Sister Munira Salim Abdalla, chief administrator for the Islamic Ummah of Fredericksburg, asks a law enforcement official to intervene during a heated dispute in a public meeting regarding the possible construction of a new Islamic Center of Fredericksburg site in Spotsylvania County, Va., on Nov. 17. (Peter Cihelka/Free Lance-Star via AP)

Since the Nov. 13 attacks on Paris, many U.S. leaders have unleashed discriminatory rhetoric in the name of counterterrorism. Thirty-one governors said that Syrian refugees were not welcome in their states. Jeb Bush suggested that refugees should be allowed into the United States if “you can prove you’re a Christian.” Donald Trump said that he would “strongly consider” shutting down American mosques and that he wants “surveillance of certain mosques if that’s okay.” Claiming “there is no meaningful risk of Christians committing acts of terror,” Ted Cruz criticized the Obama administration “because they pretend as if there is no religious aspect to this.”

Of course, most religions — including Christianity — have been manipulated to inspire or excuse terrorism. And people of most nationalities — including Americans — have committed terrorist acts. So the critical question is not whether the Islamic State has a national or religious aspect; a great deal of terrorism does. The question is: Will policies that discriminate against people on the basis of nationality or religion help or hurt efforts to counter the Islamic State’s terrorist threat?

Many excellent scholars — both before and since 9/11 — have produced research that tells us about the relationship between discrimination and counterterrorism.

Here’s what we know. To be most effective, counterterrorism policies need to make an explicit distinction between the individuals who genuinely threaten others with terrorism, on the one hand, and on the other, the broader populations those terrorists claim to represent. Counterterrorism efforts — especially using force — should narrowly target only the former, as much as possible.

Groups that commit terrorism often hope to provoke a violent overreaction against the community they claim to be defending. Even though most people in that community are nonviolent, such a reaction might force them to turn to the terrorist group for their own defense, swelling its ranks and realizing its ambition for greater political power.

In other words, counterterrorism policies that discriminate against Syrians or Muslims in general are precisely what the Islamic State wants. Such an approach will, quite literally, help the terrorists win.

How do we decide what makes someone a terrorist?

When we see people acting badly, our first instinct is to blame their bad behavior on something that has nothing to do with us, so that we don’t look bad by association. Due to what psychological studies term the “ultimate attribution error,” we blame the bad behavior of people we see as “others” on personal characteristics that make them intrinsically different from ourselves. In contrast, we blame the bad behavior of people who are like “us” on external pressures and constraints (rather than on traits that we share).

This explains why many non-Muslim Americans instinctively decide Islam is the reason the Islamic State attacked Paris, but would never attribute the Oklahoma City bombing to the fact that Timothy McVeigh was American. Moreover, this attribution error is especially likely after intense conflicts and when emotions run high — such as after a terrorist attack.

Blaming terrorism on people unlike ourselves is psychologically easy. But it’s also dangerous.

Here’s how. First, since most people in every major demographic category are not terrorists, attributing terrorism to such a broadly shared characteristic gets the causes of terrorism wrong. Policies based on that way of thinking won’t work effectively, because they’re not aiming at the actual problem.

Second, we waste time and energy when we assume that far more people are dangerous than actually are. Drawing a crass link between terrorism and all Syrians or all Muslims, for example, implies that we’re in danger from every single one of some 17 million Syrians or 1.6 billion Muslims (including more than 150,000 Americans of Syrian descent and more than 2.7 million Muslim Americans). That’s absurd. Needless to say, the overwhelming majority of both Syrians and Muslims are not terrorists, just like the overwhelming majority of any other nationality or religion.

Third, we falsely conclude that others like us can’t commit terrorism because they don’t share the trait that “causes” it. This can lead us to ignore individuals or groups who may actually be quite dangerous, just because they look or speak in a familiar way.

How to help terrorists: Assume that everyone in their ethnic or religious group is a terrorist, too

Here’s the larger problem. When we fall for the universal attribution error and target the whole group that the terrorists (falsely) claim to represent, we hurt our own society and help the terrorists.

Many terrorist groups have claimed to represent a religious or ethnic population when most of those people actually rejected their violence. But in cases where the terrorists succeeded in provoking a discriminatory backlash against the population at large, those nonviolent people were forced to turn to the terrorists for protection. In other words, a xenophobic reaction to terrorism can actually help the terrorists’ recruiting and fundraising, allowing them to grow far more powerful than they could have without it.

