May 12, 2017 3:55PM

More Data, More Problems

In recent years, criminologists, law enforcement organizations, government agencies, and other criminal justice experts have been experimenting with various methods of data collection to improve American criminal justice. For example, some researchers look at recidivism—that is, how likely a person who has been incarcerated will end up back in jail or prison—to stem the tide of mass incarceration. Others have turned to "hot-spot policing" to better focus limited police resources on preventing new crimes in highly specific, high-crime areas.  Each method typically has its strengths and weaknesses, and much can be learned from new techniques.

But more data isn't always a good thing. After a long battle with the Sun-Times, the Chicago Police Department released its "Strategic Subject List." From the report:

“We have 1,400 individuals that drive this gun violence in this city,” police Supt. Eddie Johnson said in August, assuring the public his department was keeping tabs on the people on its closely guarded “Strategic Subject List.” “We’ve gotten very good at predicting who will be the perpetrators or victims of gun violence.”

Yet the list is far broader and more extensive than Johnson and other police officials have suggested. It includes more than 398,000 entries — encompassing everyone who has been arrested and fingerprinted in Chicago since 2013.

Nearly half of the people at the top of the list have never been arrested for illegal gun possession. About 13 percent have never been charged with any violent crime. And 20 of the 153 people deemed most at risk to be involved in violent crime, as victim or shooter, have never been arrested either for guns or violence.

What's more, the data isn't being used in the way the designer intended. 

“Let’s say you’ve never been shot or been arrested,” [Yale professor Andrew] Papachristos says. “But if your friends have been shot, you are at a greater risk of being shot.”

But Papachristos, a Chicago native, now distances himself from the way the police are using the Strategic Subject List in Chicago, noting that his work focuses on identifying potential victims, not on predicting the chances someone will shoot another person.

It is impossible to know exactly how the CPD uses this information, but if indices of potential victimization are interpreted as reasons to suspect violent behavior, officers may treat those people in the greatest danger as the most dangerous individuals, flipping the protective function of policing on its head. 

Secret lists, especially ones with hundreds of thousands of people on it, will not reduce crime or make the public safer. Data collection and analysis can be very useful in finding better ways to treat hard-to-solve problems, but putting a bunch of names on a list is akin to building a haystack to find a needle. 

There are many more problems with the data and you should read the whole Sun-Times report about it here

If you'd like to learn more about how data can shape what we know about policing, check out this panel I moderated last year at our most recent criminal justice conference.