Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 1 of 124 DISPARITY REVIEW – PART II Disparity Review – Part II Developing a Deeper Understanding of Disparities Identified in Part I December 2019 SEATTLE POLICE DEPARTMENT 1 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 2 of 124 DISPARITY REVIEW – PART II EXECUTIVE SUMMARY The Consent Decree included several requirements related to bias-free policing, including the general mandate that the Seattle Police Department (“SPD”) “deliver police services that are equitable, respectful, and free of unlawful bias, in a manner that promotes broad community engagement and confidence in the Department.” To that end, it also required updates to SPD’s Bias Free Policing policies and training. The Consent Decree also stated that, in consultation with the Community Policy Commission (“CPC”), SPD should consider whether to revise SPD Manual 5.140. In 2014, after consultation with the Department of Justice and the Seattle Police Monitor, SPD amended Policy 5.140 (Bias-Free Policing), which now contains a “Disparate Impacts” provision, thereby satisfying the requirements of the Consent Decree. See Dkt. 118 at 5-8. Under this policy, SPD committed to eliminating policies and practices that have an unwarranted disparate impact on certain protected classes of people. SPD recognizes that even in the absence of intentional bias, the long-term impacts of historical inequality and institutional bias can result in disproportionate enforcement activities. With that in mind, the Department is committed to identifying and eliminating unwarranted or unnecessary disproportionate enforcement while protecting public safety and public order. As noted by the Monitor in his Tenth Annual Systemic Assessment on Stops and Detentions, there is disparity in the City of Seattle’s law enforcement activities with respect to stops and detentions, which are police-civilian contacts that involve an officer stopping and detaining, for a limited time, an individual. Known as a Terry stop,1 such contact is authorized under law and policy for purposes of investigating, based on an officer’s reasonable suspicion, whether the individual is engaging, has engaged, or is about to engage in criminal activity. SPD’s analysis of its stops and detentions, in various reports including the Part I Disparity Review (“Part I Report”), confirms disparities in stops and detentions. The question, therefore, becomes what factors – be they policies, trainings, or shared information – contribute to these disparities. Both the Monitor and the SPD sought to establish an acceptable methodology for identifying where disparities in police-community interactions exist, and the scale of any identified disparities. This work is essential as there is no commonly agreed upon metric for disparity in the academic or professional literature. In their published reports, the Monitor and the SPD have demonstrated the usefulness of the technique of Propensity Score Matching (PSM), which uses regression to “score” how similar events are to each other across a variety of factors and match 1 In Terry v. Ohio, 392 U.S. 1 (1968), the United States Supreme Court held that such brief detentions are authorized under the Fourth Amendment when, under the totality of circumstances, an officer has reasonable suspicion to believe that criminal activity is afoot. SEATTLE POLICE DEPARTMENT 2 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 3 of 124 DISPARITY REVIEW – PART II them for comparison. For example, PSM was used to match a Terry stop where the stopped individual was Black to a stop where the stopped individual was White, but all other known factors (available in fielded data) were as similar as possible. In the current report the SPD builds upon its findings in the Part I Report as well as the Monitor’s work in the Tenth Annual Systemic Assessment of Stops and Detentions. Those two predominantly quantitative reviews confirmed common knowledge – there is widespread disparity in policing, just as there is in almost every aspect of society.2 In the Part I report, SPD found (i) a disparate rate of frisks for Asian, Black, and Hispanic individuals compared to Whites during a Terry stop; (ii) a lower rate of finding a weapon during those frisks for Asians, Blacks, Hispanics, or American Indian/Alaska Natives compared to Whites during a Terry stop; and (iii), a higher rate of pointing firearms at Asian, Black and Hispanic individuals compared to Whites. What neither SPD’s Part I Report nor the Monitor’s Tenth Annual Systemic Assessment of Stops and Detentions could establish, however, were insights into what drives these rates of disparity and what could be done within the police department and community to lower these rates. Accordingly, SPD consulted with Dr. Jack McDevitt3 to attempt to answer these questions. To do so, SPD developed a method to review randomly selected incidents of Terry stops where the subject was frisked and uses of force (“UOF”) where a firearm was pointed at a subject (Type I UOF). These reviews were conducted both internally and, in partnership with the CPC, with a variety of community groups. In addition, SPD collaborated with Charles C. Lanfear4, an outside expert, and attempted to apply additional statistical techniques – Logistic Regression and Structured Equation Modeling (SEM) – to isolate the specific quantifiable factors of stops and 2 David Grusky, Charles Varner, and Marybeth Mattingly, Pathways: State of the Union 2017, (Stanford Center on Poverty & Inequality) (https://inequality.stanford.edu/publications/pathway/state-union-2017). 3 Dr. Jack McDevitt is the Professor of Practice in Criminology and Criminal Justice and the Director of the Institute on Race and Justice, at Northeastern University, College of Social Sciences and Humanities. 4 Charles Lanfear is an instructor at the University of Washington in both the Center for Statistics and the Social Sciences, and the Department of Sociology. Mr. Lanfear holds a bachelor of arts degree in sociology, a master’s degree in public policy, and a master’s in sociology. He currently is completing his doctoral studies in sociology at the University of Washington. SEATTLE POLICE DEPARTMENT 3 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 4 of 124 DISPARITY REVIEW – PART II uses of force that were most associated with the decision to frisk and the decision to point the firearm, including the race of the subject(s) in the interaction. SUMMARY OF MAJOR FINDINGS The report consists of two distinct sections of analyses: a new qualitative5 approach to examining disparity in police-community interactions in Seattle, and continued work on better isolating contributory factors, quantitatively, in any identified disparity. The primary report is a qualitative assessment of fifteen (15) randomly selected incidents (where officers frisked an individual during a stop, or when officers pointed a firearm at an individual). This report is based on both internal, SPD critical reviews of the incidents, and external, direct feedback from various community members who reviewed a sample of those same incidents. This report also includes a secondary, technical report (Appendix A), that provides a comprehensive overview of the results of standard logistic regressions (used to examine which elements of the stop are predictive of whether a frisk will occur, and which elements of an incident are predictive of whether a firearm will be pointed at a subject). The Department had intended for the logistic regression work to be supplemented by Structured Equation Modeling (SEM), as SEM allows the consideration of factors that are not inherently included in the available data and hired an expert to conduct this work. SPD’s expert, Charles C. Lanfear, determined during his work on the logistic regressions, that SEM would not add any additional explanatory insights.6 As additional data become available, SPD may conduct an SEM analysis in the future. The primary, qualitative section of the report, found consistent issues about how the department trains and monitors the use of a witness/victim description to identify and stop potential subjects. The critical assessments of internal SPD sworn and civilian staff and the community sessions highlighted interactions where the degree to which the stopped individual “matched the “Qualitative research is a type of social science research that collects and works with non-numerical data and that seeks to interpret meaning from these data that help understand social life through the study of targeted populations or places.” Crossman, Ashley, (October 24, 2019) An overview of Qualitative Research Methods, www.thoughtco.com/qualitative-research-methods-3026555 5 6 The full explanation of why the SEM approach ultimately was not included can be found on page 56 in the Appendix. SEATTLE POLICE DEPARTMENT 4 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 5 of 124 DISPARITY REVIEW – PART II description”7 left many reviewers wanting a stronger match. Concerns were also raised about the pointing of firearms when no weapon was visible and the subject appeared to be generally compliant (as well as pointing of firearms in crowded locations when no weapon was readily visible). Internal SPD reviewers and community members identified the same concerns around the size of the deployed response to some of the incidents. They wondered whether a significant police response could play a role in both officer decision making and escalating an incident, thereby requiring a higher level of force to be used. Finally, there were a variety of concerns around the style and manner in which officers engaged subjects, and the role that might have played in all decision-making – by both officers and subjects. Specifically, there were discussions around the role that trauma and stress, and a lack of connection/trust to the communities involved influenced officer behavior. The conclusion of the report outlines how SPD is using this feedback to consider modifications of protocols and training to potentially decrease identified disparity rates. The technical report (included in Appendix A), while presented separately for those interested in the continuing effort to better isolate quantifiable factors associated with officer decision making and not generally discussed in body of this report, highlights additional areas of interest. The outside expert first found that Terry stop hit rates, defined as a stop resulting in an offense report, arrest, citation, or referral for prosecution, for all racial/ethnic groups were relatively similar (ranging from 61.5% to 68.6%), as was the distribution of different types of stop dispositions across demographic groups. Additionally, the expert found that hit rates were significantly lower for on-viewed incidents as opposed to dispatched calls. Of particular note, the expert found that 62% of all stops occurred when the officer had a “known description” of a person they were looking for. This is similar to the qualitative review process finding that “matching a description” was an important element in frisking. In relation to the issue of safety and firearm pointing, the expert found that White subjects were more likely to have a weapon hit than non-White subject. 7 This refers to the description provide by a 911 dispatcher or another officer to the officer(s) who made the decision to stop. Descriptors included perceived race, perceived gender, clothing, color and length of hair. SEATTLE POLICE DEPARTMENT 5 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 6 of 124 DISPARITY REVIEW – PART II Of clearest concern was the expert’s finding that the more “out of place”8 the subject is, the more likely he or she is to be stopped and frisked. The out of place effect reinforces the need for the Department to evaluate the call-taking/dispatching segment of police response to mitigate bias. The finding that how the interaction started – based on a 911 call or the officer on-viewing a situation – is a factor in the likelihood that a broadcast subject description is the basis for a stop, requires additional examination around how descriptions are received, communicated, and acted upon. That is, the type of call influences whether the officer makes a stop simply because the subject matches a description of a person of interest, instead of because observable behavior, is an area where disparity may be introduced and needs additional examination. For the Department, the finding that frisk, hit, and weapon recovery rates differ by precinct and beat strongly calls for deeper examination of why there are not similar outcomes across the City. In summary, based on these novel approaches to understanding disparity, the SPD has committed to the following actions – full details are in the body of the report – to begin to address the identified disparities and continue this work. 1. Amplify the training and guidance around how much of a match between the description of a suspect and the appearance of the subject there must be to constitute a “match” to initiate a stop, and safety frisk, if warranted (and how specifically that match must be described in order to appropriately documents a stop and/or frisk) 2. Review policies, trainings, and protocols for the pointing of firearms 3. Develop enhanced procedures and trainings for 9-1-1 call takers and 9-1-1 dispatchers to improve their ability to recognize and mitigate implicit bias “Neighborhood racial context may also interact with an individual’s race, Officers may believe that certain individuals “belong” in specific neighborhoods and increase their social control against individuals who deviate from this pattern (Novak and Chamlin 2012; Stewart et al. 2009). This form of racial profiling is commonly referred to as “out-of-placeness,” “out-of-place policing,” or “racial incongruity” (Brunson and Weitzer 2009; Fagan and Davies 2000; Novak and Chamlin 2012; Steward et al. 2009). Individuals who are “out of place” may be perceived as more suspicious. (Gaston 2019; Fagan and Davies 2000) or possibly in the area for nefarious purposes (Meehan and Ponder 2002). Conversely, people “out of place” may be easier to identify for officers who have been dispatched with a specific subject description.” Appendix, page 35. 