From 1972 to 1991, a Chicago detective named Jon Burge led a group of police officers in torturing confessions out of suspects. They called themselves the “Midnight Crew,” and their behavior eventually resulted in the jailing of Burge and the creation of a reparations council to pay the victims. More recently, the Baltimore Police Department’s Gun Trace Task Force was found to have planted evidence, assaulted innocent citizens, and committed overtime fraud.
Many of the most egregious examples of police misconduct arise from tightly knit groups of officers like these. That’s no accident. Recently released data from the Chicago police department shows that misconduct spreads from officer to officer like an infectious disease. And the same behavior that leads cops to violate the rules often predicts whether they will participate in a shooting.
In 2009, the Invisible Institute sued the city of Chicago to reveal in-depth information on the complaint histories of selected Chicago police officers. After a drawn-out legal battle, the Invisible Institute prevailed and acquired the complaint histories of all officers since 1988. They then processed, standardized, and augmented that data with information on police shootings, uses of force, and a complete duty roster of all officers. In total, the data covers more than 30,000 officers and almost 23,000 complaints between 2000 and 2018.
Because complaints can list multiple officers at once, it’s possible to determine that more than one cop was present at the scene at the same time. Complaints listing multiple officers link those cops together, and by assembling thousands of officers across tens of thousands of complaints, it’s possible to build a giant social network of police interactions.
The illustration above visualizes such a social network. Dots represent officers, linked by lines of complaints. Most officers register few complaints and sit on the outside of the network. But a small portion of officers at the center of network behave differently than those on the outside.
About 1,300 of Chicago’s cops fall into clusters of linked police officers who together have been the subject of at least 100 citizen complaints against them. The list of police within this group reads like a Who’s Who of Chicago police misconduct: From Jerome Finnigan, who led a corrupt unit of cops and plotted to kill a fellow officer, to Raymond Piwnicki, who harassed black citizens using racist language. Officers within this group show not only higher rates of complaints, but also participate in more uses of force and even more shootings. On its own, such a pattern could simply mean that these particular officers are more likely to be street cops or assigned to high-crime divisions. But these central officers are also more than five times as likely to figure in an incident that results in a civil payout by the city for misconduct, according to data on lawsuits involving police officers gathered by the Chicago Reporter.
Within police departments, it’s often well-known that some cops break the rules. In the Chicago Police Department, it was an “open secret” that Burge and his crew extracted confessions using illegal means. Far less serious conduct, such as a reputation for pushing the boundaries, can also get around. Seth Stoughton, a former police officer turned college professor at the University of South Carolina, said “I learned that some, maybe one or two officers when I was [on the force], tended to do things right at the edge of what is acceptable procedure.”
Reputations like Stoughton’s colleagues in turn attract or repel other officers. For example, complaints tend to list officers with more similar use of force rates than if you were to pick officers at random from the department. That could be a result of cops seeking out assignments with others like them. Cops at the center of the Chicago network of 1,300 problem officers were about six times more likely than the department at large to work in one of Chicago’s gang units. Other cops who transferred to those units showed more use of force and greater total civil suit payouts than randomly selected officers. “It’s like magnets. And good officers don’t want to work with [bad ones], because they’ll get in trouble themselves,” said Sam Walker, a professor emeritus of criminal justice at the University of Nebraska and consultant to many police departments. Because high-complaint officers attract like-minded colleagues, they tend to be surrounded within the network by others like them.
Officers prone to misconduct do more than draw in others like them. The data shows that they also may be teaching their colleagues bad habits. Using the Invisible Institute’s data I picked out all the more than 12,000 officers with low complaint rates before 2008 (the year the Independent Police Review Authority, a new police oversight board, became operational). Then I split those cops into two segments: the 863 who had been listed on a complaint with officers at the center of the network, and 12,815 who hadn’t. The officers who had been exposed to the contagious, misconduct-prone cops at the center of complaint networks went on to show complaint rates nine times higher over the next ten years than those who hadn’t.
Their behavior often escalates beyond complaints to more serious violence. The same cops who are exposed to other high complaint officers go on to be listed on four times as many uses of force per year in the next few years. They also commit shootings at rates more than five times higher than their colleagues who weren’t exposed to misbehaving officers.
