如何避免枪击悲剧重演？

GUY RAZ, HOST: And as police begin to piece together a picture of the gunman, Adam Lanza, they will also be looking at possible motives. Here in the studio with me is NPR science correspondent Shankar Vedantam.

And, Shankar, you have reported in the past about building profiles of these kinds of assailants. I mean, usually, we're talking about men. We're talking about often about white men. Does what we know about Lanza fit that profile of a mass shooter?

SHANKAR VEDANTAM, BYLINE: Well, I mean, there's a lot of information that still coming in about Lanza. And I think it's fair to say that some of that information has been conflicting so far, so I don't think we have a full picture of him as yet. It is the case that if you look at the patterns of many of these gunmen, they do happen to be men. They're almost invariably young men, usually young white men. Many of them have long-standing interests in guns. Many of them have reported mental disturbances that troubled people in some ways. Many of them are often described as being loners.

But I'm not sure necessarily these add up to a profile. I think they add up to a pattern but not necessarily a profile. And what I mean by that is they - when you look back at events in the past, you can see the events had similarities to them. But when you talk about a profile, a really effective profile is something that allows you to say here's where the next shooters going to come from, here's where the next school is going to be targeted. And we just simply can't do that right now.

RAZ: If there are patterns, why can't we identify those people who may potentially be a risk?

VEDANTAM: Well, I think if you were to catalog all the people in the United States who are young white men with an interest in guns or, you know, violent video games or who have a history of being bullied, people who may be troubled in some way, you know, that list would run not in the hundreds, but it would run into thousands or the tens of thousands. And the vast majority of them, of course, would never dream of doing something like Lanza is supposed to have done on Friday. So from a scientific point of view, the problem with building a profile is that these are very, very unusual events.

Now, they don't seem like unusual events because they're so salient in our minds right now. But when it comes to these highly unusual events, it's very hard to build predictive models, which is what a good profile is, that tell you what's going to happen next.

RAZ: So if these are rare events, what can authorities do?

VEDANTAM: Well, law enforcement officials increasingly are moving to a model that is called a threat assessment model, as opposed to a profiling model. And it's based on a central insight, which is when you look at many of these recent school shootings or mass shootings, they've often taken time to plan. There's been a lengthy planning period because you just can't pull these things off on the spur of the moment.

And as a result of that, there are often other people who know something about the plan. Maybe they have hazy information or disjointed information. And one of the ideas of the threat assessment strategy is to encourage people to come forward with information, even if they feel that information is incomplete or doesn't make complete sense.

And the idea is that once you get actual information, tangible information, you then go through a process where you say, is this just a person who's upset? Or does this person actually have a plan? Does the person actually have the capacity to carry out the plan? Is there a track record of this person having done anything violent? And the idea is you don't go from zero to 60 in one second. You've taken the information and then say, very carefully, is this a serious threat? What can I do it about it?

RAZ: That's NPR's science correspondent Shankar Vedantam. Shankar, thanks.

VEDANTAM: Thanks so much, Guy.