The Contagiousness of Police Violence

November 16, 2018
Thibaut Horel, Trevor Campbell, Lorenzo Masoero, Raj Agrawal, Andrew Papachristos and Daria Roithmayr*
Explanations for unlawful police violence focus on individual “bad apple” officers or deviant top-down departmental culture. Recent research suggests that violence may diffuse through social networks, much like a disease spreadson networks of interaction. We investigate whether police violence is contagious—whether an officer’s exposure to earlier police shootings by network neighbors increases the probability that the officer will engage in future police shootings. Drawing on data from Chicago, we construct and analyze dynamic patterns of diffusion of shooting on police professional networks. We find structural anddynamic evidenceconsistent with a dynamicof contagion in police-involved shootings, even after controlling for homophily. Fitting a dynamic model of contagion to police network data from Chicago, wefind that a singleshooting event at the beginning of the study period gives riseto0.5additional offspring shootings during the eight year period under study. Contagion appears to affect a significant number of shootings—it contributes to the occurrence of 141 of 488 (29%) police-involved shootingsin our study. Most remarkably, within two years, exposure to a single shooting more than doublesa network neighbor’s probability of a future shooting.Our findings suggest that interruptingthe transmission of violence in officer networksmay well be an important avenueto reduce police shootings. IntroductionThe October 2018verdict against police officer Jason Van Dyke for killing Laquan McDonald is the first time in the last fifty yearsthat aChicagoofficer has been convicted of murder in an on-duty shooting. Nationwide, police officers shoot between 900 and 1000 civilians every year. Police use fatal force most often against civilians armed with knives or guns, but unarmed black men constitute a significant number of civilian deaths—36 in 2015 alone. Mental health also plays a significant role in fatal force: one in four civilians shot by the police are reported to be experiencing some form of mental health distress at the time of their encounter with police.1
*Author affiliations and contributions: Harvard University(Horel), University of British Columbia (Campbell), Massachusetts Institute of Technology (Campbell, Masoero and Agarwal), Northwestern University (Papachristos) and University of Southern California, Gould School of Law (Roithmayr). Corresponding author, Daria Roithmayr, droithmayr@law.usc.edu. Concept and dataset (DR); Design (TH, DR, AP); data analysis on network features and inter-shootingdelays (TH); Data analysis on Hawkes model (TC, LM and RA); Drafting manuscript (DR); Editing manuscript (AP, TH and TC). Special thanks to the Invisible Institute’s Rajiv Sinclair and Andrew Fan, and to George Wood, Sinclaire Ewing-Nelson, and Jennifer Wu for invaluable data assistance and support.1Fatal Force, Washington Post dataset for 2017 (published January 2018, https://www.washingtonpost.com/graphics/national/police-shootings-2017/)(last accessed October 4, 2018).

Full paper can be accessed here:

1The Contagiousness of Police ViolenceThibaut Horel, Trevor Campbell, Lorenzo Masoero, Raj Agrawal, Andrew Papachristos and Daria Roithmayr*Explanations for unlawful police violence focus on individual “bad apple” officers or deviant top-down departmental culture. Recent research suggests that violence may diffuse through social networks, much like a disease spreadson networks of interaction. We investigate whether police violence is contagious—whether an officer’s exposure to earlier police shootings by network neighbors increases the probability that the officer will engage in future police shootings. Drawing on data from Chicago, we construct and analyze dynamic patterns of diffusion of shooting on police professional networks. We find structural anddynamic evidenceconsistent with a dynamicof contagion in police-involved shootings, even after controlling for homophily. Fitting a dynamic model of contagion to police network data from Chicago, wefind that a singleshooting event at the beginning of the study period gives riseto0.5additional offspring shootings during the eight year period under study. Contagion appears to affect a significant number of shootings—it contributes to the occurrence of 141 of 488 (29%) police-involved shootingsin our study. Most remarkably, within two years, exposure to a single shooting more than doublesa network neighbor’s probability of a future shooting.Our findings suggest that interruptingthe transmission of violence in officer networksmay well be an important avenueto reduce police shootings. IntroductionThe October 2018verdict against police officer Jason Van Dyke for killing Laquan McDonald is the first time in the last fifty yearsthat aChicagoofficer has been convicted of murder in an on-duty shooting. Nationwide, police officers shoot between 900 and 1000 civilians every year. Police use fatal force most often against civilians armed with knives or guns, but unarmed black men constitute a significant number of civilian deaths—36 in 2015 alone. Mental health also plays a significant role in fatal force: one in four civilians shot by the police are reported to be experiencing some form of mental health distress at the time of their encounter with police.1*Author affiliations and contributions: Harvard University(Horel), University of British Columbia (Campbell), Massachusetts Institute of Technology (Campbell, Masoero and Agarwal), Northwestern University (Papachristos) and University of Southern California, Gould School of Law (Roithmayr). Corresponding author, Daria Roithmayr, droithmayr@law.usc.edu. Concept and dataset (DR); Design (TH, DR, AP); data analysis on network features and inter-shootingdelays (TH); Data analysis on Hawkes model (TC, LM and RA); Drafting manuscript (DR); Editing manuscript (AP, TH and TC). Special thanks to the Invisible Institute’s Rajiv Sinclair and Andrew Fan, and to George Wood, Sinclaire Ewing-Nelson, and Jennifer Wu for invaluable data assistance and support.1Fatal Force, Washington Post dataset for 2017 (published January 2018, https://www.washingtonpost.com/graphics/national/police-shootings-2017/)(last accessed October 4, 2018).