Stoughton credited part of the infectious quality of misbehaving officers to the process of training young cops. While officers learn the rules of policing at the academy, the probationary period provides hands-on training in the first few months on the job. In that time, many of the procedures they were told to follow in the academy get discarded. “On their first day, every cop hears some variation of ‘Forget everything you learned at the academy’,” Stoughton said.
Bob Verry, a retired police chief and current internal affairs investigator in New Jersey, likened the process of learning misconduct to the “broken windows” theory of policing, in which small violations escalate to larger and larger crimes. “Officers start out with minor things — forgetting their tie clip — and then that becomes forgetting to shine their shoes. … They get away with one punch during an arrest and it just goes on from there,” Verry said.
The data is rich with examples of young officers whose trajectories bent toward misconduct after exposure to bad influences. One rookie cop joined the department in 2001 and was assigned to the 9th police district, on the South Side of the city, after training. In his first three years, he received two complaints; both times, he was exonerated.
In 2004, just after completing his third year on the job, that officer was accused of using excessive force. He was listed on the complaint with four other officers, two of whom had multiple other complaints to their names. From that point on, his complaint rate skyrocketed. Citizens filed several allegations against him over the next five years. His use-of-force rate increased as well, from less than two per year to six in 2014. That officer was Jason Van Dyke, who shot Laquan McDonald that same year and is now on trial for McDonald’s murder.
In theory, patterns of bad behavior like Van Dyke’s should be detected and corrected by supervisors. Complaints by civilians and other officers should trigger official investigations, and officers beset by numerous allegations should be sent to counseling or suspended. But the departmental investigation process is dysfunctional, and the vast majority of civilian complaints do not yield any discipline for the accused officer. When complaints are filed by other cops, discipline is much more likely, but according to data obtained by the Invisible Institute, officers at the center of the network are less likely than others to have complaints filed against them by other cops. The “blue wall of silence,” the tendency among cops to protect their own, appears to safeguard these officers more than others.
“We know that officers who are more in tune [with] or endorse the code of silence … are also more likely to use force, and less likely to use communication in interactions with the public,” said Scott Wolfe, a criminologist at Michigan State University, referring to published research. “They’re more likely to pull a gun and more likely to shoot a gun.”
Using the same data from the Invisible Institute, researchers at the University of Southern California, Yale, Harvard, and the Massachusetts Institute of Technology found in a working paper that networks of complaints and shootings can be used to tell that police violence is contagious. “Violence might spread when officers learn from each other scripts for trying to manage risky civilian encounters or encounters in which they lose control,” said Daria Roithmayr, a professor of law at USC and an author of the paper.
Previous results from two of the researchers had shown that violence behaved in a contagious fashion among citizens of Chicago. Without help or endorsement from those researchers, the Chicago police department used these results as the basis to build an algorithm that would predict which citizens were at higher risk of being involved in a shooting, and then contacted them to try to prevent that violence. The researchers believe that using the contagious nature of violence to target individuals — cops or citizens — is “extremely hard” and “just wouldn’t work well in practice.”
“Once we understand more about contagion, there might be implications for policymakers in terms of how police tasks are assigned or how police units are structured,” Roithmayr said. She cautioned, “There is much more to understand about contagion before we can begin meaningful policy prescription.” Currently, police departments do not formally take an officer’s social history into account when judging whether to reassign an officer or otherwise prevent them from engaging in misconduct. But Verry noted that internal affairs investigators typically look into the other officers on the scene of a complaint.
Illinois Attorney General Lisa Madigan recently filed the draft of a legally binding document that would lay out a series of reforms for the Chicago Police Department to implement in the next few years. In this draft consent decree, one of the required improvements is that the department create a system to pick up on bad behavior before it becomes more serious. A required element for that system is the capacity “to identify group- and unit-level patterns of activity.” The special attention paid to identifying misconduct among cliques within the department suggests that the architects of the consent decree may be aware of the social nature of misbehavior exhibited by groups like Burge’s Midnight Crew.