8 SEATTLE POLICE DEPARTMENT 6 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 7 of 124 DISPARITY REVIEW – PART II 4. Address “disparity-associated9” issues involving officer professionalism 5. Continue the work on identifying and responding to disparate impacts by continuing to partner with the CPC in developing and holding incident review community sessions, as was trialed during this analysis. INTRODUCTION Under Paragraph 223 of the Consent Decree, the Court retains jurisdiction over this matter “until such time as the City has achieved full and effective compliance and maintained such compliance for no less than two years.” On January 10, 2018, the Court entered an order finding the Department to be in “full and effective compliance” as of the date of the Order, thus commencing the two-year “sustainment period.” Dkt. #439. The Court further ordered the parties and the monitor to “meet, confer, and prepare a plan for discharging their obligations under the Consent Decree” during this two-year period. On March 13, 2018, the Court entered an order approving the Sustainment Plan developed pursuant to the Court’s January 10th order. This plan, and an attached matrix of deadlines, became the governing documents for this Sustainment Period. To a large extent, the Sustainment Plan follows the scope of Consent Decree itself, as it requires SPD to self-assess those core topical areas that, collectively, cover the entirety of SPD’s obligations as set forth in Section III of the Consent Decree (“Commitments”) and assessed, comprehensively, in Part I . The Sustainment Plan also includes commitments by SPD not specifically called for under the consent decree. SPD is committed to continuing the work it has done to further itself as a “learning” institution and driver, nationally, of innovative and data-informed policing. This analytic project is a part of that promise. Included within the Sustainment Plan is SPD’s commitment to continue its emphasis on impartial, bias-free policing, to regularly examine disparities across its data to ensure that this emphasis is carried out in practice, and to provide the Court with a fuller view of “disparities with respect to stops, searches, and seizure; use of force; and other law enforcement activity.” SPD also demonstrates that commitment with its Bias-Free Policing policy (Seattle Police Manual, 5.140) and requiring minimizing bias training that is coordinated with the state’s Criminal Justice Training Center. As noted later in the report, the community groups – and some internal reviewers – consistently noted concerns about how officers managed scenes. These concerns were not specific to the decision to stop, frisk, or point a weapon; rather, they were about how the subjects were treated and whether that would contribute to perceptions that disparity or bias was involved. 9 SEATTLE POLICE DEPARTMENT 7 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 8 of 124 DISPARITY REVIEW – PART II Accordingly, SPD completed its Part I Report to reach agreement with the parties to the consent decree on a methodology to confirm which type of police-community interactions had disparity and which displayed the highest rates of that disparity. In the Part I Report, the SPD confirmed that the PSM methodology was both a statistically appropriate and replicable approach for this work, which the department could continue to use without the need for external assistance. As such, SPD currently is building the technique into its Data Analytic Platform so that updates to disparity analytics can be inherent in all data analyses and management reviews. SPD is building the PSM R-package into the DAP/Tableau environment so that analyses like those in Part I can be routinely run on these data without having to re-setup the entire analytic package in the Rsoftware. In the current report, SPD sought to build on the information produced in the Part I Report. The findings in the Part I Report confirmed that, while there were consistent disparities across all the examined interactions, obvious remedies for the disparity could not be found. The goal of this report was to look beyond the data., SPD committed to focusing this report on gathering qualitative insights into those interactions that were associated with the highest levels of disparity in the Part I report, so that recommendations could be made on changes to policies, procedures, trainings, and overall employee education. The core of the Disparity Review - Part I Report is a description of the qualitative work undertaken – both internally and with a variety of community groups – to critically review randomly selected incidents of Terry stops where the subject was frisked and Uses of Force where a firearm was pointed at a subject (Type I UOF). Additionally, with the PSM methodology now approved, this report also includes a technical report focused on a more traditional complete assessment of the data trends on stops, detentions, and firearm pointing. This technical report presents basic descriptive statistics and logistic regressions examining the role that subject characteristics and other factors play in the decision to frisk and point a firearm. The introduction of the SEM approach was planned, but the currently available data models did not support its final inclusion in this report. BACKGROUND SPD collects and reports out on its stops and detentions and uses of force. To provide the readers of this report with a synopsis of the overall picture of Terry stops and firearms pointing during the time period of this study, SPD provides the following information10. It is important to note that the SPD does not use “stop and frisk” as an enforcement tactic. Thus, while some police departments may look at officer stops as a proxy for proactivity, the SPD does not view the number of stops as a measure of its officers’ proactive policing. Additional data related to SPD’s stops and detentions may be found in Appendix F, beginning at page 111, and on SPD’s website. 10 SEATTLE POLICE DEPARTMENT 8 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 9 of 124 DISPARITY REVIEW – PART II During the study period, January 1, 2016 through June 30, 2018, the Seattle Police Department responded to over 978,000 calls for service. SPD officers conducted a total of 19,511 Terry stops during that same time period. SPD conducted weapons frisks in 21.6% of all Terry stops and recovered a weapon in 20% of those frisks. In other words, SPD frisked subjects in only 21.6% of all Terry stops and recovered weapons in 20% of those frisks. In 2017, of the Terry stops that could be associated with a CAD11 event, 75.7% of those Terry stops were the result of a call for service, meaning SPD officers were responding to a 9-1-1 call from the public. In 2018, over 70% of all Terry stops city-wide were responsive to a request for service from the public. In 2017, 23.1% of all stops resulted in an arrest; an additional 37.9% of all Terry stops resulted in a police report being written. In 2018, 27.5% of all stops resulted in an arrest, an increase of 41.1% from 2017, and 38.3% of all stops resulted in a police report being written. Just over half of all Terry Stops recorded during the study period were of White subjects (50.5%). Black subjects accounted for 30.6% of all stops, Hispanics accounted for 4.8%, Native American/Alaskan Natives accounted for 3.3%, and Asians/Pacific Islanders accounted for 2.8%. Hispanic subjects had the highest frisk rate (27.0%) with a weapon hit rate of 18.7%. Though White subjects were frisked at the lowest rate (18.2%), they accounted for the highest weapon hit rate (25.2%), (see Table 1). Table 1: Stops, Frisks, and Weapon Hit by Subject Race Male subjects accounted for more than three quarters of all Terry Stops (77.8%) during the study period. They were frisked at a rate of 24.5% with a weapon hit rate of 20.9%. Female subjects were frisked at a rate of 10.9% with a hit rate of 12.3% (see Table 5). A “CAD event” is a unique incident, given a unique identifying number, logged in response to a call from the public (“Call for Service”) or a report from an officer in the field, “on-view,” of an incident or event requiring their response. CAD Events are classified as “DISPATCH” when in response to a Call for Service and “ONVIEW” when reported by an officer in the field. 11 SEATTLE POLICE DEPARTMENT 9 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 10 of 124 DISPARITY REVIEW – PART II Table 2: Stops, Frisks, and Weapon Hit by Gender During the study period, there were a total of 4,392 use of force (UoF) incidents that involved 917 incidents of firearms pointing (20.9% rate). Three quarters (76.7%) of all uses of force reported during the study period were Level 1 use of force incidents, the lowest level of force. These incidents had a rate of firearm pointing of 26.2%. Level 2 use of force incidents accounted for 21.6% of all uses of force and were associated with a 3.0% rate of firearm pointing (see Table 3). Table 3: Use of Force Count and Firearm Point Rate by UoF Incident Type The majority (70.7%) of use of force incidents occurred during a dispatched call for service while 29.2% occurred when an officer initiated the contact (Onview). The firearm point rate for dispatched calls was 19.7% while the rate for Onview firearms pointing was 23.9% (see Table 4). Table 4: Use of Force Count and Firearm Point Rate by Call Type The largest percentage (39.4%, 1,732) of all uses of force involved a White subject, but incidents with a White subject had the lowest rate of firearm pointing (16.6%). Hispanic subjects had the SEATTLE POLICE DEPARTMENT 10 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 11 of 124 DISPARITY REVIEW – PART II highest rate of firearm pointing use of force incidents (37.8%) but accounted for just 3.4% (148). Black subjects were involved in 32.1% (1,408) of all uses of force with a firearm pointing rate of 26.6% (see Table 5). Table 5: Use of Force Count and Firearm Point Rate by Subject Race As this data shows, and as SPD found and reported in its Part I Report, disparities exist in the rate of frisks for Asian, Black, and Hispanic individuals compared to Whites during a Terry stop; the rate of finding a weapon during those frisks for Asians, Blacks, Hispanics, or American Indian/Alaska Natives compared to Whites during a Terry stop; and the rate of pointing firearms at Asian, Black, and Hispanic individuals compared to Whites. With this information in hand, SPD sought to determine whether it could further analyze these incidents to inform operational and policy changes. SEATTLE POLICE DEPARTMENT 11 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 12 of 124 DISPARITY REVIEW – PART II METHODOLOGY Qualitative Review A. Case Selection Based on the findings in the Part I Report, SPD concluded that the best path forward was to examine a sample of Terry stop (frisk) and Use of Force (firearm pointing) reports for the same time period as the Part I Report. During the study period, January 1, 2016 through June 30, 2018, the Seattle Police Department responded to over 978,000 calls for service. SPD officers conducted a total of 19,511 Terry stops during that same time period. SPD conducted weapons frisks in 21.6% of all Terry stops and recovered a weapon in 20% of those frisks. Given the length of incident videos and reports, SPD determined that randomly selecting twenty (20) incidents of each type (Terry frisks and Firearms Pointing), for a total of forty (40) incidents to be reviewed, was a sufficient qualitative review sample.12 The SPD identified these forty (40) incidents through a stratified random sample to ensure that an equal number of officer on-view and officer dispatched calls were included for each incident type, as well as an equal number of White and non-White subjects. Initially, SPD intended to include a cross-section of incidents across the two-and-a-half-year study period.13 SPD then reviewed the written reports and video footage for each of these forty (40) incidents to determine whether they were appropriate for internal and external review. This initial examination focused on assessing whether the officer’s behavior was governed by policy/training, thereby restricting the officer’s discretion.14 This initial review found that all but 12 The random, stratified selection was drawn from the population of incidents that were matched in the Part I propensity score analyses. This population consists of every non-white frisk/weapon pointing incident and its corresponding PSM-matched white counterpart. The sample was constructed to ensure it was proportional to the white/non-white distribution, As this was a pilot of a type of analysis the SPD is not familiar with any other department doing, and given the timelines, this sample size was selected as a compromise to get a good sample, along with the needs to have a compendium of information that could be examined, in detail, in a reasonable amount of time given the desire to watch videos, read reports, etc. 13 SPD later determined that the difficulty of obtaining, reviewing, and redacting in-car video (the only video available in 2016 and most of 2017), along with the enhanced content on body-worn video, and the significant changes in policies/training since 2016, justified over-sampling additional 2018 and 2019 cases. 