Modeling Pathways toward Radicalization

What conditions lead an individual or group toward committing political violence? Is it possible to accurately forecast who will become radicalized or even estimate when they might resort to violence?

These and similar questions weigh heavily on the minds of security specialists and decision makers around the world. At Los Alamos National Laboratory, scientists in the International Research and Analysis Group (IAT-1) are merging social, economic, political, cultural, and media data with social science in computer models to better anticipate terrorism and political violence. IAT-1's work aids the U.S. Department of Defense and the U.S. Intelligence Community in analyzing the indicators and warnings of political violence.

Oslo, Norway Explosion pic

On July 22, 2011, a car bomb exploded in the executive government quarter of Oslo, Norway, shown above. The explosion killed eight people and wounded several others.

What Is a Terrorist?

A terrorist is one who practices terrorism, but terrorism itself is difficult to define. "Terrorism" is an emotionally and politically charged term, thus making it difficult to provide a precise definition.

"I associate the term 'terrorism' with politically motivated violence intended to instill fear in a population so the population will then influence policy and decision makers," says Edward P. MacKerrow, director of IAT-1.

MacKerrow explains that terrorist groups have a violence target and an influence target. The violence target is the population the group attacks, but that population may or may not be considered the enemy. A suicide bomber may attack a government office because he or she actually sees the staff members inside as enemies. But in some cases, terrorists destroy civilian locations, such as restaurants and markets, without seeing the civilians there as enemies. In that situation, the terrorist group hopes the targeted civilians will, out of fear, pressure the government (the influence target) to acquiesce to the terrorist group's cause. The underlying goal of this tactic is to co-opt an unwilling population into helping the terrorist organization achieve its stated goals.

An example of an attack meant to influence policy took place in Madrid, Spain, in March 2004. Three days before Spain's general elections, an al-Qaeda–inspired terrorist cell carried out a series of bombings against Madrid's Cercanias (commuter train) system. The ruling party, Prime Minister José Maria Aznar's Partido Popular, supported the U.S. invasion of Iraq, but this was a very unpopular policy with most Spaniards. Assuming the bombings would be viewed as al-Qaeda's retaliation for Aznar's support of Bush, Blair, and Operation Iraqi Freedom, the terrorists hoped the Spanish, fearing additional attacks, would vote Aznar's party out of office.

In fact, Aznar's party did lose the election, but ironically, Spanish political analysts think the loss was caused as much by Aznar's actions after the bombings as by the bombings themselves. The Aznar government sought to obscure the connection to al-Qaeda until after the election and so misrepresented the attack as the work of the Basque separatist organization, the ETA (Euskadi Ta Askatasuna). That deception was uncovered.

Bush, Blair, Aznar photo

Prime Minister (1996–2004) Jose María Aznar of Spain (right) converses with President George W. Bush (2001–2009) and Prime Minister Tony Blair of Britain (1997–2007). Aznar's relationship with the United States and Britain, particularly his early support of the Iraq War, may have inspired an al-Qaeda sympathizer cell to execute a series of terrorist bombings against Madrid's commuter train system. –Photo by Staff Sgt. Michelle Michaud

The Making of a Terrorist

One condition that can lead to the making of a terrorist is the lack of political means to have complaints heard and addressed. Such barriers are common even in democracies, where the suspicion that one's views are being ignored can lead to political violence.

"We have lots of kids graduating college who can't find jobs. That's what happened in Cairo. That's what happened in Madrid. You don't want those kinds of riots here."
–Michael Bloomberg, mayor of New York City

For example, in early 2011, a series of riots took place in the United Kingdom protesting the government's new fiscal austerity policies. On March 26 British education secretary Michael Gove said, "Of course people will feel a sense of disquiet, in some cases anger, at what they see happening, but the difficulty we have as the government inheriting a terrible economic mess, is that we have to take steps to bring the public finances back into balance." However, the difficulty was not only in the stringent nature of the government's economic response but also in the process by which it was constructed and implemented. Various organizations, including some trade unions, argued that the British government was pushing deep spending cuts too rapidly and without consulting the public—they thought the public's opinions were not listened to and considered. Political violence was a result, and such a situation could, in turn, lead to a terrorist response: fighting "a broken system."

"If political channels are open, so everyone can participate equally in the process, the chances of a radical group having to resort to violence in order to be heard are lower," says MacKerrow. "However, if such channels limit who can participate, then those who are excluded from the process will feel great humiliation, which triggers deep emotions such as anger and hate—especially if they are rejected as a result of their group identity."

One way to assess the trajectory of a given group is to first understand why its members are together—what their core issues and concerns are. How and to whom have they communicated their issues and complaints, and how have the issues been received? Have they been acknowledged and perhaps addressed? For example, have these groups petitioned the government or marched in public events? And what have the government and the public done as a result of such activities?

