broad consensus about the type of outsiders Americans prefer.

Of course, our research is limited. In an online survey with experiments, we could not present all the ways in which Americans usually encounter information and arguments about refugee policies. But often, Americans encounter these arguments online. Further, our experimental interventions borrow directly from real-world efforts, which increases our confidence in the validity of our findings beyond the survey experiment.

Encouraging Americans to “walk a mile” in the shoes of a refugee may take advocates a long way.

Claire Adida is associate professor of political science at University of California at San Diego. Follow her @ClaireAdida.

Adeline Lo is postdoctoral research associate in the department of politics at Princeton University. Follow her @adelineylo.

Melina Platas is assistant professor of political science at New York University Abu Dhabi. Follow her @melinaplatas.