Terrorism Research Before and After 9/11

Gary LaFree It's a real pleasure to be here and I want to thank John for the invitation; it's a real honor, in fact, to recognize this important anniversary.

I want to start by acknowledging, actually, the research support we've received actually from the Department of Homeland Security Science and Technology, University Programs, Human Factors at DHS, and also the National Institute of Justice. A lot of the work I'm going to be talking about this morning was underwritten by one or all of these organizations. The real topic, though, and to kick this off and think about the social science role, is that, if you think about it, on the day after 9/11, we didn't know as nearly much about terrorism and its constructs and its causes as you might think we should have. We had no comprehensive global terrorism database at that point in time, and for some really basically bureaucratic reasons. The U.S. State Department was always in charge of collecting terrorism data for the United States, and because of the way it was set up, it never collected data on domestic terrorism. Now that we've had the ability to collect domestic terrorism data, we realize that basically it's about nine times more common than transnational. So none of that was being tracked by anybody. We really didn't have very good information also on terrorist organizations themselves. I'm sure there was quite a bit of classified information on individual organizations, but not the kind of scientific, empirical work that has been done by criminologists for example on gangs. And we also knew very little, we had very little systematic data, on what governments were doing to counter terrorism and how successful these various efforts were. In fact, these three are sort of rank ordered from the best to the worst in terms of what we did know the day after 9/11.

And just to pause for a second about this and think about, here you are trying to fight terrorism, and you don't really know how much of it there is or what its characteristics are. You sort of imagine trying to fight cancer when you don't know how much cancer there is, or trying to deal with unemployment when you don't know how much unemployment there is.

My favorite quote to sum up where we were the day after 9/11, or maybe the harshest, comes from a British guy, who says the following: "Terrorism research exists on a diet of fast food research: Quick, cheap, ready-to-hand and nutritionally dubious." So, it was a pretty grim situation in many ways, and even though there were quite a few people in the social sciences that had begun to look at the area.

So, I want to start really by complimenting NIJ, because NIJ came to our rescue back in 2002, when Laura Dugan and I first stumbled onto a very large uncomputerized database on terrorism. And in fact, NIJ, early on, ended up funding a variety of very important projects in the social sciences involving terrorism. They funded projects looking at terrorism databases, at improving criminal justice's responses to terrorism, at assessing potential high-risk terrorism targets, links between terrorism and other crimes, and terrorist organizations' structure and culture. And the room's too big for me to tell easily, but I'm guessing some of the people listed on this slide are probably present in the room today, and if they are, I'll thank them if I refer to any of their research in a second.

So one of the outcomes for us, because of this beginning NIJ grant, is then we were able to successfully develop a Center of Excellence proposal for the Department of Homeland Security, and in 2004 we were able to start a center that was specifically studying the social and behavioral science understanding of terrorism, which is abbreviated as the START Center and which I have been directing ever since. The START Center has basically looked at the life cycle of terrorism. So we look at how individuals join terrorist groups from a psychological standpoint, from a sociological standpoint; we look at group creation; once groups have formed, we look at how they evolve, we look at their trajectories over time; we look at why some groups disappear rapidly and other groups seem to get more dangerous over time; and we also look at, if there is a terrorist strike, what's the best way for government to get back on its feet as rapidly as possible, to communicate risk to the public and have the public believe that communication, and so on.

So what I would like to do with the few minutes I've got this morning is talk about those three areas where we were not in very good shape the day after 9/11, and give you some examples of how far we have come in the last 10 years. I'll start with building a global terrorism database.

The approach that we took, that my colleague Laura Dugan and now probably hundreds of students have been working on this in the past 10 years, is to have built an unclassified database on all terrorist events around the planet, going back to 1970. It uses—in the old days we used to use newspapers, we used to use hard copies, we used to use news services; increasingly we use the internet to keep this thing up to date. In about a week, actually, we're going to be releasing the most recent version of it. It's actually online; you can sort it, you can look for specific examples and so on.

And this is providing a wealth of social science information about terrorism, including being able to track rates over time. When we were first able to produce the table in front of you, for example, the solid blue line tracks terrorism going back to 1970, we found that actually, contrary to what a lot of people might have been thinking on 9/11, terrorism actually peaked in the late 1980s, early 1990s; with the collapse of the Soviet Union, terrorist attacks around the world declined rapidly, partly because many of the groups using terrorism at that point had Marxist/Leninist-type objectives. And then it was actually at a pretty low point right before 9/11, and now has gone back up pretty dramatically in the last few years. We're getting ready to add the last two years' data, which is going to push these lines much further. In the old days, people used to talk about terrorists wanting a lot of people watching, not a lot of people dead; you can see the dashed line under the solid line is fatalities relating to terrorism, and you can see those lines are pretty far apart during much of the 1980s and 1990s. They're much closer together now. So in other words we're seeing fewer attacks but more casualties, so that some people say terrorists not only want a lot of people watching these days, but a lot of people dead.