2Police shootings that involve excessive force do great harm. They erode the legitimacy of law enforcement and createa relationship of profound mistrust between officer and civilian.2Like gunshot violence for civilians, police shootings tend to cluster in socially and economically struggling neighborhoods in communities of color.Social movements like Black Lives Matterhaveorganized around the issue of police brutality, andhave worked to make visible the distinctly racial profile of police shootings.Several high-profile killings of civilians have spurred nationwide protests and sparked debate over police use of force.Data from Chicago provide a useful up-close look at the structure and dynamics of police shootings. Between 2010 and 2016, Chicago police engaged in 435 shootings. Officers killed 92 people and wounded 170. Almost 80 percent of the 262 people shot by Chicago police were African-American. Latinosaccounted for 35 of the shooting victims, nearly 14 percent of the total. Only fourteenof those shot were white, less than 6 percent ofthe total.3Sortingout which of these shootings were justified and which involved excessive force is extremely difficult. Almost all police shootings in Chicago have beenadjudicated by department officials and independent review authorities as justified.4At the same time, a recent investigation by the Department of Justice in the wake of Laquan McDonald’s killinghas cast doubt on such findings. The DOJ reportfound fault with a number of police practicesthat create a high risk of excessive force. Ofparticular concern was the practice of shooting civilians during a foot chase when they arerunning away from officers anddo not pose a threat to officers or the public, and shooting into cars that are driving away but are not posing an imminent threat to the public.5Theoretical literature on the causes of police violence falls into two categories. The first focuses on the so-called “bad apple” micro-level theory of violence, which attributes violence to the deviant traits (like the authoritarian personalities) of individual officers. On this view, violence is generated bottom up. A second macro-level category ascribes police violence to top-down police department incentives, set up by departmental leaders and enforced by way of both formal and social approval and/or punishment.Recent research on the contagiousness of violence suggestsa third possibility: that violence is also generated at the meso-level, spreading from officer to officer much like a disease. A literature on the diffusion of violence supports this possibilityas an additional source of excessive force. Research on violent uprisings and gangs suggests that under certain conditions, people who are exposed to earlier 2See Tammy R. Kochel, Assessing the initial impact of the Michael Brown shooting and police and public responses to it on St. Louis County residents views about police, 4–6 (2015), http://opensiuc.lib.siu.edu/cgi/viewcontent.cgi ?article=1001&context=ccj_reports [http://perma.cc/RK99-DVLW]. See generally Anthony Bottoms & Justice Tankebe, BeyondProcedural Justice: A Dialogic Approach to Legitimacy in Criminal Justice, 102 J. Crim. L. & Criminology 119, 132 (2012); Tom R. Tyler, Enhancing Police Legitimacy, 593 Annals Am. Acad. Pol. & Soc. Sci. 84, 85 (2004).3Id.4See A Historic Murder Conviction, The Chicago Tribune, October 6, 2018. 5U.S. Department of Justice Civil Rights Division and US Attorney’s Office, Northern District of Illinois, Investigation of the Chicago Police Department, January 13, 2017 (DOJ Report).