14 An example of such a restriction would be the service of a high-risk warrant. SPD policy and training require an officer to have their firearm at low/high-ready or on-target when serving high-risk warrants In these incidents, the SEATTLE POLICE DEPARTMENT 12 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 13 of 124 DISPARITY REVIEW – PART II six (6) of the incidents were either not discretionary or had issues with useable video footage. So, an additional twenty-two (22)15 incidents were pulled using the same methodology. Tables 6A and 6B, below, present details of why certain randomly selected incidents were not included in the intensive internal or external review. The juvenile stop was excluded so as not to even risk disclosing any juvenile information. Table 6a. Reason for Terry Stops with frisks to be excluded from review Terry % of Excluded Incidents Video Issue Visible/ Witness Weapon High-Risk Stop / Encounter 5 4 4 Consented to Search / Admitted to Weapon Possession 5 23% 18% 18% 23% Incident Known to Subject Arrest History Total 3 1 22 14% 4% 100% Table 6b. Reason for Firearm Pointing incidents to be excluded from review Warrant Service Visible/ Witness Weapon High-Risk Stop / Encounter Known Subject History / Shots Fired Call Juvenile Subject Total decision to frisk or point the firearm was driven by policy and/or the behavior of the subject and was not a discretionary action. For the safety of the officers and the subjects/bystanders, the Department is not considering changing those practices. It would have been disingenuous to ask for feedback on these incidents. Given that the random pull of 40 had a “success” rate of 15%, the department then began examining, one at a time, additional random cases to hopefully get to 20 each – as originally desired. With the additional pulls not resulting in a shift in success rate, the decision was made to stop at 62 once 10 “viable” incidents were found. Reviewing enough additional incidents to get the 6 incidents up to 40 was not feasible. This alone was a finding. Also, the experience of finding these cases made it clear there was no possible way that the community groups could review more than 10 incidents, given the amount of material that needed to be examined. 15 SEATTLE POLICE DEPARTMENT 13 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 14 of 124 DISPARITY REVIEW – PART II Assault on Officer Pointing % of Excluded Incidents 6 6 12 2 3 1 30 20% 20% 40% 7% 10% 3% 100% In both the Terry stops and firearm pointing incidents, most of the reviewed incidents were excluded from further review because there was a visible, or witness reported, weapon; the nature of the interaction was high-risk (i.e., identified stolen vehicle, eluding vehicle, certain warrant service); or, the subject stated they had a weapon or agreed to be frisked so they could use their pockets. In these incidents, the decision to frisk or point the firearm was driven by policy and/or the behavior of the subject. Ultimately, fifteen (15) incidents out of a total of sixty-two16 (62) were assessed by the internal SPD team. Ten of those were selected for community review, and 7 of the 10 ultimately were reviewed and discussed in the community working group. An index of the total sample (Appendix B) is provided for review of the general nature of the cases that were sampled and ultimately selected. 16 Ultimately, 62 incidents were randomly selected and screened in order to identify a sample of fifteen (15) incidents for review. During the screening, SPD learned that, in most of the incidents, the involved officer’s discretion as to whether to conduct the frisk or point his or her firearm was strictly limited by safety-related policy or training guidelines. (This finding itself is important. As a result, SPD is going to review the policy and training guidelines to determine if any changes can be made consistent with national best practices. See discussion below at page 27.) Because the purpose of the internal and external review was to analyze and discuss officer’s decision-making and the potential role of racial bias, the incidents in which the officers’ lacked discretion were eliminated as not suitable for review. Following the internal review of the fifteen (15) incidents, ten (10) of them were planned for the external review – primarily due to time/capacity concerns, but also due to the in-depth internal review revealing additional policy/training guidelines. A review of the incidents with the CPC eliminated one (1) incident for a final total of nine (9) incidents intended for the external review. In the community workgroups, there was enough time for community members to review all nine (9) and discuss seven (7). SEATTLE POLICE DEPARTMENT 14 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 15 of 124 DISPARITY REVIEW – PART II B. Internal and External Review After the sample of fifteen (15) incidents was finalized for internal review, a survey instrument (Appendix G) was shared with a team of five (5) internal SPD members, both sworn and civilian, to structure their reviews of the incidents. The instrument asked the SPD team to assess their perceptions of safety, appropriateness, dignity, and fairness from the perspective of the officer(s) involved, the subject(s) involved, and any community member(s) who may have witnessed the event. Based on the internal review of the incidents, two (2) incidents were not included in the sample to review with community members. One was a suicidal caller who had stated he wanted to kill himself with a knife. Officers arrived, held him at gunpoint briefly, and had him drop the knife. The other incident involved a fleeing vehicle, where the subject hit other vehicles, bailed from the car, and then fled. The officer ended up taking the person into custody at gunpoint, with the help of witnesses. This incident was excluded as it was a high-risk situation. An additional incident was included in the sample, but all the groups exhausted their scheduled time before addressing that incident, so it was not subjected to external review. The SPD sought the City of Seattle’s Community Police Commission’s (CPC) partnership for their expertise in engaging community and facilitating working meetings focused on addressing issues in law enforcement. In a series of planning sessions, the SPD and the CPC decided to hold a series of smaller meetings, where community members could caucus together to enhance openness and safety in shared experience, as opposed to a large-scale meeting. The plan was to hold four (4) meetings back-to-back in one week, with the following communities each having a session: (a) African-American/Black, (b) Asian/Pacific Islander, (c) Native/Indigenous, (d) Latinx. Due to final scheduling issues, three total meetings were convened, and the Latinx and Native/Indigenous participants agreed to co-host their session. The three (3) meetings were all held from 6:30 p.m. to 8:30 p.m. during the week, at a local community college. The SPD/CPC team chose to compensate the community members, at a rate of $30 per hour, given the time and nature of the work to which they were being asked to commit. Additionally, hot and cold food and beverage were provided. The CPC was responsible for the direct outreach to community members to schedule them for each session. A mutuallydeveloped “invite” (Appendix C) was shared/posted for recruitment and information. In total, twenty-five (25) community members participated in these first-of-their-kind reviews. In the meetings, a facilitator, experienced with working in the community of the evening, began each session with some mutual work to create space for the discussion to follow. During this time, an SPD representative explained the purpose of the meeting, including the intent and results of the Part I Report. SPD generally explained that the intent of the meeting was to help the Department assess these incidents from the perspective of those experiencing the interactions, so that any institutional bias might be overcome. The SPD representative then explained the general context of each incident, shared the privacy redacted body-worn and/or SEATTLE POLICE DEPARTMENT 15 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 16 of 124 DISPARITY REVIEW – PART II in-car video (pausing and revising as requested), and shared the redacted hard-copy officer written report. The facilitator then led a discussion around each incident. During this discussion, a sworn SPD supervisor was available to answer technical/procedural questions that were raised by the participants. Two (2) of the sessions were able to view and discuss two (2) different incidents, each, while one session was able to view and discuss three (3) incidents, for a total of seven (7) incidents reviewed by the community participants. The goal had been to review 9 incidents, but two of the groups only had time to do two each. Quantitative Review The exact methodological steps are outlined in the technical report (Appendix A), but in this section the intent and general approach of each section is provided. Given the extensive reporting on the results of the PSM analysis in the Part I Report, that work is not repeated here. A. Descriptive Statistics With the PSM methodology approved following with the Part I Report, in this report the SPD sought to produce a more traditional quantitative report. As such, the technical report (Appendix A) is structured to present a logical progression of complexity in telling the story of what the available quantitative data allow. SPD is fully aware – as was the motivation behind the drive to have the PSM methodology accepted – that presenting basic descriptive statistics would ignore the reality that more is affecting every interaction between officers and community members than the demographics of those involved. However, it does provide a useful baseline for demonstrating the importance of including other factors in more advanced techniques. The technical report begins with general descriptive statistics on the occurrence and trends of stops, frisks, and firearms- pointing, looking across demographics and temporal and geographic factors. Much of this work also can be done by the general public using the department’s Terry Stops and Detentions and Use of Force databases. B. Regressions The next step in a logical evolution of understanding quantitatively what is leading to different rates of interactions is controlling for all available factors. This initially was tried in the Part I Report – and confirmed by the Office of the Inspector General – and the regression modeling could not account for enough of the variance to produce significant results. However, to document that the department did attempt to examine the relationships among these factors, the SPD reports on a collection of fixed effects logistic regressions. C. Structural Equation Modeling (SEM) After assessing the feedback provided during the Part I Report, SPD decided that it needed to attempt some additional quantitative analyses focused on the decision to frisk or point a firearm – particularly in a way that could best handle latent (unmeasured data) variables, as well as the bi-directional influence of factors. More simply, SEM allows for the analysis to infer, mathematically, variables that are not able to be included in a regression model. All the work SEATTLE POLICE DEPARTMENT 16 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 17 of 124 DISPARITY REVIEW – PART II done both by SPD, and in other areas around disparity, highlights that existing data explain a relatively small amount of what is affecting officer decisions. Applying the SEM methodology to the issue of disparity in law enforcement is relatively new, so SPD identified an expert in comparative demography and computational statistics at the University of Washington. This expert found that the results of the fixed effects logistic regression indicated that it was an equally appropriate (and easier) tool, with the currently available fielded data, than the SEM approach, so the SEM work ultimately was not included. Suggestions were provided for how to do the necessary work to permit a meaningful SEM model in future reports. RESULTS Qualitative Data Collection and Outcomes A. Internal Review The five (5) internal SPD reviewers each completed the same 15 incident assessments, each with five (5) response prompts (Appendix D. Their responses were content analyzed to generate both quantitative and qualitative summaries of how the raters viewed the incidents. There was a total of seven (7) Terry frisks incidents and eight (8) firearm pointing incidents. An electronic survey tool was used to collect these responses. a. Terry Frisk Reviews Content The internal raters were prompted to comment on their perception of the sense of safety of officers, subjects, and community, as well as on the dignity and fairness of the incident. For the Terry incidents each of the general content of these responses are summarized below. i. Officer Safety The raters generally identified four elements that they believed most significantly contributed to the officers’ feelings of safety in the frisk incidents that they reviewed: (1) whether the stop involved a person(s) believed to have a weapon; (2) whether the subject(s) was in a vehicle; (3) whether the incident took place during the day or night; and, (4) if the subject(s) was following commands, particularly keeping their hands visible and away from their pockets/waistband. In one particular incident, the raters also identified the fact that the officer was alone when the nighttime stop was initiated. The internal term focused on officer feelings of safety and the subject’s behavior in response to commands designed to ensure they did not have access to a weapon. Most important was the officer’s prior or concurrent knowledge that the subject(s) likely had a weapon. If the caller, an eyewitness, or the officer(s) themselves