We can also do much more specific examples now of how, of what the terrorist threat really looks like. I just brought one case that was published in Criminology and Public Policy last year that I worked on with Sue-Ming Yang and Martha Crenshaw. What we're showing here is 53 foreign terrorist organizations that were developed by the U.S. State Department as posing the most serious threats to the United States. So we looked at just what these 53 groups were doing, where they were attacking and so on. On the left side of the chart, you see attacks by these 53 groups on U.S. targets. These are not attacks just on the U.S. homeland, but they're on U.S. targets anywhere in the world, so it could be U.S. embassies, businesses and so on. And what was so striking about these findings—it's the first time we were able to generate them—is how few of these attacks were actually aimed at the U.S. So essentially what you've got is groups in Pakistan, for example, that are predominately attacking Pakistani targets, and predominately killing Pakistani citizens, in a ratio of something like 10 to 1. And we found this quite amazing, actually, that even groups—and these are not a random sample of terrorist groups, but groups that specifically are interested in doing the U.S. harm, nevertheless most of the time they attack close to home and they attack local targets.

We also were able to look at terrorism attacks by these groups over time, using a methodology very familiar to criminologists, called trajectory analysis. I don't know if Dan Nagin is in the room today, but I'm sure many of the people in the room use trajectory analysis, and it seems to work pretty well when we apply it to the study of terrorist groups. And we find that attacks against the United States break pretty neatly into three waves. You see the green line, where you have attacks mostly in the 1970s, and these groups pretty much disappear by the 1980s. You see a purple line where the attacks are mostly from the 1980s and early 1990s and then disappear. Then on the far right you see a red line that's just starting to get some traction, and if we extended these data to 2010, which we're in the process of doing, you'd see that red line being much, much bigger. I also want to point out, though, where you get these three fairly neat waves, you can also see, hugging the bottom of the x-axis, is a blue line. We call this group "sporadic"; they're sort of the one-hit wonders, and they're a very interesting group because they're very lethal, they cause more casualties than any of the other groups except for the 21st-century group, in the red, and they're very hard for us to say much about statistically, because they go away rapidly. They only have one or two strikes before they go away.

Just to give you a sense of the underlying groups behind these three waves, and then the sporadic group, here are some of the key groups involved. I just picked two from each case. So, examples of the 21st-century boom: I think it'll be little surprising that al-Qaeda and the Taliban are the two most important groups. From the 1980s, the two most important groups were the Shining Path and the Farabundo Marti National Liberation Front. From the '70s, we have the Red Brigades and the People's Liberation Forces. And two examples of the sporadic groups that are not as commonly known to everybody but that are also quite dangerous are the Popular Liberation Army and the Moro Islamic Liberation Front.

So I would argue we have learned a whole about the basic structure of terrorism around the world in the last 10 years. We also have made a lot of progress in terms of understanding terrorist organizations: how they work, how they function, what makes them lethal. And I just have time for a couple of examples.

I'm going to start out with a spider-looking construction, which is done by two of our researchers at SUNY Albany, Asal and Rethmeyer, and what they've done is used open sources to track all of the connections between terrorist organizations over time. That big spider-looking thing in the upper left-hand corner is al-Qaeda. And I don't have time to go into huge detail—and by the way, all of this stuff, a lot of this is on our website, if you'd like more information, or if you send me an email I'd be happy to share it—but one of the key things we're finding about networks is they're a little bit like business organizations. The better networked organizations are, the more likely they are to target the United States, and the more likely they are to have lots of fatalities.

And just to cut to the chase, we've found that 11 percent of terrorist groups are responsible for 70 percent of fatalities, and they are the ones that are best networked. We're actually finding a very interesting parallel for just groups in the United States: Better networking seems to work in terrorism just like it does in the business world.