"If people are not given an equal opportunity to express their opinions and issues, they can feel they are not respected because of who they are—their group identity," says MacKerrow. "An important indicator of potential violence is when we see a given group splinter into factions, where some new factions form because they feel the peaceful approach is not working and the political system is broken. Since the current political process has not worked for them, some of the factions may increase their potential for violent strategies of influence."

As an example, MacKerrow points to the recent tragedy in Norway. On July 22, 2011, Anders Behring Breivik, disguised as a police officer, opened fire on participants of a youth event sponsored by the ruling Norwegian Labour Party. Breivik killed 69 people. He was also responsible for a car-bomb explosion in Oslo earlier the same day, which killed eight people and wounded several others.

Hate Group Chart

According to the Southern Poverty Law Center, the United States has seen an alarming increase in the number of known domestic hate groups, from 602 in 2000 to 1,002 in 2010. Many of these groups seek to return to an idealized "golden age" in which ethnic relations reflected a more clearly defined dominant majority.

"When you look at Norway in terms of its socioeconomic and political systems, it is the best place in the world," says MacKerrow. "It has a very high quality-of-life index and is very democratic. And yet we have the tragic incident that just happened—a 'lone-wolf' terrorist incident. Although this type of incident is nearly impossible to predict, there were conditions before the event that could have been used to anticipate the potential for violence. For example, would it have been possible to sense that some people in Norway felt that their country was losing its identity because of too many immigrants? Could some have felt a sense of invasion?"

Indeed, Breivik did feel that way, and on the day of the attacks, he posted a 1,518-page manifesto, "2083—A European Declaration of Independence," outlining his militant anti-immigration and anti-Islam far-right ideology. But he had never hidden his views. He had been expressing them for years on Internet forums

"Radical groups are not shy about stating their goals, claims, and objectives," says MacKerrow. "These are often publicized in the open on websites and blogs and in videos. It is well known that there are some groups in Western Europe (and in the United States) who are strongly against immigration. But just because groups make such claims or are against a given policy or social condition does not mean all will pursue violence. The challenge is to know which ones will. To address this challenge, we use years of political-violence research to develop computerized models of indicators and warnings: checklists of observables to look for. As evidence is collected about these observables, the forecast indicators are adjusted."

Much like cults, terrorist groups encourage or intimidate people to abandon their families and adopt the terrorist organization as their new family. The organization can use its family-like roles and relationships to focus a great amount of peer pressure on its members, compelling them to do acts of terrorism.

In some instances, terrorist organizations target specific types of people who have characteristics that make them receptive to the organization's ideas. For example, some may look for university students who are close to dropping out and are disgruntled over their efforts to forward political or societal change.

The existence of such vulnerable people, coupled with the growth of groups that think they cannot engage in conventional societal and political processes, are the first conditions society must be aware of when attempting to identify a possibly violent or hate-driven group.

A high rate of unemployment is also a condition for creating hate groups and political violence. "I am concerned about the large numbers of unemployed recent college graduates with significant student loan debt in this country," notes MacKerrow.

Members of the American Nazi Party

Members of the American Nazi Party demonstrate outside the U.S. Capitol. Economic downfall, polarization between different identity groups, unpopular wars, and immigration issues are very important indicators for political violence.

In a New York Daily News article dated September 16, 2011, New York City Mayor Michael Bloomberg voiced his concerns that student unemployment could lead to riots if Washington fails to create more jobs, saying, "We have lots of kids graduating college who can't find jobs. That's what happened in Cairo. That's what happened in Madrid. You don't want those kinds of riots here."

In Cairo, rioting Egyptians demanded the ousting of President Hosni Mubarak. Among the grievances that led to protests and riots were economic issues, such as high unemployment. In Madrid, protesters blamed the Spanish government for spending millions on the recent visit of

Pope Benedict XVI instead of allocating more funds to address the country's widespread unemployment. In March 2011, the youth unemployment rate in Spain stood at 43.5 percent, the highest in the European Union.

MacKerrow points out that there are other, rapidly increasing conditions in the United States that could lead to violence. "Economic downfall, polarization between different identity groups, the increasing GINI index [a measure of income inequality], unpopular wars, corruption, and immigration issues are very important indicators of the potential for political violence. We have seen a huge increase in the number of known hate groups here, from 602 in 2000 to 1,002 in 2010. Many domestic terrorist groups seek to return to an idealized 'golden age' in which ethnic relations reflected a more clearly defined dominant majority."

protests in the United Kingdom pic

In early 2011, a series of protests took place in the United Kingdom, with demonstrators confronting police. Various organizations, including some unions, argued that the government was pushing deep spending cuts too rapidly, without consulting the public.

Preventing Terrorism

MacKerrow believes it is possible to prevent violence, but only if society is open to the democratic process, even for radical or extreme ideas.