stated they observed a weapon, the officer was going to conduct a frisk as safely and quickly as possible to remove the weapon from the interaction. SEATTLE POLICE DEPARTMENT 17 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 18 of 124 DISPARITY REVIEW – PART II ii. Subject Safety The raters generally identified four elements that they believed most significantly contributed to the subjects’ feelings of safety in the incidents: (1) the officers’ ability to calmly and professionally control the scene so no aggressive actions were necessary; (2) the officers’ skill at explaining what was happening and why; (3) the deployment of any firearms/weapons as an indicator that the officers were ready to use significant force; and, (4) the overall size of the police response, which seemed to have a dual effect, with a large response for some creating a sense of control and safety, and a large response creating a sense of anxiety and fear about how dangerous the officers must have assumed the incident to be.17 In one incident several raters noted that the officer’s decision to seemingly “surprise” the subject by coming from behind him, not announcing himself until he was right behind the subject, and then loudly announcing he was SPD. Several raters noted that while this made sense in decreasing the opportunity for the subject to potentially access a weapon, it might have startled the subject into doing something that could have escalated the incident. iii. Community Perception The raters generally identified four elements of the incidents that likely would have affected a community member’s perception of the incident and sense of safety: (1) the ability of the officers to quickly and calmly have control of the scene; (2) the extent to which it was obvious the officers were explaining what was happening to the subjects; (3) the size of the response, with the larger the size generally seen as escalating community concerns around the safety of all those involved in the incident; and, (4) the resolution of the incident, with the raters feeling that particularly in the larger, more sudden responses, a community member would be concerned/confused if the official resolution apparently involved no official enforcement action. In two incidents, several raters focused on how officers – whether many of them or a solo officer – quickly exited their vehicles, often leaving them in the middle of the street, to then go and apprehend the subject. This quick action, coupled with a lengthy detention, left several raters concluding that a random community member would be unclear why there was such significant action for ultimately the subject to be released. 17 In an examination of disparity, it was an important empathetic step to consider the incident from the perspective of the subject to understand their behaviors and reactions. The intent was to help identify if specific officer actions may be contributing to subject behaviors that lead to the decision to frisk/point. SEATTLE POLICE DEPARTMENT 18 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 19 of 124 DISPARITY REVIEW – PART II iv. Subject Fairness and Dignity The raters generally identified five elements of the incidents that might have affected the subject’s feelings about the fairness and dignity of the interaction: (1) the consistent effort of the officers to politely explain what was happening was seen to increase these positive perceptions; (2) the length of the detention was an important factor as it was seen to show an effort to not connect them to anything other than what they were looking for; (3) the presence of weapons obviously was seen as something that could lessen feelings of dignity especially if the stopped subject was not connected to what the officers were looking for; (4) the resolution of the incident was important, as even with professional conduct it is difficult to maintain a sense of fairness and dignity if the subject was directly involved in the original issue; and, finally, (5) several raters noted that being told you “fit the description” seemed to cause concern rather than serve as a means of explaining the stop. In one incident, the raters concluded that the subject completely understood why the frisk and overall interaction had occurred, as he admitted to brandishing the knife at the other subject. When subjects were not in an apparent behavioral health crisis, their knowledge that they, in fact, had been seen with a weapon, seemed to make the subjects not be surprised that they were being stopped. v. How would incident have been perceived The raters generally focused on the size of the response, the presence of weapons, and the overall apparent resolution of the incident as important factors in how the community would have perceived the incident. Raters additionally noted that it would be difficult for any general community member to understand what was happening if they had not heard the details from the dispatch call, or known what/who the officers were looking for if it was an officer on-view action. The biggest factor in how it likely was perceived was how it was resolved, with raters consistently expressing that community members would have questions/concerns if there was a large response, a long detention, and then the subject was released with no official action. This would lead to a conclusion that the person had been misidentified, with the potential explanation being a biased identification. b. Firearm Pointing Reviews Content The internal raters were prompted to comment on their perception of the sense of safety of officers, subjects, and community, as well as on the dignity and fairness of the incident. For the Firearm Pointing incidents each of the general content of these responses are summarized below. i. Officer Safety The raters generally focused on four elements of the incidents for their interpretation of the officers’ likely sense of safety: the type of the call – particularly a call where a weapon was believed to be present; the behavior of the subjects – particularly fleeing from the officers or SEATTLE POLICE DEPARTMENT 19 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 20 of 124 DISPARITY REVIEW – PART II reaching for something in pockets or a car; the time of the day of the incident, with darkness being a significant contributing factor; and, the visible presence of a weapon. More specifically, raters noted that the facts from the dispatched call, or what was observed directly by the officer, combined with the behavior of the subject primarily drove the decision to point the firearm. One rater noted some concern that with a large response it may prime the officers to assume something serious is about to happen and they may be more likely to point their firearm. Another rater had some concerns about the decision of officers in two incidents to have their firearm pointed while there were bystanders very close to the subject, especially while running. ii. Subject Safety The raters generally focused on four elements of the incidents that may have affected the subject’s sense of safety: the tone and manner in which the officers explained what was happening and why; the scale of the police response and presence of multiple firearms; whether the subject knew the officers suspected or saw them in possession of a weapon; and the tactics of the officers, including using the patrol vehicles to ensure the immediate area was well-lit if it was dark, so that they would not be concerned about what they could and could not see. More specifically, raters credited the officers in almost every incident with explaining what was happening, giving clear instructions on what they wanted the subject to do, and how the officers positioned themselves and their equipment, as reasons why these incidents did not feel like someone was going to discharge their firearm. In one incident, raters had an issue surmising the feelings of the subject considering their state of behavioral crisis. In another incident, the focus was on the location of the incident and the presence of bystanders, such that raters felt the subject may have been concerned that the officers would be concerned about not just their personal safety and be more likely to directly engage the subject. iii. Community Perception The raters generally focused on four elements of the incidents for insights into a community member’s likely sense of the incident: the presence of a visible weapon – either by the subject(s) or the officer(s); the overall scale of the response by the officers; the calm and professional nature of the officers’ commands (and whether the subjects were calm and following the commands); and, the officers’ ability to control the scene, including how long the incident took to resolve. More specifically, raters noted community members were likely to have increased concerns about the ability of an event to have a peaceful outcome the longer it goes on. Additionally, raters commented on the intent of the officers’ commands and whether it was clear they were trying to prevent injury to the subjects. Raters indicated that when subjects immediately followed the officers’ clear commands, community members felt better about the likely outcome of the incident. There were concerns over how the community would be affected by a large-scale response with many visible officer weapons, concluding that community members would assume SEATTLE POLICE DEPARTMENT 20 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 21 of 124 DISPARITY REVIEW – PART II a serious event with an increased likelihood of a force-involved resolution. In three of the events, there were concerns around how the community would react to the deployment of firearms, including a long gun, with bystanders in the immediate vicinity (including officers running with the firearm in their hand through a crowd). iv. Subject Sense of Fairness and Dignity The raters generally focused on four elements of the incidents for insights into the subjects’ likely feelings of fairness and/or dignity: how the officers treated the subjects – including the use of derogatory language, making fun of language challenges, and offering medical treatment as quickly as necessary; the officers efforts to consistently explain what was happening and why; how quickly the interactions were able to be resolved – this was particularly true for those individuals who apparently were not the individuals the officers thought they were; and, how the officers made it clear they were looking for something/someone specific – often a gun/shooter – and were not concerned about finding other things, such as drugs. More specifically, raters noted an officer is not always going to be one-hundred percent certain, and the subject’s sense of fairness and dignity and how they are treated is dependent on the officer explaining, professionally, why they were selected. Additionally, when officers were committed to ensuring the subjects were physically unharmed and that they were allowed to continue on as soon as it was obvious they were not who the officers were looking for, subjects seem more accepting of the encounter. v. How would incident have been perceived The raters focused on how the general community would have perceived these incidents if they were not aware of the reason why officers were engaged at all. The main concern was that for someone trained in the law and procedures, the officers in all of the situations were responding appropriately. However, for others the sheer number of officers and guns that were pulled out, seemingly randomly, is disconcerting for the public. This was particularly true for the incident where one officer was pointing a rifle at a subject who was not holding a weapon, from only five to six feet away. There also were concerns around the language officers used in certain incidents, and if a community member heard this, they would perceive it as unprofessional and potentially indicative of disdain or bias toward the subject. Finally, given there were incidents where guns were drawn and people were frisked, and then allowed to walk away – the community would be very confused by these incidents, overall. B. External Community Review After the internal SPD team reviewed the incidents, the SPD team worked with the CPC to screen the incidents to ensure those presented would both minimize secondary trauma or victimization and allow for meaningful discussions around policies/trainings. Ultimately, seven (7) incidents were chosen for review during the community meetings. The community meetings were structured to caucus similar community members with each other, except for the Latinx and Native/Indigenous communities who agreed to meet together SEATTLE POLICE DEPARTMENT 21 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 22 of 124 DISPARITY REVIEW – PART II due to scheduling issues. From the random selection of the full sample, and the appropriateness review of the final sub-sample, all attempts were made to present incidents that represented the community in the room, though some incidents that included White subjects were included to prompt a discussion about how those incidents were conducted differently than the experience of the communities in the room. Some of the groups did not feel that the subjects were obviously (visibly) members of a non-White community and had concerns about whether the incidents accurately reflect police interactions with subjects who are clearly non-White. Two of the group sessions, African American/Black and the Latinx/Native/Indigenous meetings, reviewed two incidents (videos and officer reports), each. The Asian/Pacific Islander group was able to review three incidents (videos and officer reports). The summary of the discussion comments, and SPD’s responses to the questions, is included in Appendix E. a. “We Always Match the Description” In each of the community