The Information Game: Police-Citizen Cooperation in Communities with Criminal Groups by Andrew Cesare Miller MSFS, Georgetown University (2011) BA, The George Washington University (2007) Submitted to the Department of Political Science in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY May 2020 © Andrew C. Miller, XXMM. All rights reserved. The author hereby grants to MIT permission to reproduce and to distribute publicly paper and electronic copies of this thesis document in whole or in part in any medium now known or hereafter created. Signature of Author....................................................................................................... Department of Political Science May 21, 2020 Certified by.................................................................................................................... Roger Petersen Arthur and Ruth Sloan Professor of Political Science Accepted by................................................................................................................... Fotini Christia Professor of Political Science Chair of the Graduate Program Committee The Information Game: Police-Citizen Cooperation in Communities with Criminal Groups by Andrew Cesare Miller Submitted to the Department of Political Science on May 21, 2020 in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Political Science Abstract Criminal groups—gangs, mafias, and drug cartels, among others—likely cause more deaths than interstate war, insurgency, and terrorism combined. This violence and the lack of accountability for perpetrators present a major challenge to states’ cen- tral mandates of providing public safety and administering justice. States fall short of their mandates, in part because they struggle to gain cooperation from citizens. This study is about what I call The Information Game: the competition between the police, which want citizens to come forward with information about violence, and criminal groups, which want citizens to stay silent. I present cycle of silence theory, which posits that collective misperceptions prevent communities from reaching their full potential of police-citizen cooperation. Akin to terrorism, fear generated by criminal group violence makes retaliation appear to be more likely than it is. The violence has the underappreciated but potent second order effect of pushing citizens who are willing to cooperate to hide their disposition from others. Cooperation thus appears to citizens to be less of a norm than it is. I also take new methodological approaches—namely, fielding the first large-scale virtual reality experiment—to test realistically and ethically strategies aimed at promoting cooperation. The results show that providing access to anonymous tip lines, creating awareness of community cooperation norms, and in some circumstances, exposing citizens to police officers of the same ethnicity increase citizen information-sharing with the police. Employing a multi-method research design, this study draws on original surveys in Baltimore, Maryland (N=650) and Lagos, Nigeria (N=1,025) as well as proprietary survey data of criminal justice experts (N=2,700) and citizens (N=109,000) in 113 countries provided by the World Justice Project. I pair the quantitative analysis with first-hand observation as well as interviews with more than 150 citizens, state authorities, and criminal group affiliates. Thesis Supervisor: Roger Petersen Title: Arthur and Ruth Sloan Professor of Political Science 2 Contents List of Figures 6 List of Tables 7 Acknowledgements 8 Introduction 11 I Theory 26 1 The Information Game 27 1.1 Criminal Group Violence . 33 1.2 Citizen Information . 44 1.3 The Information Game . 52 1.4 Threatening the Rule of Law . 60 2 Cycles of Silence 66 2.1 A Threshold Framework . 70 2.2 Cycle of Silence Theory . 78 2.3 Limits of Code of Silence Theory . 93 3 Winning the Game 108 3.1 Cooperation despite Distrust . 111 3.2 Cooperation Strategies . 121 3.3 Empirical Strategy . 135 II Baltimore 145 4 The Game in Baltimore 146 4.1 Background: “The Greatest City in America” . 149 4.2 Method: Baltimore Community Safety Survey . 158 4.3 Results: Baltimore’s Cycles of Silence . 175 4.4 Alternative Theory: “No Snitching” Code . 192 3 5 Baltimore’s Quiet Cooperators 202 5.1 Background: Policing in Baltimore . 205 5.2 Method: Experimental News Reports . 213 5.3 Results: Cooperation in the Shadow of Freddie Gray . 219 III Lagos 226 6 The Game in Lagos 227 6.1 Background: “The Real Lagos” . 230 6.2 Method: Eko Community Survey . 239 6.3 Results: Lagos’s Cycles of Silence . 250 6.4 Alternative Theory: Sons of the Soil? . 262 7 Fighting Fear in Lagos 272 7.1 Background: Policing in Lagos . 274 7.2 Method: Virtual Reality Experiment . 282 7.3 Results: Generalizable Strategies . 290 Conclusion 298 Bibliography 312 4 List of Figures 1.1 Average Annual Deaths by Violence Type . 35 1.2 Homicide Rates Worldwide . 37 1.3 Criminal Organization Power and Homicide Rates . 39 1.4 Percent of Homicides without Convictions in Mexican States . 