Skip to main content
ABC News
Past Terrorist Attacks Helped Trump Capitalize On Anti-Muslim Sentiment

Before President Obama made a statement on the terrorist attacks in Brussels today, presidential contender Donald Trump had already weighed in: first on Twitter and then on Fox News and NBC, which interviewed him by phone.1

Trump would “close up our borders” in response to Brussels, he said in the Fox News interview, adding that the U.S. would have to be “very, very vigilant as to who we allow into this country.” The comments echoed proposals issued by Trump after previous acts of terror. Following the series of terrorist attacks in Paris on Nov. 13, Trump appeared to propose (and then partly pulled back from) a national database to register Muslims. Then, after the attacks in San Bernardino, California, on Dec. 2, Trump called for a “total and complete shutdown of Muslims entering the United States.”

We might expect more rhetoric along these lines from Trump because the Paris and San Bernardino attacks appeared to boost his standing in national opinion polls, as well as the amount of media attention he received. And Trump’s rhetoric on terrorism will likely continue to focus heavily on immigration — according to exit polls, Republican voters aren’t particularly keen on Trump’s crisis-management skills, but most favor his ban on Muslim immigrants, and immigration overall is one of the top draws for Trump voters.

Below, you’ll find two versions of the FiveThirtyEight national polling average as it ran from June 16 last year, the day that Trump entered the presidential race, through Nov. 12, the day before the Paris attacks. One version (the gray line) is essentially the same as our regular national polling average,2 which is deliberately designed to be conservative and slow-moving. The other (the red line) is a “high sensitivity” version that places a much greater premium on poll recency; it’s noisier but is probably better able to capture the public response to developing news events.


As you can see, Trump’s national polls had stagnated in the mid-to-high 20s in the two months before Paris; Ben Carson was approaching him in national surveys and pulling slightly ahead of Trump in Iowa. Meanwhile, the news cycle had become much less Trump-obsessed than usual. According to a forthcoming study we’ll be publishing of headlines at, the political news aggregator, Trump was the lead political story on only three of the 50 days before the Paris attacks. Google search interest in Trump was also considerably down from its peaks in August and September.

But the Paris and San Bernardino attacks were associated with an uptick in Trump’s numbers. According to our high-sensitivity polling average, Trump improved from 28 percent of the vote just before the Paris attacks to 32 percent on Dec. 1, the day before the San Bernardino attacks. His numbers then rose further, to about 35 percent by mid-December.


It’s possible the timing of Trump’s polling bounce was coincidental, but that seems unlikely given his focus on both Islamic terrorism and immigration since the start of his campaign. Furthermore, this period was associated with a sharp rise in news coverage for Trump, which has tended to both reflect and reinforce his gains in the polls. Following the relatively limited coverage of Trump in October and November, he was the lead news story for six straight days on Memeorandum after he first floated the Muslim ban on Dec. 7. Google searches for Trump more than doubled after he proposed the ban.

Exit polls in states that have voted so far suggest that immigration and anti-Muslim sentiment may be bigger factors in Trump’s gains than his ability to handle terrorism itself. On average in states where exit polls have been conducted, 38 percent of voters who said terrorism was their top issue voted for Trump — the same as his overall level of support in these states. But 56 percent of voters who listed immigration as their top issue went for Trump.

Alabama 43% 44% 61%
Arkansas 33 32 53
Florida 46 47 60
Georgia 39 41 57
Illinois 39 40 69
Iowa 24 21 44
Massachusetts 49 50 74
Michigan 37 39 62
Mississippi 47 43 56
Missouri 41 39 63
Nevada 46 36 62
New Hampshire 35 29 53
North Carolina 40 36 60
Ohio 36 34 68
Oklahoma 28 34 29
South Carolina 33 31 51
Tennessee 39 44 50
Texas 27 34 35
Vermont 33 43 70
Virginia 35 34 42
Trump supporters care more about immigration than terrorism

Results for immigration voters in Oklahoma and Vermont are inferred based on answers in other categories

Source: National Election Pool Exit Polls

In some states, the exit polls have also asked two other relevant questions: First, whether voters favor “temporarily banning Muslims who are not U.S. citizens from entering the U.S.,” and second, which candidate they think would best “handle an international crisis.”3

Trump has performed well on the Muslim ban question, winning close to half of the roughly two-thirds of Republicans who agree with the ban, while underperforming with those who don’t. But Republicans may have concerns about whether Trump passes the commander-in-chief test: In all eight states where the exit polls asked the question, fewer Republicans said they preferred Trump in an international crisis than voted for him overall.

Alabama 43% 40% 52
Arkansas 33 30 38
Georgia 39 36 45
New Hampshire 35 30 45
South Carolina 33 27 41
Tennessee 39 35 45
Texas 27 26 37
Virginia 35 31 44
GOP voters aren’t into Trump for his crisis-handling skills

Source: National Election Pool Exit Polls

So if the past is any guide, Trump will escalate his attacks on Muslims and immigrants in the coming days. His opponents, especially Hillary Clinton, will press him on his steadiness in office and his fitness to handle a crisis. After a couple of days of coverage from Brussels, the news coverage will revert to being all about Trump.

Listen to the latest episode of the FiveThirtyEight politics podcast.



  1. NBC also conducted a telephone interview with Democratic front-runner Hillary Clinton.

  2. A very minor difference is that we re-ran the numbers to include polls conducted as of the date listed in the chart, even if they hadn’t been publicly released by then. For example, a poll conducted from Oct. 13 to Oct. 15 but released to the public on Oct. 18 would ordinarily not appear on our chart until Oct. 18. In these versions, however, it’s included in the average beginning on Oct. 15.

  3. In the table below, I only include states where both questions were asked, instead of just one or the other.

Nate Silver is the founder and editor in chief of FiveThirtyEight.