3violence will engage in violence at a later time at a greater rate than those who are not exposed.6In addition, other literature suggests that social learning plays an important role in police department dynamics. Studies suggest that officers frequently learn from each other key scripts about how to exert control over civilians or reduce the risk of a civilian encounter.7In this paper, we investigate the contagiousness of police violence in police networks, usingChicagoas a case study. More specifically, we investigate whether an officer’s exposure to a shooting by a neighbor on his network increases the probability that he will engage in a shooting in a subsequent encounter with a civilian. We analyze data from civilian complaints and shootings, collected by the Chicago Police Department (CPD) and Chicago’s Independent Police Review Authority (IPRA) between 2008 and November 2015. Drawing from this data, we construct networks of officers who interact with each other, to test whether violence—here, police shooting—is contagious. We generate networks by connecting officers who interact with each other; in network parlance, we draw “edges” between the “nodes” of officers who are “tied” or “connected.” Here, we construct a layered network that connects police officers who engage in shooting. We first construct a network ofCPD officers who are listed together on the same civilian complaint (the “complaint” network), to map officers who interact with each other in responding to calls with civilians. We assume that being listed on the same complaint constitutes evidence of anexisting relationship among the officers listed, a relationship that might mediate the transmission of violence.8We then layer onto the complaint network a separate network of police-involved shootings for which the shooting officers are connected on the complaint network. Layering shooting data onto the complaint network allows us to create this separate network consisting of shooting officers who have been listed together on a civilian complaint and are statistically likely to have had an existing relationship during the shooting. We analyzethis separate shootings network. In particular, weinvestigate the relationship between an officer’s exposure to violence by his neighbors on theshootings network and his future acts of violence.6Henk W. Houweling and Jan G. Siccama, The Epidemiology of War, 1816-1980, 29 J. Conflict Res. 641 (1985) (violent uprisings cluster and diffuse across time and space); Ben Green, Thibaut Horel and Andrew Papachristos, Modeling Contagion Through Social Networks to Explain and Predict Gunshot ViolenceinChicago, 2006 to 2014, 177 JAMA Intern Med.326 (2017)(diffusion of gunshot violence exhibits evidence of contagion).7For a full review of the literature, see Daria Roithmayr, The Dynamics of Excessive Force,UNIVERSITYOFCHICAGOLEGALFORUM:Vol.2016 , Article 10.8Empirical studies support the idea that co-offender relationships are highly correlated with existing relatively stable relationships between co-offenders. Jean Marie McGloin and Holly Nguyen, the Importance of Studying Co-Offending Networks for Criminological Theory and Policy,” in PROC.THIRD ANNUAL ILLICIT NETWORKS WORKSHOP, Montreal, Quebec, October 2011; Jean Marie McGloin, Christopher J. Sullivan, Alex R. Piquero, and Sarah Bacon, Investigating the stability of co-offending and co-offenders among a sample of youthful offenders,” 46 Criminology 1 (2008). See also Albert J. Reiss, Co-offending and criminal careers, in CRIME AND JUSTICE:AREVIEW OF RESEARCH(1988); Albert J. Reiss and David P.Farrington, Advancing knowledge aboutco-offending: Results from a prospective longitudinal survey of London males, 82 J. of Crim. Law and Criminology 23 (1991).

1The Contagiousness of Police ViolenceThibaut Horel, Trevor Campbell, Lorenzo Masoero, Raj Agrawal, Andrew Papachristos and Daria Roithmayr*Explanations for unlawful police violence focus on individual “bad apple” officers or deviant top-down departmental culture. Recent research suggests that violence may diffuse through social networks, much like a disease spreadson networks of interaction. We investigate whether police violence is contagious—whether an officer’s exposure to earlier police shootings by network neighbors increases the probability that the officer will engage in future police shootings. Drawing on data from Chicago, we construct and analyze dynamic patterns of diffusion of shooting on police professional networks. We find structural anddynamic evidenceconsistent with a dynamicof contagion in police-involved shootings, even after controlling for homophily. Fitting a dynamic model of contagion to police network data from Chicago, wefind that a singleshooting event at the beginning of the study period gives riseto0.5additional offspring shootings during the eight year period under study. Contagion appears to affect a significant number of shootings—it contributes to the occurrence of 141 of 488 (29%) police-involved shootingsin our study. Most remarkably, within two years, exposure to a single shooting more than doublesa network neighbor’s probability of a future shooting.Our findings suggest that interruptingthe transmission of violence in officer networksmay well be an important avenueto reduce police shootings. IntroductionThe October 2018verdict against police officer Jason Van Dyke for killing Laquan McDonald is the first time in the last fifty yearsthat aChicagoofficer has been convicted of murder in an on-duty shooting. Nationwide, police officers shoot between 900 and 1000 civilians every year. Police use fatal force most often against civilians armed with knives or guns, but unarmed black men constitute a significant number of civilian deaths—36 in 2015 alone. Mental health also plays a significant role in fatal force: one in four civilians shot by the police are reported to be experiencing some form of mental health distress at the time of their encounter with police.1*Author affiliations and contributions: Harvard University(Horel), University of British Columbia (Campbell), Massachusetts Institute of Technology (Campbell, Masoero and Agarwal), Northwestern University (Papachristos) and University of Southern California, Gould School of Law (Roithmayr). Corresponding author, Daria Roithmayr, droithmayr@law.usc.edu. Concept and dataset (DR); Design (TH, DR, AP); data analysis on network features and inter-shootingdelays (TH); Data analysis on Hawkes model (TC, LM and RA); Drafting manuscript (DR); Editing manuscript (AP, TH and TC). Special thanks to the Invisible Institute’s Rajiv Sinclair and Andrew Fan, and to George Wood, Sinclaire Ewing-Nelson, and Jennifer Wu for invaluable data assistance and support.1Fatal Force, Washington Post dataset for 2017 (published January 2018, https://www.washingtonpost.com/graphics/national/police-shootings-2017/)(last accessed October 4, 2018).