We also can take this kind of information and start to do some predictions about which groups turn to terror. Fortunately, many groups with grievances pursue those grievances in nonviolent and legitimate ways, but unfortunately some do not. This project, which is headed by John Wilkenfeld at the University of Maryland, looked at organizations around the world and why they turn to terror. They found that organizations with a democratic ideology are less likely to turn to terrorism. They also found the four variables that are listed on your screen are important in terms of predicting whether a group will turn to violence. If they have a separatist cause, so in other words they're trying to get their own territory; if they use rhetoric justifying violence—and I found this remarkable, based on criminology. You know the old Jack Katz book: In a bank robbery, the bank robber always declares that this is a robbery, as if you didn't know that it was a robbery. In every single case he looked at, somebody says, "This is a robbery." It turns out that terrorist organizations are like that too: If they tell you they're going to use violence, by gosh they tend to use violence. We also found if they have foreign support, I mean this explains in large part why a huge drop-off in certain kinds of terrorism with the collapse of the Soviet Union, and also state repression tends to move groups toward violence. You take all these factors into account, and you take groups with a democratic ideology, there's an 89 percent likelihood of engaging in terrorism when all of those characteristics are present.

Let's turn it now specifically to terrorist attacks in the United States from 1970 to 2010—and this is work being done right now by Brent Smith and Kelly Damphousse, who have actually been supported over the years by several NIJ grants—this shows a heat map of terrorism in the United States. It's a little misleading, because you know the western counties tend to be bigger—it doesn't mean that there's a lot more terrorism in the West, it means we have bigger counties in the West.

But if you divide the kind of terrorism we have to be concerned about in the West, and I've done it somewhat crudely here, where we're looking at left-wing, right-wing, and international attacks, we find striking differences across the country in terms of where these threats are likely to happen. Among all attacks among the United States, and we're counting around 3,000 going back to 1970, international attacks are much more likely to be found here in the Northeast, left-wing attacks are more likely to be found in the Midwest, and right-wing attacks in the South. And I've got another slide that demonstrates this perhaps a little more clearly.

You can see the Northeast, the blue is international attacks, the vast majority of international attacks, like the 9/11 attack, take place in the Northeast. You can see the left-wing attacks are concentrated in the Midwest and to a lesser extent the West, and right-wing attacks are especially concentrated in the South and to a lesser extent the West. And we can do a lot more of these sorts of detailed social science profiles of terrorist organizations, both globally and in the United States.

We also can take this work and start to look at the important dependent variable. One of the things about organizations that are set up to look at terrorism, like the CIA, like the State Department, is they often concentrate on just half the dependent variable, because that's their job. They concentrate on terrorist organizations, but they don't concentrate on groups that could have been terrorist organizations but for whatever reasons were not—so groups with grievances that did not turn to violence. And you can see this work from Steve Chermak and Josh Freilich, who've also been funded in the past by NIJ, that there are some important differences in U.S. white supremacist groups in terms of whether they use violence or not. Some of them are not entirely expected. For example, those that use violence tend to be much more narrowly focused on racial ideology; those that use nonviolence have a much broader perspective. You can see as you go through this list some important differences in terms of countering this particular form of violence.

So I think we've learned a whole lot about groups and how they're put together and what makes them more dangerous over time, why they disappear, why they get more violent, why they attempt to acquire, sometimes, more serious weapons. Let me turn to a third category. Of the three, we probably know the least about governments' actions to counter terrorism and how successful those actions are. Governments, of course, are notoriously close-lipped about wanting to share these sorts of secrets; there's certainly no uniform crime report for these sorts of actions, but we've done now lots of projects that look at these issues to try to figure out, after the government tries a particular kind of policy, does it improve things? Does it make things worse? Or does it have no effect? And a lot of the work we're doing in this area is based on the fact that we now have time-series data, and I just brought one example.

This is work that was published in the journal Criminology with Laura Dugan and Raymond Korte two years ago, and I think it's typical of the kind of work we're doing in this area. We took all of the terrorist attacks of the IRA in Northern Ireland and elsewhere, mapped them over time, starting with the British occupation in 1969 and ending with the major ceasefire after 2003, and then we looked at all of the main things the British tried to do to counteract these violent episodes, and then we tried to control for lots of other kinds of explanations: Were there economic changes, were there changes in crime, was the IRA just responding to attacks by the loyalists in a kind of tit-for-tat motion. So we had probably 20 other kinds of control variables in a multivariate model, and here's what popped out.

We found that of the six major things tried by the Brits to stop terrorism in Northern Ireland, the things on the right-hand side actually made things worse, usually significantly worse; the things on the left-hand side made them better. So out of these six, we could only find one thing where we were able to demonstrate that actual terrorism, controlling for a lot of other explanations, declined over time. So we can do these sorts of things, these sorts of analyses, using time-series data now, for any sort of situation the government tries, basically, of course subject to all of the usual kinds of statistical limitations.

