March 13, 2017

What Americans Think About Refugees is More Complicated Than You Think



President Trump’s reissued travel ban does not merely restrict travel from six Muslim countries; it drastically slashes the number of refugees resettled into the U.S. The new ceiling, set at 50,000, is the single lowest cap on refugees ever set in the history of the modern refugee program, which began in 1980. The four-month pause on resettlement is longer than the pause President Bush implemented following the 9/11 terrorist attacks.

I have previously written about the disapproval of the first iteration of the travel ban. I focused on the travel restrictions rather than on the changes to refugee policy.

This piece is a close examination of public opinion polling on the question of refugee resettlement specifically. This reveals a more fragmented picture. Americans’ attitudes toward refugees aren’t overly welcoming, but neither are they overly hostile.

The most accurate way to summarize public opinion on refugees is: it’s complicated.

The Bad

Looking back at polls from the last few years, whiz kid Harry Enten from FiveThirtyEight, concluded that “Americans seem OK with lowering the number of refugees accepted by the U.S., but outright bans are not likely to be popular.”

Enten is right that Americans have historically supported reductions in refugee resettlement. Old Gallup polls purport to track public support for refugees (more on this later) and the results do not portray Americans in most humanitarian light. They show that recent minority support for accepting 10,000 Syrian refugees was higher than the support for a plan to accept Jewish children fleeing the Nazis years ago.

This historic anti-refugee bias of public opinion should give us immediate pause in how good a guide public support is in the first place, especially considering all of the economic, national security, and humanitarian reasons in favor of resettlement.

However, untangling public opinion polling complicates the narrative, revealing a messier story than appears at first glance.

And the Ugly

The survey results below show much more variation than the simple summary we so often hear and repeated above—that Americans reject refugees. In fact, most polls show net support for Syrian refugee resettlement. Even if we did want to rely heavily on the polling in crafting policy, it’s not clear which polls we should use or how we ought to interpret the conflicting answers.

polls1

Some polling questions, like the Brookings poll which shows double digit net support for refugees, give basic and unbiased information about the refugee program, which we should care about if we are interested in informed opinion.

For instance, polls can give the definition of what a refugee is or describe the process by which refugees are admitted, and this apparently increases support for resettlement.

Other polls, however, posit hypothetical policies that were never seriously proposed as described, like the Bloomberg poll that brings up the idea of a religious test for refugees. This decreases overall support reported by the poll.

polls2

This polling done on specific policy proposals is even harder to draw conclusions from because it is ambiguous what opposition to the policy represents. A person may respond that he opposes admitting 10,000 Syrian refugees because he thinks the number is too high, or because he thinks it too low. This problem plagues all of the Gallup polling which purports to show consistent historical opposition to refugees.

And, without sufficient background knowledge of the relative size of proposed admissions and the refugee program as a whole, respondents are more likely to report that they are opposed to the plan. Not to mention, the mere association of a policy with the President who proposed it can skew the results.

For all the ambiguity, we see widespread reporting on these problematic polls as if they were clear-cut and definitive.

Still other polls were widely reported on but should be dismissed entirely since the survey questions contained factually incorrect information. For instance, a 2016 Rasmussen poll supposedly showed a rejection of Obama’s proposed policies but falsely claimed in the question that Obama wanted to bring in 110,000 refugees in one year from Africa and the Middle East aloneleading to an unsurprising reduction in the support for refugees.

Other polls, like the 2016 GlobeScan poll, used unconventional methods and generous interpretation, exaggerating American support for refugees.

All this is to say, polling reveals a complicated picture. It is inaccurate to say Americans have been overwhelmingly open to refugees but it is also inaccurate to say they have been consistently opposed to allowing refugees.

The oversimplification of public opinion gives ideologues on either side the chance to say their case is on the side of the “the people” or alternatively, gives some cause to dismiss the public altogether. But the truth is far messier, and public opinion often trails elite opinion rather than vice versa.

Ultimately, polling is at best an imperfect guide to public opinion, which remains elusive. More honestly, it is a distraction, when we ought to judge policy on its merits and not on transient and indeterminate popularity.