"When a group perceives that a societal or political process is not addressing its needs, society must work to ensure that the group is at least heard," says MacKerrow. "Now, such a spectrum of opinion is huge, and there will be organizations that will feel they can never participate in a societal or political process on equal footing. In such cases, we must be vigilant and watch for warning signs of violence, such as their acquisition of and training with weapons. Another warning sign is estrangement. For example, has a group begun to ostracize some people and indoctrinate others, as cults do?"

MacKerrow says, "Understanding what can drive a group to violence, and possibly preventing that violence, comes down to social awareness, and to the idea that, as much as possible, society must be accepting of other cultures and customs. A lack of integration can lead to humiliation and polarization. But integration is a two-way street, an immigrating culture wishes to keep some of its customs and mores, but it should also be willing to integrate with the other cultures of that area."

Computer Modeling for Simulating the Behaviors of Terrorist Groups

To further our understanding of violent groups and how they may act, LANL scientists are turning to the computer for help. They are developing computer models to simulate terrorist behavior. But simulating human behavior is much more difficult than simulating physics problems.

One method being used for the terrorist problem, a social science problem, is known as agent-based modeling, which contrasts with the equation-based modeling commonly used in physics. Equation-based modeling simulates physical systems, either natural (like ocean currents) or man-made (like industrial machinery). The variables that affect the system (for example, material content or temperature and pressure over time) are identified and represented in sets of equations, which are then solved and their results compared with data gathered from real-world measurements.

Agent-based modeling simulates complex social systems made up of interacting "agents," which represent either individuals or groups. The researcher gathers information on how people or groups have behaved over time and encodes that information into the model's algorithms, then performs simulations to see the patterns of system behavior that might play out under possible future situations

"Computational social-science methods such as agent-based modeling are a more natural way to model social systems than is an equation-based approach," says MacKerrow. "The end result is a computer-generated ensemble of scenarios we can, in principle, analyze to try to anticipate and mitigate potential future events."

Scenario analysis is much different than prediction; that is, it can tell us something about the possibility of something happening but not give us a definitive probability—exactly how likely it is for something to happen and when it might happen.

According to MacKerrow, the difficulty lies in the lack of data from controlled experiments. "We can compensate by looking at historical events," he says. "For example, we look at the conditions of a country before a time of revolution. Perhaps a certain group blew up a building because (1) the group petitioned the government on a certain political issue and was ignored, (2) the group then marched in protest and several of its members were arrested and mistreated, and (3) the main group disbanded, but a splinter group with more-aggressive tactics emerged."

Such causal relations can be built into models for simulating a scenario with given initial conditions. Comparing the simulation results with historical events known to have involved similar conditions allows for a very crude "validation" approach. The computer simulations are run thousands of times and compared with key historical events. Counterfactual analysis must also be considered though—would those key events have occurred, and what other events might have occurred, if the historical conditions had been slightly different?

"The trouble is that the real-world history of any event is singular; so we really have only one historical scenario to benchmark against," says MacKerrow. "We do not have a set of different, controlled, system-wide social experiments to validate against. However, controlled experiments on how decisions are made under uncertainty, how reciprocity and cooperation occur, and how cross-cultural interactions occur are increasing. We're learning from experimental and behavioral economics, and that's a good start."

Experimental economics is a field that does do controlled experiments, putting humans in a controlled setting to measure social and behavioral processes, testing and challenging existing social science axioms.

A Science of Radicalization

MacKerrow and his team are working to help decision makers understand how social, economic, and psychological factors can be used to anticipate social violence and terrorism. With such knowledge, decision makers will be better equipped to assess the behavior of groups as they begin to escalate their level of aggression.

But key questions remain: Can we predict violence before it happens? Can we know who will become a terrorist and when that will happen? Social science still can't explain why one particular person would choose violence, but more people's attitudes are becoming available—for example, on the Internet—providing unique new data for researchers. The direct drivers behind violent choices will continue to be studied at LANL as new data become available.

"I firmly believe that we need to look at why some people decide to resort to violence when others do not—as much as we look at the social conditions that create the possibility for violence. That will be necessary if we're to reduce political violence in the long run," MacKerrow says. "Computer modeling of social phenomena is challenging the social sciences to come up with improved theories of human behavior. The end goal is not the computer model, but a better scientific understanding of the radicalization pathways and how to create nonviolent options for groups and individuals."

-Octavio Ramos Jr.

National 9/11 Memorial photo

The National 9/11 Memorial at Ground Zero in New York City. The World Trade Center Building One is in the background.


In this issue...

Contact Us | Careers | Bradbury Science Museum | Emergencies | Inside LANL | Maps | Site Feedback | SSL Portal | Training

Operated by Los Alamos National Security, LLC for the U.S. Department of Energy's NNSA © Copyright 2016 LANS, LLC All rights reserved | Terms of Use | Privacy Policy