42 1.5 Homicide Clearance Rates in the United States . 44 1.6 Crime Reporting and Perpetrator Accountability . 48 1.7 Violence Exposure Pyramid . 49 1.8 Distinguishing Criminal Groups from Other Non-state Actors . 54 2.1 Full Potential Cooperation Will in Communities . 75 2.2 Cycle of Silence . 79 2.3 Violence and Risk Perception . 87 2.4 Cycle of Silence Effect on Community Cooperation Will . 91 2.5 Criminal Organization Power and Community Safety . 102 2.6 Criminal Organization Power and Extortion . 106 3.1 Citizen Perceptions of the Police . 119 3.2 Case Contextualization . 139 4.1 Drug-related 911 Calls and Arrests . 156 4.2 Baltimore Interviews . 161 4.3 Baltimore Survey Sample Communities . 163 4.4 Respondent Cooperation Will . 168 4.5 Retaliation Overestimation . 177 4.6 Fear and Retaliation Overestimation . 181 4.7 To Whom Cooperators Reveal Information-sharing . 184 4.8 Standardized Coefficients on Cooperation Will . 187 4.9 Collective Misperceptions and Information-sharing . 188 4.10 Southern Park Heights and Belair-Edison Cooperation Will . 190 4.11 Citizen Perceptions of Drug Crew Impact on Community . 198 5.1 Resident Perceptions of the Police . 209 5.2 Experimental Vignette Scenes . 214 5.3 Information-sharing Summary Statistics . 218 5.4 Treatment Effect on Information-sharing . 220 5.5 Trust in Remaining Anonymous . 222 5 6.1 Lagos Interviewees . 240 6.2 Lagos Sample Market Areas . 243 6.3 Respondent Cooperation Will . 251 6.4 Retaliation Overestimation . 252 6.5 Fear and Retaliation Overestimation . 255 6.6 Standardized Coefficients on Cooperation Will . 261 6.7 Collective Misperceptions and Cooperation Will . 262 7.1 Information Box at an NPF Divisional Headquarters . 279 7.2 Virtual Reality in Use . 283 7.3 Experimental Vignette Scenes . 284 7.4 Information-sharing Summary Statistics . 286 7.5 Treatment Effect on Information-sharing . 291 7.6 Co-ethnicity and Police Fairness . 296 6 List of Tables 3.1 Study Hypotheses . 138 4.1 Survey Respondent Demographics . 165 4.2 Community Descriptive Statistics . 173 4.3 Collective Misperceptions and Cooperation Will . 186 5.1 Experimental Vignette Variations . 215 5.2 Treatment Effect on Information-sharing . 219 6.1 Market Area Summary Statistics . 244 6.2 Survey Respondent Demographics . 247 6.3 Collective Misperceptions and Cooperation Will . 260 7.1 Experimental Vignette Variations . 285 7.2 Treatment Effect on Information-sharing . 292 7.3 Results Summary . 300 7 Acknowledgements It is not for me to judge the quality of my dissertation, but I can say without question that it is stronger thanks to the support of my advisors, colleagues, friends, family, and many others who helped me along the way. The best decision of my academic career has been to ask Roger Petersen, Fotini Christia, Evan Lieberman, and Richard Nielsen to advise my dissertation. They provided the essential balance of pushing me when I was not thinking big enough and supporting me when I faced challenges—of which there were many. This is especially true for Roger, my committee chair. He was there from the outset as I brought him through a maze of research questions before finding a path forward. His guidance throughout was instrumental in ensuring that I asked an important question, theorized about it in a systematic way, and anchored my research in reality. This dissertation would also rest on a much shakier intellectual foundation without his and others’ groundbreaking scholarship that preceded it. Through a combination of enthusiasm and intellect, Fotini convinced me that ideas I once thought too ambitious were possible. Evan’s meticulous and incisive input on everything from minor logistical questions to 300-plus page drafts was invaluable. And, Rich provided critical insights on my project’s framing and analysis as well as on how to navigate the political science discipline. Most of my research took place in Baltimore, Maryland and Lagos, Nigeria. I am indebted to many people in both cities. I am grateful to the Baltimore residents who I interviewed for entertaining my seemingly endless questions in their living rooms, front steps, and back porches. I am equally grateful to the police officers who I accompanied on patrols for the candid conversations late into the night. I also thank Harbor Craft Productions for their outstanding work on the video vignette. The vignette was not a typical film project, but the Harbor Craft team got it done with skill and tact. Lagos, a sprawling mega-city, can be intimidating. When I first arrived there, I had a lot of questions and very few answers. I am grateful to Shelby Grossman, who I consider the dean of Lagos’s expat research community, for generously sharing advice on how to get my feet on the ground. I thank Shelby and Leonardo Arriola for entrusting me to manage an expensive, large-scale survey. A special thank you goes to the Eko Community Survey (ECS) team, who worked tirelessly to get the survey done safely and rigorously while having some fun doing it.