2Police shootings that involve excessive force do great harm. They erode the legitimacy of law enforcement and createa relationship of profound mistrust between officer and civilian.2Like gunshot violence for civilians, police shootings tend to cluster in socially and economically struggling neighborhoods in communities of color.Social movements like Black Lives Matterhaveorganized around the issue of police brutality, andhave worked to make visible the distinctly racial profile of police shootings.Several high-profile killings of civilians have spurred nationwide protests and sparked debate over police use of force.Data from Chicago provide a useful up-close look at the structure and dynamics of police shootings. Between 2010 and 2016, Chicago police engaged in 435 shootings. Officers killed 92 people and wounded 170. Almost 80 percent of the 262 people shot by Chicago police were African-American. Latinosaccounted for 35 of the shooting victims, nearly 14 percent of the total. Only fourteenof those shot were white, less than 6 percent ofthe total.3Sortingout which of these shootings were justified and which involved excessive force is extremely difficult. Almost all police shootings in Chicago have beenadjudicated by department officials and independent review authorities as justified.4At the same time, a recent investigation by the Department of Justice in the wake of Laquan McDonald’s killinghas cast doubt on such findings. The DOJ reportfound fault with a number of police practicesthat create a high risk of excessive force. Ofparticular concern was the practice of shooting civilians during a foot chase when they arerunning away from officers anddo not pose a threat to officers or the public, and shooting into cars that are driving away but are not posing an imminent threat to the public.5Theoretical literature on the causes of police violence falls into two categories. The first focuses on the so-called “bad apple” micro-level theory of violence, which attributes violence to the deviant traits (like the authoritarian personalities) of individual officers. On this view, violence is generated bottom up. A second macro-level category ascribes police violence to top-down police department incentives, set up by departmental leaders and enforced by way of both formal and social approval and/or punishment.Recent research on the contagiousness of violence suggestsa third possibility: that violence is also generated at the meso-level, spreading from officer to officer much like a disease. A literature on the diffusion of violence supports this possibilityas an additional source of excessive force. Research on violent uprisings and gangs suggests that under certain conditions, people who are exposed to earlier 2See Tammy R. Kochel, Assessing the initial impact of the Michael Brown shooting and police and public responses to it on St. Louis County residents views about police, 4–6 (2015), http://opensiuc.lib.siu.edu/cgi/viewcontent.cgi ?article=1001&context=ccj_reports [http://perma.cc/RK99-DVLW]. See generally Anthony Bottoms & Justice Tankebe, BeyondProcedural Justice: A Dialogic Approach to Legitimacy in Criminal Justice, 102 J. Crim. L. & Criminology 119, 132 (2012); Tom R. Tyler, Enhancing Police Legitimacy, 593 Annals Am. Acad. Pol. & Soc. Sci. 84, 85 (2004).3Id.4See A Historic Murder Conviction, The Chicago Tribune, October 6, 2018. 5U.S. Department of Justice Civil Rights Division and US Attorney’s Office, Northern District of Illinois, Investigation of the Chicago Police Department, January 13, 2017 (DOJ Report).

Advertisements

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 )

Connecting to %s