sessions, one of the main issues discussed was the common explanation/justification that the individual stopped, frisked, and sometimes targeted with a firearm, “matched the description” of a person officers were looking for, particularly a person believed to have a weapon. In one incident (#1 in Appendix A), a group of young adults is hanging in and around a car in a grocery store parking lot, after dark. Patrol officers, acting on intelligence from gang detectives looking for a shooting suspect, initiate a stop, remove everyone from the vehicle, and handcuff and frisk some of them. At one point during the incident, one of the young people, upon hearing the officer’s explanation for why they were stopped – “we are looking for a suspect in a shooting, and you matched the description”—responds with her own statement of “we always match the description.” In another incident, where officers are looking for a shooting suspect based on a witness’ description, officers stop a man whose shoes, and potentially his apparent racial/ethnic identity, were the characteristics that matched the given description broadcast over the radio. In the video the officers note that several of the other factors do not match before they initiate the stop. In the other three Terry incidents, the officer(s) witness the subject engaging in potentially illegal behavior and subsequently conduct safety frisks. In the discussions around “matching the description,” community members focused on whether officers must match a certain proportion of elements of the subject’s description. SPD personnel explained that there are frisk elements (articulable, reasonable suspicion) that an officer must document, but there is not a decision matrix or threshold for how closely a person must match a given description. The SPD team went on to explain that because it is easy to quickly change one’s outer clothes, and that in both of the incidents under review they were looking for a shooting suspect, the preciseness of the match can often be something like the shoes or an officer’s reasonable belief that the subject is a known person. A community member pointed out that the shoe description in the particular incident was a color scheme that matched a brand of shoes perhaps more commonly worn by a particular community – a fact that could artificially increase the likelihood that a member of that community would be incorrectly stopped. One community member specifically asked about how were 911 call takers and dispatchers trained, and what SEATTLE POLICE DEPARTMENT 22 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 23 of 124 DISPARITY REVIEW – PART II were their procedures around ensuring that they are not just passing along caller bias that may lead to official disparity. The community further had concerns about the general nature of descriptions for an event – even a shooting – that occurred hours, to days, prior to the stop. The discussion centered around both the logical conclusion that the immediate threat had lessened as more time passed, and the suspected location of where the suspect might be found naturally would expand and continue to pull in more and more members of the subject’s community. The community stressed that as more time passed, the specificity of the description should increase to help ensure only the true subject was involved in any justice system contact. An additional interesting facet of the conversation was that the outcome of the stop and frisk – finding the person or a weapon – would have had little effect on the group’s reaction to the video. The ends did not justify the means – as “this happens all the time,” and you can either “find what you’re looking for,” “give someone a bad experience with the police,” or “someone could have gotten hurt if the young subjects had not reacted so calmly.” Was it worth it? b. Overwhelming Force The other consistent operational comment/concern that the community members had was the seemingly routine practice by SPD to have a large number of officers respond to each incident that was reviewed – from a routine traffic stop to a report of a man with a gun. The discussion generally focused on understanding if this was standard practice and whether the Department understood that the large-scale response was likely to escalate the potential anxiety, risk, and/or trauma perceived by the subject. There were questions of whether that number of officers would respond to these situations if the person was White, and this did somewhat play out in one Terry incident where a young adult White man is stopped for suspected narcotics activity, and only the two officers in the car respond. SPD explained that since this person was not known to have a weapon, where in other incidents they were specifically believed to, and this was a two-officer car, there was initially no reason for the officers to have additional back-up. The SPD explanation emphasized how much of the post-consent decree training centered on having a minimum number of officers present during certain types of calls to help prevent an officer getting into a situation where they might feel a use of force was their only option. c. Wrapped into this discussion was a conversation around how it seemed very important for one officer to give directions to the subject(s), as a confusion around directions from multiple officers could quickly get a subject into a dangerous position. SPD explained that when the incident in question occurred, the Department had already shifted its training to a “contact and cover” model, where the lead officer on the call is the person who speaks to the main subject, while other officers maintain scene integrity and security. So, as a supervisor, they would have addressed the issue with a reminder of the training standard. Overall, the community members had many concerns about the number of guns (ranging from one to seven) that were routinely drawn and pointed at subjects in the videos. There were SEATTLE POLICE DEPARTMENT 23 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 24 of 124 DISPARITY REVIEW – PART II questions about having guns drawn and pointed, especially when running, when bystanders were in the immediate area. SPD explained that when responding to a dangerous call, especially a weapons call, SPD is going to respond with many officers to ensure the scene is safe and no one else, including themselves, gets hurt. There were additional questions about why an officer, a good distance away from a subject believed to have a weapon, but with none visible, has their firearm drawn and pointed at the subject. SPD explained the research/tactical training based on the Tueller finding that an armed subject can close a distance of twenty-one (21) feet in a second and a half (1.5 seconds). This finding – the 1.5 seconds – suggests that an average person would not be able to draw a firearm and neutralize a threat in that amount of time. SPD focuses on finding cover and creating distance, but if a subject is credibly believed to have a deadly weapon, an officer will have their weapon drawn, at the least. This conversation also involved a discussion on how different firearms are brought to the scene. In one incident, a “threats with a gun” call, one officer is about five (5) feet from the subject pointing a rifle at the subject. The community members were distressed that a rifle was used in such close quarters. SPD explained that only certain officers are rifle-certified and if they are on a call, before getting to the scene and exiting the vehicle, they will have formulated a plan for which firearm to have as their primary weapon. If they decide the rifle is appropriate, they are committed to it and are not going to drop it, fling it over their back, or otherwise switch weapons. In each of the firearms pointing incidents, there was a significant discussion about the level of stress in/tone of voice the officers exhibited. Specifically, there were concerns about it feeling like officers were coming to the scene exhibiting obvious signs of stress and trauma and sounding/behaving like they “were ready to kill someone.” Community members wanted to know what training/care officers received around managing their stress/trauma. They also wanted to know what systems were in place to help manage what calls an officer was sent to, knowing what they recently had dealt with during their shifts. SPD explained that generally the very aggressive, particularly cursing, behavior is not everyone’s desired tactic, but that in some situations, with individuals whom officers know respond only to that level of engagement, officers sometimes use it. The work the department is doing on officer wellness, including smart call dispatch was discussed. d. “Disparity-Associated” Comments A majority of the other comments focused on how the interactions could be done more respectfully and safely. These are important elements to overall community trust, but were not focused on changes that could lead to reductions in event-specific disparity. Primarily, these comments involved ensuring the officers explained what they were doing, why, and when – including officers not reaching for a visible pocket knife without letting the subject know what they were going to do. There was concern here that sudden incursions into the personal space SEATTLE POLICE DEPARTMENT 24 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 25 of 124 DISPARITY REVIEW – PART II of community members, particularly those with limited police exposure, would increase the subject’s anxiety and the chance that the subject might unexpectedly react, which might result in a more forceful response from the officer(s). These issues were paired with questions around why, in some cases, there seemed to be no specific reason for frisking the subject. In a traffic stop case, where the subject had unclear ownership paperwork, the subject was frisked when asked to get out of the car to better talk about the ownership issue. In this case, the officer noticed a knife clipped to the side of the subject’s pocket and quickly removed it. SPD explained that a visible weapon like a knife always is going to be removed and that was why this subject was frisked, not because of the overall encounter, while acknowledging that the officer could have informed the subject that he was going to remove the knife to ensure everyone’s safety. Other comments centered around always treating the subject with respect, and how some seemingly meaningless comments served no purpose other than to lessen the position/humanity of the subject, and further increase the power dynamic difference between the officer and subject. In one incident, the community members noted that the nature of the interaction changed based on the Field Training Officer (FTO) repeatedly asking the new officer “if everything felt right.” The community noted that this seems a perfect example of how an implicit bias could be passed on, unofficially. SPD explained that the FTO does need to explain to the officer in training things to be aware of to keep themselves safe. In this incident, the discrepancy in the ownership paperwork was a potential sign that the vehicle was stolen – a factor that would make the scene riskier. There were additional concerns around how officers generally interact with White and non-White individuals. A community member noted that officers tend to take the “guardian” approach to White individuals, in that they try to make them feel safe and speak to them with respect. NonWhite individuals experience the “warrior” approach where everyone is seen as a threat that needs to be controlled. The community stressed that the only way to overcome this was to diversify the department – in demographics and in the backgrounds (including where they grow up) of the officers. This discussion included questions around how the department handles people who have limited to no understanding of English, and how that affects every stage of the justice system. One community member stressed that having more women on the department would help with treating people with respect and not using force, as they felt from their experience and the women included in the reviewed incidents, female officers were more calm and resorted to force less. Finally, there was a discussion around what the department was doing at the end of incidents to restore the dignity of the subjects involved. Individuals, especially those who were not involved in the behavior leading to the police interaction, are stressed, demoralized, traumatized, and vulnerable following any interaction where they are frisked, and especially when they have a firearm pointed at them. The community members clearly wanted the department to consider SEATTLE POLICE DEPARTMENT 25 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 26 of 124 DISPARITY REVIEW – PART II that just because the officer(s) gets to drive off to the next call, the community members involved do not, and that compounds any direct or indirect effects of disparate impacts. NEXT STEPS & RECOMMENDATIONS The Seattle Police Department is committed to continuing these avenues of inquiry into where disparity is occurring in its interactions with the public, and what the possible causes of that disparity are. The SPD is now equipped and experienced with both quantitative and qualitative tools that it will continue to use and improve to address these issues. A. Qualitative Findings SPD commits to continuing to hold incident review community sessions, as was trialed during this analysis. The department concluded that the feedback was extremely valuable. The community members generally reported being appreciative that the work was being done, though there was understandable skepticism about if any substantive changes would be made due to community feedback. Additionally, many of the community members expressed