Cities, crime, and completely irresponsible presentation of data

So the web site NeighborhoodScout.com has presented its list of the “Top 25 Most Dangerous Neighborhoods in America.” Using FBI crime data reported by census tract, NeighborhoodScout.com lists the specific census tracts with the highest predicted rates of violent crime per 1,000 residents.

Cities with neighborhoods on the list include Atlanta, Chicago, Detroit, Flint, Greenville, Houston, Indianapolis, Memphis, Nashville, Rockford, Saginaw, Spartanburg, St. Louis, Tulsa, and West Memphis.

One benefit of presenting the data in this way is the possibility of disabusing people of specious generalizations about whole cities. No–Chicago is not dangerous, though some neighborhoods in Chicago have been facing a wave of violent crime.

But there are at least two problems with the presentation of this data:

  1. Chicago and Detroit are each home to four of the most dangerous census tracts in the country. For less than careful readers, this can give the impression that Chicago and Detroit both face similarly scaled violent crime challenges. But Chicago is approximately four times larger than Detroit, so such an impression would be a serious misunderstanding.
  2. Far worse is NeighborhoodScout.com’s labeling of their fourth column, “My Chances of Becoming a Victim Here (in one year).” This is utterly irresponsible. The numbers do not in fact represent my chances or your chances of becoming a victim in any given census tract. As Andrew V. Papachristos, Robert Wood Johnson Health & Society Scholar at Harvard University, has shown, victims of many violent crimes tend to be members of certain networks. In his brief paper, “The Small World of Murder,” he writes specifically of Chicago, “The vast majority of homicides… are anything but random. Homicide victims and offenders tend to be criminally active and more than two-thirds know each other…. Murders are ‘connected.’ Most murder victims and offenders live in a distinct social world in which most people are just a few handshakes away from each other.” Papachristos specifically takes aim at the idea that all people in a given neighborhood are equally likely to be victims of violent crime. Presentation of risk in the way that NeighborhoodScout.com has done is “vastly overstated because it suggests that all residents are equally likely to be killed. The reality [on the west side of Chicago] is quite different: about 70 percent of the homicides over the last five years occurred in a network of only 1,500 people, all of them with criminal records.” In short, this is what might be called “Murder by Network.” In determining the risk of becoming a victim of violent crime, being in certain networks is a far more significant variable than is geography. So, no, the NeighborhoodScout.com numbers do not represent my risk of becoming a victim. In fact, they don’t represent anyone’s risk of becoming a victim. If you’re a part of these networks to which Papachristos refers, you’re much more likely to be a victim. If you’re not part of these networks, then you’re orders of magnitude less likely to become a victim.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s