For example, the Tamil Tigers were only a small group when they attacked a Sri Lankan army patrol on July 23, 1983. But that attack sparked “Black July,” a wave of anti-Tamil riots that quickly spread across the country and left 400 Tamils dead and more than 100,000 homeless. The Sinhalese majority blamed all Tamils for the Tamil Tigers’ violence. So where did Tamils turn for protection? To the Tigers, whose recruitment boomed, starting a Sri Lankan civil war that would last for the next 26 years.

Government counterterrorism policies that aim too broadly can bring on the same result. The British combated the Provisional Irish Republican Army during the 1970s and 1980s with policies that targeted far more Catholics in Northern Ireland than those actually in the terrorist group. For instance, the British government armed Protestant-dominated local police and paramilitaries, held without trial and aggressively interrogated numerous Catholics whom their local communities knew full well had no connection to the PIRA, and used trials in which judges could convict someone with only a confession or the testimony of a former accomplice. As happened with the Tamils and the Tamil Tigers, this heavy-handed reaction drove many Catholics in Northern Ireland — including many who would otherwise have rejected terrorist violence — toward the PIRA as their only possible defense.

The right kind of discrimination: Discriminating between terrorists and non-terrorists

Counterterrorism policies need to be discriminating to be effective, but in exactly the opposite way from what’s been suggested by the rhetoric coming from some of our leaders. Rather than discriminating against the broader population a terrorist group wants to represent, counterterrorism policies need to effectively discriminate between those actually responsible for the terrorist threat and the vast majority who prefer nonviolence and pose no real threat.

Terrorism is not the product of a particular ethnicity or religion. Instead, research has shown that a group’s use of terrorism has more to do with institutional and social psychological factors or a misplaced belief it will effectively promote its goals. But terrorist groups only thrive over the long term if they can command enough community support.

As a result, the most effective counterterrorism policies apply force in a highly discriminating way, specifically targeting individuals known to pose a threat, and leaving the broader population free and peacefully integrated into society. Doing so drives the terrorists apart from their desired constituency instead of driving them together. As Audrey Cronin’s study of counterterrorism strategies concludes, “Marginalization from their constituency is the death-knell for modern groups.”

In contrast, broad-brush counterterrorism policies are both ineffective and dangerous. The United States tried such policies after 9/11, using a range of ethnic profiling strategies that, as David Cole and Jules Lobel observe, failed to convict any terrorists. Worse than simply wasting resources, though, xenophobic overreactions generate support for the terrorists. Robert Art and Louise Richardson conclude their edited volume on democracy and counterterrorism by noting that “terrorist strategists have long been aware of the value of provoking governmental overreactions that play back into the terrorists’ hands” and that the “indiscriminate use of force” almost always backfires.

So how do we apply this knowledge to fighting the Islamic State?

The Islamic State claims to have established an Islamic caliphate, but it continues to be condemned by Muslim countries, Muslim scholars and many other terrorist groups claiming to represent Islam.

The refugee exodus from Syria confirms that the Islamic State rules by fear. Its best chance to stabilize its rule is to make its population believe they have no alternative by getting other countries to close their doors to them. Broad-brush discrimination against Syrians and Muslims reinforces the Islamic State’s narrative regarding a war between Islam and the West. By contrast, when Western countries welcome Syrian refugees, they undermine that narrative.

This strategy of compassion will help defeat the group. With fewer subjects, the Islamic State will lose opportunities for the extortion and taxation that are its primary source of revenue.

In fact, the Islamic State may have attacked Paris precisely because it realized its own weakness, strategically lashing out in the hope that Europe would close its doors to refugees and force them to stay under Islamic State control. As Seth Jones wrote recently, “The history of insurgent groups is sobering: Most increase terrorist activity when they lose territory.”

President George W. Bush understood that Islam was not the cause of al-Qaeda’s terrorism. On Sept. 20, 2001, he labeled al-Qaeda “a fringe movement that perverts the peaceful teachings of Islam,” declaring, “The enemy of America is not our many Muslim friends. … Our enemy is a radical network of terrorists.” President Obama also understands this, which is why he said, using another name for the Islamic State, in no uncertain terms, “ISIL is not ‘Islamic.’… The vast majority of ISIL’s victims have been Muslim. … ISIL is a terrorist organization, pure and simple.”

Both presidents are concerned for Americans’ safety — but both know that there is no safety in fear and hatred that discriminates broadly rather than accurately. Governors and future presidents can learn from them.

Richard W. Maass is assistant professor of political science at the University of Evansville. His current book project examines domestic constraints on great-power annexation, and he has published research in the journals Diplomatic HistoryTerrorism and Political ViolenceInternational Security and Historical Methods. He is on Twitter @richardmaass.