a desire to have more time. These were the community members who stated they had hoped they would be reviewing more intense incidents where lives were on the line – so they could make a real difference. SPD and CPC both expressed a willingness, in the future, to coordinate such sessions, ensuring that support services were in place to respond to any secondary trauma/re-victimization. a. Addressing Disparity in Terry Frisk Rates The SPD commits to amplifying its training on articulable, reasonable suspicion for stopping an individual based on the description from a witness/victim. SPD acknowledges that eyewitnesses often can be mistaken, especially in a stressful situation involving a potential weapon, and there is a delicate balancing of community safety from finding the weapon and individual rights infringed by a search. This work also will examine the current policies and procedures around the length of time between the incident and role that the amount of time since the incident must play in a decision about the reasonableness of a stop. This work also will involve additional analysis and implicit bias training and revised procedures for 911 Call Takers and Dispatchers. It is important that these first responder have just as many tools and insights as officers to recognize when they are encountering direct bias and when they may be transmitting implicit biases. b. Addressing Disparity in Firearm Pointing Potential actions that could address the disparity involved in this type of use of force are harder to identify as in each of the incidents pulled in the total sample, and the community-review sample, the officers were acting according to established safety procedures. That is, the incidents SEATTLE POLICE DEPARTMENT 26 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 27 of 124 DISPARITY REVIEW – PART II involved a call where the individual was known to, or alleged to, have a weapon, the officer could see a weapon, or the incident began as a high-risk encounter (fleeing vehicle or certain types of warrant services). In these incidents, it is essential for the safety of the general public, and the officers, that a firearm be at the ready. SPD will review if these types of incidents require the firearm to be targeted at a person from the outset, as opposed to held in a ready position. SPD also will continue to review current and future firearms pointing incidents to better understand if there are instances where the pointing is not clearly in line with safety procedures. c. Addressing Disparity-Associated Issues in Terry and Firearm Pointing Incidents The Seattle Police Department is committed to having a department that reflects the community it serves. Between 2014 and 2018 the department increased its non-White hires from 22% to 36%. As part of its continuing efforts to fully staff the Department with the highest qualified individuals, SPD is pursuing a variety of diverse recruitment and promotional efforts, including those targeted to hire individuals from across Seattle. There were both consistent praise and concern around the verbal style used by officers in the incidents. In some, the officers were professional and calm, while in others officers cursed or were seen as overly forceful, disrespectful, and escalating. SPD is committed to reinforcing the training it currently provides to all officers around LEED (Listen, Explain, Equity, Dignity) through a its ongoing in-service trainings. The City of Seattle is home to one of the most diverse – in terms of languages spoken – zip codes in the country. Accordingly, the city has access to language line and other translation services. The policing service, however, is not an easy deployment of the language line model. From the call taking to field contacts, the translation services are quick, raw, and often stressful. Community members had questions around how can someone accept a Miranda notice when they cannot clearly understand English – as seemed to happen in one incident. This concern also carried over to a discussion on how can subjects who are not English speakers make complaints about bias. SPD explained that in both of these situations – a frisk and a firearms-pointing – a sergeant will be on scene and they are there to take that complaint. SPD is pursuing technological solutions beyond what are currently available through city resources, and the SPD strongly acknowledges the community recommendation that a more diverse police service will allow for more direct face-to-face translation and relationships. One recommendation for continuing these sessions was to ensure that there were voices of people who had been through the system included. As the SPD continues to work with its community connections and the CPC to have more of these sessions, SPD commits to find ways to engage with these community members in a way that is fair, unharmful, and productive. This will require more deliberative planning and the identification and inclusion of resources that can be supportive of all those in the room. SPD will do this work with its partners and anticipates that it will produce additional recommendations for addressing disparity. SEATTLE POLICE DEPARTMENT 27 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 28 of 124 DISPARITY REVIEW – PART II Finally, as noted earlier, community members discussed whether SPD should take action to restore the dignity of individuals following these types of interactions. SPD is committed to the previously mentioned LEED principles, as well as the overall goal of reducing – and hopefully eliminating – disparity. If there are resources that can be deployed on scene, or actions and statements that officers can make to help make individuals feel whole after these interactions, SPD wants to know what those are, and will work to determine how to best include them in operational guidelines. B. Quantitative Findings The SPD already has initiated the work necessary to integrate the PSM methodology into its existing analytic dashboards. In the time since the SPD launched its SeaStat data-driven management program (September 2014) for increasing public safety, the use of these dashboards has become so ingrained in the daily work of supervisors, Since early 2017, SPD has been replicating this success with companion dashboards focused on force, disparity, supervision, and other areas covered under the consent decree. As the PSM methodology, and future statistical work, is built into these dashboards, discussions about disparity and what can be done about it will be as frequent as what can be done to reduce crime. The logistic regressions covered in the technical report reinforced the issues around the out of place phenomenon – both on the part of the community, the call takers/dispatchers, and the officers. The department will dig deeper into the differences across beats and precincts, for possible causes and remedies. CONCLUSION This qualitative review of incident types with an identified disparity rate highlighted several key trends in how the events begin, unfold, and resolve. Some of these trends suggest areas of policy, training, information sharing, or other strategies, that could potentially decrease the disparity. The intent of this qualitative review – complimented by additional quantitative work – was to go beyond the information captured in data fields of department reports. Overall, the feedback from the internal SPD reviews and external community meetings offered the most potential related to the decision to frisk. This discussion focused on how similar an identified subject is to a description provided by a witness/victim. There were concerns around a minimal degree of match, especially if it was related to one common item of clothing or if there had been a significant amount of time since the subject was last seen. The feedback on the pointing of firearms involved a discussion mostly on how the officers interacted with the subjects, in terms of professionalism and de-escalation. In both the internal and external review, almost all the firearm pointing incidents had a preceding factor that legitimized the pointing – a high-risk warrant/stop, a report of a weapon, or an officer viewing a weapon. There were concerns around how tactics for approaching a potentially high-risk vehicle may contribute to a number of frisks and firearms pointing. SEATTLE POLICE DEPARTMENT 28 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 29 of 124 DISPARITY REVIEW – PART II From these discussions, the SPD has committed to a collection of actions: 1. Amplify the training and guidance around how specific a subject’s description must be to initiate a stop, and safety frisk, if warranted a. Examine guidance on the relationship between the amount of time since the subject was seen and the specificity of the description b. Examine how an allegation of the suspect possessing a weapon is communicated c. Examine the circumstances under which the witness/victim gave the description 2. Review policies, trainings, and protocols for the pointing of firearms a. Review the classifications in the department’s risk-assessment matrix, used for serving high-risk warrants b. Review procedures around how to initially engage and approach vehicles during high-risk stops c. Review procedures on warrant services d. Emphasize current team tactics training principles, particularly around use of firearms 3. Analyze further the 9-1-1 call taking and dispatching process. Develop enhanced procedures and trainings to reduce the transmission of implicit bias a. Enhance and expand implicit bias training for all 911 Call Center employees b. Review best practices around the transmission of subject descriptions c. Review procedures/policies around not dispatching officers to a call that is about an obviously non-criminal/non-emergency issue, especially when it is apparent an element of the call is someone “being out of place” based on the caller’s statement around their appearance. 4. Address “disparity-associated” issues involving officer professionalism and wellness a. Amplify existing LEED training to ensure officers always are professional and using de-escalation techniques b. Identify and pilot “instant translation” services to better inform and engage all communities c. Continue to diversify the police department d. Identify and establish procedures for positively resolving incidents where there is a formal engagement, particularly those identified as having disparate rates 5. Continue the work on identifying and responding to disparate impacts a. Continue to attempt additional quantitative approaches to better diagnose key factors underlying disparity b. Continue to hold and improve additional community feedback sessions around these issues, including identifying a method for bringing the voices of those who have been through the system into the room SEATTLE POLICE DEPARTMENT 29 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 30 of 124 DISPARITY REVIEW – PART II c. Examine the differences in disparities across precincts/beats, and conduct a temporal trend analysis, as well The Seattle Police Department also commits to an on-going identification and examination of disparity rates in its major forms of interacting with the community, including an analysis of its 2019 stops and detentions and uses of force to validate the 2018 findings. As part of its commitment to ongoing improvement and self-assessment, the SPD has built the propensity score matching (PSM) methodology into its analytic platform (DAP) to make this work a natural part of its performance reviews. The Department also commits to an annual, published review of this work on analyzing and responding to disparity, including updates on implemented strategies to lower disparity and any evidence of their success. This work will include continued community sessions to review incidents – as the Department works with the CPC and community to learn from the pilot sessions conducted during this review and continue to improve the process. SEATTLE POLICE DEPARTMENT 30 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 31 of 124 Appendix A Terry Stop Analysis Charles C. Lanfear 11/23/2019 Abstract This study examines how police officer decisions to conduct investigative stops differ by subject race at the incident level. The primary outcome in these analyses is the incident-level probability that a Terry stop results in the record of an offense, referred to in the literature as a “hit”. Secondary outcomes are alternate measures of hits, such as the probability of a stop resulting in an arrest or of a frisk resulting in recovery of a weapon. Under the assumption similarly situated subjects of different races receive similar dispositions, racial differences in the probability of a hit are indicative of differences in the discretionary threshold to stop or frisk subjects. These analyses use geolocated Seattle Police Department records matched to US Census Bureau American Community Survey (ACS) data to control for local population composition. The results indicate mixed evidence that minority subjects are subject to lower thresholds for Terry stops than white subjects. Hit rates on non-white subjects were associated with officer discretion and were lowest for weapon recoveries in frisks where discretion in the outcome is absent. This finding casts doubt on the assumption underlying standard hit rates that dispositions are unrelated to race for similarly situated subjects. Rather, these results are consistent with differential treatment in stop disposition and the decision to frisk. Introduction The literature on differential treatment in discretionary police stops is well established (see Neil and Winship 2019). The situational circumstances surrounding Terry stops have been explored in the literature since the Terry v. Ohio case in 1968 and Floyd et al. v. City of New York