Study: Data predicts spread of ‘bad apple’ officers’ influence
CHICAGO – For years, researchers and police officials across the country have mined data such as arrest and shooting records to examine how criminals influence each other’s behavior – and in the process try to predict who might be the next gunman or victim.
A study to be released Thursday takes the same approach but looks at whether data on police themselves can help determine if officers with complaints of using excessive force can influence a colleague’s chances of being accused of similar conduct.
The answer is yes, said the study’s co-author, Andrew Papachristos, a Northwestern University sociologist who has done extensive work on studying networks of gun offenders, including in Chicago.
“Who we’re connected to affects what we feel, think and do,” Papachristos told the Chicago Tribune. “This is true for both good and ‘bad’ behaviors. Criminologists have long known that networks play an important role in deviance and violence. … We are just starting to take the steps to see if these same ideas apply to ‘police deviance.’ Turns out they do.”
The study, to be published in the journal Criminology & Public Policy, is among the first scholarly attempts to turn policing data inward and examine whether it’s possible to predict potential bad behavior by officers, Papachristos said.
The use of data to predict everything from where to deploy officers to how to identify those most prone to violence has been a staple of the police for years, but not without controversy.
Still, the study comes at a time when Chicago police face intense pressure from a federal consent decree to usher in massive reform to address years of constitutional violations that have eroded public trust and cost the city tens of millions of dollars in settlements.
Among the challenges for the department will be to create an “early warning system” to uncover problem officers.
Papachristos said the study’s conclusions suggest the department needs to take a broader look than just an individual officer’s history.
“If you are going to build an early intervention system that only looks for bad apples, that will only go so far,” he said. “How we pair and assign officers matters – a lot. Officers with a history of abuse have a pretty strong influence on subsequent behavior of other officers.”
Relying on Chicago police misconduct data published by the Invisible Institute, the study’s authors examined more than 8,600 officers named in at least two separate complaints between 2007 and 2015. They found that being named in any type of complaint with other officers with histories of excessive force put officers at a higher risk of receiving similar complaints in the future.
The more officers in the group with histories of excessive force, the higher the risk, the study found.
The research was also able to examine race, age and gender of accused officers, leading to another key finding: The presence of female officers in the group reduced the chances of anyone receiving complaints of excessive force in the future.
“In other words female officers may have a beneficial social influence in police networks,” the report said.
Torreya Hamilton, a Chicago attorney who has filed numerous civil rights lawsuits against groups of officers, welcomed the research, saying she has long wondered if the Police Department used its internal data to identify problem officers.
“I think the Police Department has all the data they need to do something like this,” Hamilton said. “The question isn’t could they use it? The question is will they?”
The notion that officers can negatively influence each other should not come as a surprise. The elite Special Operations Section (SOS) was disbanded in 2007 after officers in the unit faced criminal charges alleging widespread corruption, including false arrests and robberies. And rogue groups of patrol officers have periodically faced criminal charges.
But Papachristos cautioned that those are extreme examples that are worthy of further study to better understand how they happened. The current research, though, looked at everyday interactions that can have negative impact.
“Part of what we are showing is even in smaller cases … it doesn’t mean (bad behavior) doesn’t seep over,” he said. “You need to stop them before it becomes the SOS.”
Andrew Ferguson, a professor at the University of the District of Columbia’s law school who has written extensively on the use of data in policing, said the study is the first he has heard of that is using department data to try to predict bad behavior by officers.
“That raises some fascinating questions,” he said.