Articles, Blog

Can Social Networks Predict Gun Violence?


[SIRENS SOUNDING] Chicago faces an
epidemic of gun violence. This year alone, more
than 2,100 people have been shot, more
than 400 killed. The violence that we’re
seeing is really an epidemic. It’s a public health crisis. People have been using
this analogy of crime as an infectious
disease or crime as an epidemic for a long time. When politicians
use it, they use it to draw attention to these
rates that are off the chart. What I tried to do, which
I do think is unique, was figuring out what
does that actually mean? Because rates don’t move. A rate doesn’t shoot anybody. It’s people that do that. 10 years ago, as
part of a project I was working on with a street
outreach worker and actually a prosecutor, and
we were working on the same project, and one of
the outreach workers asked me, so who’s going to get shot next? And I knew what
he didn’t want me to say was, well, it was going
to be a young male between 18 and 35, and there’s a 30%
chance of him being in a gang. And so I sat down with
stacks and stacks of homicide and shooting records, and
I started making, by hand, these sort of network diagrams. And I started
thinking about, well, what applies in this case of
gun violence and what doesn’t? Crime and delinquency
is a group phenomenon. There are some things that
you do sort of on your own. But by and large, when people
do these sorts of behaviors, it’s as a collective. So if person A and person
B, say, rob a bank together, they have a tie. And then if person B, at a
different point in time, say, stole a car with person
C, they have a tie. And these different
ties link together to create chains and
clicks and clusters and ultimately these
very large networks that encompass large parts of
geographic neighborhoods or even entire cities. Then what we would do is we
would take victimization data, either from homicide
records or from the morgue or from the public
health departments, and we would layer
those into the network. And then what we saw was
the clumping and clustering of victims, which we actually
didn’t know what to expect. But what we find
is that if you’re in a network where people
around you are being shot, that your victimization
level skyrockets. Shootings form these chains,
and they cascade from one victim to another victim
to another victim. And it’s not that the shooting
causes the other shooting. It’s that people are
connected in such a way that these consequences
kind of unfold over time. And so when you put
those things together, what it tells you
is it tells you we know kind of where
it’s going to happen within social networks. We know when people
are at elevated risks, and we actually know
when it might happen. And now we’re trying to
piece these things together to see how we might leverage
that for intervention and prevention efforts. It’s hard to read
the newspaper when you’re doing this type of work. One of the things you are
constantly reminded of is that every one of those
nodes in those networks is a real person. But it also reminds me why
we’re trying to do better. Even getting that
number down a little bit means there’s a little bit
less trauma than before.

1 thought on “Can Social Networks Predict Gun Violence?

Leave a Reply

Your email address will not be published. Required fields are marked *