case in 2013. One common goal of many of these studies is to characterize the relationship between stop circumstances and the discovery of weapon or evidence of association with a crime, i.e. a ‘hit’. This analysis furthers the police stop literature by evaluating the situational outcomes of stops, arrests, frisks and weapon recoveries at the incident level in Seattle, Washington. More specifically, the focus of this project is to examine how police officer decisions to conduct investigative stops differ by subject race at the incident level. Stop Outcomes This report focuses specifically on two possible stages within a Terry stop: (1) the decision to stop an individual and the subsequent disposition of that stop, and (2) the decision to frisk a stopped individual and the resulting probability of recovering a weapon. Within these stages, officers exert discretion at three primary points: (1) the initial decision to stop, (2) the decision to frisk, and (3) the decision of a disposition (see Figure 1 on page 11). Officers have little discretion in recovering a weapon from a frisk—it is either present or not. The hit rate is a function of the decision to stop and the chosen disposition. If, conditional on evidence, officers assign identical dispositions by race, the hit rate is an indicator of the validity of the decision to stop individuals. In that case, higher hit rates are indicative of more warranted stops. If, however, officers apply more punitive outcomes to minority subjects, the observed hit rate will be artificially inflated for those groups—but only where officers possess more discretion to do so (e.g. during the decision to arrest rather than in the recovery of a weapon). In this case, higher hit rates could instead be indicative only of an increased likelihood to recommend charges for those subjects. A common method for evaluating racial disparities in policing is a hit rate test comparing the rate of contraband discovery or arrest after individuals have been stopped and searched (Neil & Winship 2019). Low 1 Full Report Page Number 31 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 32 of 124 hit rates imply that stops of individuals were unproductive, i.e. did not lead to actionable outcomes such as recovering illicit substances, illegal weapons, or an arrest. The results of a hit rate test can be used as an indicator of racial discrimination if there are substantial differences in hit rates between racial groups. These differences may indicate that the threshold for stopping individuals from one group is lower, supported with less evidence, or possibly motivated by some form of bias. Although hit-rate tests provide a readily interpretable way of examining racial disparities in policing, this approach can mask important points in the process where vulnerable populations are treated differently. Hit rate analyses are a valid method for measuring racial discrimination insofar as individuals from all racial groups are treated equally after a stop occurs. Many studies of police decisions to arrest suggest otherwise (Smith & Visher 1981; Brown 2005; Ousey & Lee 2008). Recent research in Seattle suggests that black individuals have an increased likelihood of arrest when white individuals complain to police and when a suspect is arrested in a changing neighborhood (Lanfear, Beach, & Thomas 2018). Additionally, similar hit rate results may not be indicative of equal treatment: it is possible for minorities to have the same hit rate as whites if they are stopped frequently with a low threshold and the police write reports for highly discretionary offenses such as resisting arrest. Weapons recovery is another way discretionary stops can be evaluated. One strength of this measure is that the presence or absence of a weapon is less susceptible to misused police discretion. Unlike arrests, which are in part discretionary decisions made by police officers, the recovery of a weapon requires a weapon to be present. Therefore, weapon recovery could be considered a stricter test of the accuracy of police stops than arrests. While this method provides a strong test of bias in police stops, it is used less frequently than other approaches. This may be due to data limitations; for a reliable and robust examination of individual-level factors and weapons recovery, a stop and outcome dataset would need to contain enough cases to account for this relatively rare event. An analysis of this form also limits the scope to incidents with frisks, which may have unique patterns of differential treatment not mirrored in other types of stops. One example of a frisk hit test was performed by Goel et al. (2016) using a novel computational method on stop-level data. The research team used two years of data to train a machine learning classifier to predict the ex-ante probability of weapon searches turning up an illegal firearm—that is, the probability a concealed weapon is found conditional only on information available to officers prior to the stop. The authors then predicted these probabilities for later stops and compared racial differences. They found that in stops of minorities there was a much lower probability of a gun being found, signifying officers used lower thresholds of evidence in their decisions to stop. While they found racial disparities in stops, a substantial portion of this variation was attributed to officers using a lower evidentiary threshold in high crime areas which have large minority populations. Incident-level Approach In addition to methodological challenges of hit rate analyses, a primary obstacle in measuring racial disparities in police stops is most researchers’ inability to specify a risk pool. When analyzing summed police stops at the beat-, precinct-, or city-level, researchers are interested in identifying the disproportionate risk of an individual being stopped relative to the racial composition of other people on the street and their underlying involvement in criminal behavior. Collecting information on this risk pool is rarely possible due to differences in the use of public space, the flow of individuals throughout the city, and the availability of administrative data (Neil & Winship 2019; Knox, Lowe, & Mummolo 2019). Similarly, because most data on criminal behavior is available only through police records, it is inseparable from the allocation of police resources, the rate of citizen reporting, and possible biases embedded in other areas of policing (Varano et al. 2009). By estimating the aggregate effects of racial composition of residents and contacted individuals on the stop-and-frisk- and hit rate, aggregate analyses potentially suffer from the ecological fallacy. That is, it is possible that statistical relationships detected in aggregated precinct-level data are dissimilar to those at observed at the incident-level of individual stops. The limitation of estimating a risk pool is less consequential when racial disparities are evaluated at the individual level. In these analyses the focus shifts from aggregate patterns of disparities to incidents where, 2 Full Report Page Number 32 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 33 of 124 after being stopped, race is evaluated as a possible factor shaping outcomes for similarly-situated individuals. This, of course, hinges on the assumption that individuals who are similarly situated—in context, behavior, and on individual characteristics—have similar probabilities of being involved in crime and thus should be treated equivalently by the police (Neil & Winship 2019). Identifying equivalently situated individuals is the primary challenge for stop-level tests of differential treatment. This is typically approached, as by Goel et al. (2016), by conditioning on all relevant measurable characteristics of subjects, their actions, and their surroundings. One way of accounting for situational differences during stops is controlling for individual-level variables that may differ between incidents. This analysis considers a wide-range of incident factors, such as subject characteristics, officer knowledge, stop context, officer characteristics, and local context. The following section reviews relevant literature in each of these areas. Subject Characteristics Scholars who study racial disparate policing understandably focus on how policing decisions and outcomes differ by various subject characteristics, such as race, gender, and age. In Boston, Fagan et al. (2016: 540) find that black individuals, after controlling for criminal history and gang affiliation, are more likely to be observed, questioned, and frisked than other individuals. These findings are echoed in studies by Close and Mason (2007) and Gelman, Fagan, and Kiss (2007) on initial stop and searches, Levchak (2017) and Coviello and Persico (2015) on frisks, and a multitude of studies on arrests (Baumgartner et al. 2017; Kochel et al. 2011; Lytle 2014; Mitchell & Caudy 2015). These findings are particularly pronounced for black teens, as Crutchfield et al. (2012) find that they are twice as likely to have contact with the police than other individuals with similar criminal histories. Levchak (2017) finds that men are more likely to be searched and frisked than women and Fagan et al. (2016) find that men are much more frequently the subject of a stop and search. While these gender differences may be congruent with gender gaps in offending more generally (Lauritsen, Heimer, & Lynch 2009), they also may be influenced by male officers’ reluctance to stop and frisk women (Chanin & Rojo-Mendoza 2019). Finally, young individuals are more likely to be stopped than older individuals (Levchak 2017: 389; Fagan et al. 2016). Officer Knowledge Terry stops may also be shaped by the knowledge an officer has prior to initiating contact with an individual. This information is likely comes from two sources: previous interactions with the person and/or receiving a suspect description. Prior research suggests that previous contact shapes future contacts with the same individual. During traffic stops Higgins, Vito, and Walsh (2008) find that individuals known to the police are more likely to be stopped and searched; Engel and Silver (2001) find that police arrest individuals “known to the police” at higher rates than individuals unknown to the police, and if an individual has a reputation of carrying a weapon, there is an increased likelihood police will use force during a stop (Garner et al. 2002). It is possible that these interactions are also altered by an officer’s knowledge of an individual’s criminal history. Tilyer (2014) found that officer knowledge of a subject’s criminal history increased the likelihood of a discretionary stop and partially explained the increased rate of stops for black men. The second source of officer knowledge may come from information provided by a dispatch operator. As officers are dispatched to find and stop an individual, descriptive information may be relayed from complaining citizens or witnesses via the dispatcher. Although it is impossible for the dispatcher to verify if this information is complete, actionable, or accurate, it is then passed along to officers to aid in their search. Little research has considered the differences in outcomes between dispatched and on view Terry stops. Stop Context In addition to officer knowledge, stop context may shape an officer’s likelihood of initiating a stop and/or their decision making during a stop. The impetus for a stop may come from two sources: police receive information from a dispatch operator and are directed to an area of the city or officers observe behavior in the field, 3 Full Report Page Number 33 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 34 of 124 i.e. conduct an ‘on-view’ stop. In stops that begin with a dispatch, officers may have additional information, a subject description, or a heightened sense of urgency. On the other hand, dispatch information may be provided by local residents or complainants who provide biased, incomplete, or information determined to be unimportant to police. The possible relationship between a caller’s suspicion and an officer initiating a stop can only be analyzed if this pre-arrest, pre-contact rationale is systematically documented; in most cases, it is not. On-view stops are more reliant on an officer’s interpretation of suspicious behavior, rather than information from the general public or specific guidance from a dispatcher. It is possible that spatial context, such as the racial and socioeconomic demographics of a neighborhood, is particularly influential in officer’s perceptions of disorder (Sampson & Raudenbush 2004) and suspicion of individuals. Documentation of officer behavior may also change the eventual outcome of a police stop. While officers self-report that body-worn cameras (BWC) have little impact on their behavior (Grossmith et al. 2015), evidence from experiments evaluating the influence of this technology on stop and searches offers mixed results. Two experimental trials find that officers wearing body cameras are less likely to stop people (Peterson et al 2018; Ready & Young 2015), perform an arrest (Ready & Young 2015), or receive citizen complaints (Peterson et al. 2018). On the contrary, Ready and Young (2015) find that officers with body cameras were more likely to initiate citizen contacts and issue citations, whereas Grossmith et al. (2015) find that BWCs have no impact on the number or type of stops performed by officers. It is possible that dash cameras have a similar impact on officer behavior, however the literature on discretionary police stops does not evaluate this factor. Finally, the number of officers present during a call may alter the dynamics within a stop and search, or possibly decrease or increase the likelihood a stop is initiated in the first place. Officers may wait for backup