This is terrorist attacks against the U.S. homeland specifically from 1970 to 2008. Usually when I show people these stats, they're a little surprised, because you can see there's been a big decline. We had a lot of action in the 1970s, basically; we had all sorts of action relating to racial disturbances within the United States, we had Puerto Rican separatists during that period, lots of terrorist attacks during that period, and the line goes pretty strongly down. And it's an interesting—to contrast this with another project that was just released by another person who's been funded by NIJ in the past, Kevin Strom.

This shows foiled and executed attacks against the United States, terrorism attacks, for the last 10 years. You see the green bars are foiled attacks. So there's been a real change in what's been happening with regard to foiling attacks, against terrorism. Part of the reason the executed terrorism line goes down in the previous slide is a lot more cases are going from executed to failed, or foiled, over time. So that's one of the big stories.

Which has meant big changes in how we intervene in terrorist cases. This is work being done by Brent Smith and Kelly Damphousse. They looked at changes before and after 9/11with regard to how the FBI is actually prosecuting cases, and the quick message is they've become far more proactive over time. And I'll just give a couple examples. Before 9/11, 29 percent of all indictments for terrorism by the FBI involved prevented acts; now it's up to 60 percent. And just skipping down, another really big change over time, the FBI used to use confidential informants 23 percent of the time and undercover agents 37 percent of the time—or, excuse me, 73 percent of the time confidential informants and 37 percent undercover agents. After 9/11, these numbers have dropped dramatically to 20 percent and 4 percent, respectively; there isn't time now to get these sort of informants planted, they're springing the traps much sooner, and the whole strategy is much more proactive than it was in the past.

We can also drill down and give law enforcement some much more specific information about terrorism in the United States and our responses to it. This slide's also from work done by Brent Smith and his colleagues, and it compares the preparatory behavior done by terrorist organizations over time—by the way, this was funded by NIJ as well—and Smith and his colleagues have found major difference in terms of preparation time. You can see the international groups take the longest to prepare, so with international groups, if they're just in the preparatory phase, you've got quite a bit of time to act. Right-wing groups also take a fair amount of time to prepare; there's quite a bit of time while the preparations are going on before they actually strike. On the other hand, environmental and left-wing groups strike much more rapidly; a lot of times, for these groups, the entire horizon is a single day, so from the time there's an idea to do it, the preparation, the planning, and the strike all happen very rapidly. So you get some very important differences from a law enforcement standpoint across these different types of groups.

So, I'm running out of time; I wanted to try to make the point that I think we are in good—much better shape, 10 years later, in terms of understanding the sort of social and behavioral science of terrorism. When I started running this center, I guess now six or seven years ago, I was hoping I would put myself out of business, and now I'm actually less likely to think that, and here's why. I want to just conclude with this, in terms of the future of terrorism.

What I'm showing you right now is a chart that gives the percentage of the U.S. population living in urban areas, going all the way back to 1800. And it won't surprise probably many people in the room, but what it shows is in 1800, about 7 percent of Americans lived in urban areas. Today, it's somewhere over 80 percent and continues to inch up. What are the implications of that for terrorism?

Well, let me just give you one. In 1850, the Soho neighborhood in London was the most densely populated place on the planet. I mean, this was a time when cities were still in a sense experimental; we weren't sure they were a good idea yet. Well, at that point in time, in 1850, Soho had a population of about 400 persons per acre. When 9/11 occurred, the two Trade Centers sit roughly on an acre of land, and when they're occupied, there were about 50,000 people in those two buildings. So even if there had been suicide bombers in the 1850s, it would have been incredibly difficult to do the kind of destructive attacks that were now possible in 2001.

And in fact, a lot of terrorism is directly related to city attacks. You can see, New York City alone has about 20 percent of the attacks in our database. This is actually somewhat less concentrated than places like Paris, London, et cetera. Terrorism is possible in part, or at least the incredible destructiveness is possible, because we have chosen to live in highly packed urban areas, and it's hard to see a completely successful defense for that particular point.

So I'll end with a quote from Steve Johnson, who wrote this great book I'd recommend to you, called The Ghost Map, who says, "Terrorists are probably destined to be part of human civilization for as long as there are political or religious ideologies that encourage people to blow themselves up in crowded places." We seem to have no lack in that area.

So we have a very ambitious website; if you get a chance, please visit. You'll find the articles I've been referencing very very rapidly, as well as hundreds of others are available on the website. And I will leave it at that. Thank you very much.