before initiating a stop (Stoughton 2017), where one officer performs the stop, question, and frisk and the other officer observes the surrounding area (Albrecht & Morrison 1992). Officers may also discuss their motivations for initiating a stop amongst themselves—one can imagine that this could either increase or decrease the likelihood of a stop. Officer Characteristics Officer characteristics, such as length of employment in the department and age, may also influence the outcome of a stop. Alpert, Dunham and Smith (2007) find that more experienced officers have an increased likelihood of conducting a search of the driver during a traffic stop than less-experienced officers, however officer age was a non-significant predictor of the same outcome. Work by Smith and Petrocelli (2001:14) shows a significant relationship between officer age and the driver race, with older officers more likely to stop white drivers and younger officers more likely to pull over drivers who are black or Latinx. However they did not find further evidence that officer age shapes the decision to perform a consent search or issue a legal sanction (Smith & Petrocelli 2001). Furthermore, an officer’s previous military experience may alter their perception of citizen behavior and influence their discretionary decision making during a stop. However, the review conducted here found no previous studies that evaluate this possible relationship. Additionally, an officer’s predisposition during stops may be shaped by their voluntary participation in Crisis Intervention Training (CIT), which focuses on de-escalating interactions with people experiencing mental health crises (Seattle Police Department 2019). Furthermore, because this training is optional, it may represent a selection process, whereby the officers who choose to opt into the training program may also react to stop and searches in a systematically different way that officers who chose to forego the training entirely. Local Context Scholars examining racial disparities in stop and frisk have paid particular attention to neighborhood racial composition as a possible factor shaping these discretionary interactions. For example, Gelman et al. (2007) analyzed counts of stop-and-frisks and hit rates by race, adjusting for the underlying number of all arrests by race in each precinct and modeling each class of crimes separately. They found support for elevated 4 Full Report Page Number 34 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 35 of 124 stop-and-frisks and lower hit rates in precincts with large minority populations—evidence for differential policing by precinct—and also in general across minority subjects. One possible explanation for these findings is that officers’ observations of their surroundings may influence their perceptions of individuals behaviors; Smith (1986) found that, after controlling for behavior and the type of crime, suspects in poor neighborhoods were more likely to be arrested. Gordon furthers this argument with her finding “that police drew upon symbolic ideas that emphasized the violence of black neighborhoods and the economic value of white neighborhoods in developing local strategies. As they acted in relation to these distinctions, the police amplified disparities in service provision and social control, consolidating the character of an already segregated and unequal landscape” (2019: 1). Neighborhood racial context may also interact with an individual’s race, causing officers to particularly notice when there is a mismatch between the two. Officers may believe that certain individuals “belong” in specific neighborhoods and increase their social control against individuals who deviate from this pattern (Novak & Chamlin 2012; Stewart et al. 2009). This form of racial profiling is commonly referred to as “out-of-placeness,” “out-of-place policing,” or “racial incongruity” (Brunson & Weitzer 2009; Fagan & Davies 2000; Novak & Chamlin 2012; Steward et al. 2009). Individuals who are “out of place” may be perceived as more suspicious (Gaston 2019; Fagan & Davies 2000) or possibly in the area for nefarious purposes (Meehan & Ponder 2002). Conversely, people “out of place” may be easier to identify for officers who have been dispatched with a specific subject description. As with any policing activity, Terry stops present an analytical challenge to researchers as well as a complex, multifaceted, and fraught set of interactions between officers and community members. The above discussion outlines several methodological and data challenges of examining Terry stops. It should also be noted that while careful thought has gone into the construction of the following descriptive statistics and statistical models, this exercise can only illuminate one type of knowledge on the topic. Individuals who are ‘objects of suspicion’ (Tyler, Jackson, & Mentovich 2015) are likely affected by these stops in ways the current analysis is unable to measure or evaluate. The conclusions reached in this analysis should be paired with other forms of knowledge and definitions of justice, to expand our basic knowledge of these particular police-community interactions and help better inform policing future policing practices. In the following sections, the data and methods used to evaluate possible racial disparities in police stops, frisks, arrests, and weapons recoveries are outlined. Data This study uses a combination of Seattle Police Department (SPD) records and US Census Bureau American Community Survey (ACS) measures. Terry stop data consists of records from stop templates filled out by SPD officers. These templates contain information on subject characteristics, the context of the stop, and its disposition. For the purpose of this report, stops with dispositions other than a Field Contact—in which an officer does not record an offense or sanction the subject—are considered hits. This includes disposition categories of Arrest, Citation, Referral for Prosecution, and Offense Report. These Terry stop templates were matched, where possible, to records from the Computer-Aided Dispatch (CAD) system and Records Management System (RMS). Officers must manually link a Terry stop to the CAD system for the template to be linked to CAD or RMS data in SPD records. In cases where Terry stops do not result in generation of an offense report in the RMS system (e.g. no evidence of an offense assigned by the police), officers are less likely to take additional time to link the event to the relevant CAD entry. This means that data available for both Terry stops that result in hits and those that do not are limited to what is reported on stop templates or can be linked to stop templates—CAD and RMS data are not available for all stops. This prevents, for instance, using initial call type (e.g. assault) as a predictor of hit rates. This is only problematic for estimating subject race hit rates if particular call types with much higher or lower hit rates are concentrated in particular subject race categories. There is no way to evaluate if this is the case. The location of each stop was used to match the stop to the Census block group 1 Block 1 and SPD beat it occurred groups are the smallest unit for which reliable race/ethnicity data were available. 5 Full Report Page Number 35 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 36 of 124 in. At the block group level, 5-year ACS data (end year 2017) were joined to stops to produce measures of the local context in which the stop occurred. Beats were used as fixed effects, which are further described in the methods section below. Data on officers were also joined to other incident information using officer identification numbers on the Terry stop templates. Analyzed stops were restricted to 24,021 incidents that were entered in the system between January 1, 2016 and December 31, 2018. The stops analyzed here encompass 51 SPD beats and 473 census block groups. One limitation of these data is that the locations of stops not linked to CAD or RMS data are recorded using only approximate addresses or cross-streets. Those linked to CAD or RMS data include coordinates but these coordinates are associated with the original CAD or RMS report location and may not represent where a subject was actually stopped. Consequently, recorded text addresses from the Terry templates were used in all cases. These text addresses were converted into latitude and longitude coordinates by geocoding through the Google Maps API. All addresses which the API assigned outside Seattle city limits were manually verified and an additional random sample within Seattle was taken to compare coordinates to written text. This process yielded highly accurate results, with over 90% of sampled points lying in what the author determined was the correct location. The remaining points were typically nearby or did not uniquely identify any clear location (e.g. “I-5 Southbound” or “Broadway”). Table 1 lists the measures used in the tables and models below. The table is divided into Outcomes–the results of Terry stops–and sets of measures which may be related to these outcomes. These include Subject Characteristics, Officer Knowledge, Stop Context, Officer Characteristics, and Local Context. The table defines the operationalization of each measure in these sets. The Out of Place measure is defined here due to complexity. It is possible that Terry stops may be conducted more frequently on individuals an officer perceives to be out of place in an area—for example, a black subject in a predominantly white neighborhood. To capture this, the Out of Place measure is calculated as 1 minus the proportion of the block group population of the race to which the subject belongs. That is, a black subject in a neighborhood that is 10% black would receive a score of .90, while a white subject in a neighborhood that is 80% white would receive a .20. This table excludes many additional measures which were obtained or calculated but had small and imprecise estimated relationships which did not improve model fit. Appendix Table 1 lists these excluded measures. Throughout this document, where referring to a measure specifically, the word or phrase is capitalized and italicized (e.g. Out of Place). This is done to emphasize that the term refers to a specific measure rather than its colloquial usage. Two final notes must be made regarding the data. First, some measures may be unreliable due to how the original data were entered. Data on Terry templates are recorded directly by responding officers after the stop is completed. Stop information may not be recorded immediately afterwards, particularly when a stop does not result in a disposition of an offense. This delay may cause officers to recall details of the stop inaccurately. Similarly, specific fields may not be reliably recorded due to officer time constraints, inattention, or perceived unimportance. It is also possible officers record data in a way that minimizes evidence of bias or retrospectively justifies unwarranted stops. Because of the possible unreliability of these administrative data, statistical associations between measures and outcomes should be evaluated with caution. 239 observations were removed due to missingness missingness on key measures such as subject race, gender, or age. 467 observations with locations more than 500 meters outside Seattle city limits were also removed. Those within 500 meters of city limits were assigned to the nearest beat and block group. Together, this results in a final subsample of 23,305 stops. 6 Full Report Page Number 36 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 37 of 124 Table 1. Measure Definitions Definition Outcome Any Hit Arrest Hit Frisked Frisk Hit Stop resulted in offense report, arrest, citation, or referral for prosecution. Stop resulted in an arrest. Strict hit definition. Stop featured a frisk for weapon or contraband. Frisk resulted in a recovered weapon. Subject Characteristics Subject Race Subject Gender Subject Age Officer perceived race of stopped subject. Officer perceived gender of stopped subject. Officer perceived age of stopped subject. Officer Knowledge Subject Known Subject Description Officer reported prior knowledge of subject. Officer reported receiving prior description of subject. Stop Context Dispatched Body Cam Active Car Cam Active Officers Present Stop was the result of a dispatch call, as opposed to on view. Officer worn body camera active during stop. Officer vehicle camera active during stop. Number of additional officers present during stop. Officer Characteristics Officer Experience Officer Age Officer Gender Officer Race Military Experience CIT Trained Years officer employed by SPD at stop date. Age of officer at stop date. Recorded gender of officer at stop date. Recorded race of officer at stop date., Past military experience recorded. Officer was CIT trained before stop date. Local Context Out of Place Percent Asian Percent Black Percent Latinx Percent Other Out of place indicator. See text above. Block group percentage Asian. Block group percentage Black / African American. Block group percentage Latinx. Block group percentage Other race. 7 Full Report Page Number 37 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 38 of 124 Descriptive Statistics Table 2 depicts counts of Terry stops and hits by subject race. As an example, the Hit Rate of 0.626 for White subjects indicates that an offense was recorded in 62.6% of Terry Stops of White subjects. Hit rates are fairly uniform across subject race categories. White subjects experience the highest volume of stops and a low hit rate. Black subjects experience the next highest frequency of stops and have a higher hit rate, comparable to American Indian / Alaskan Native subjects. The Unknown / Other / Multiracial category features the lowest hit rate. Table 2. Terry Stops by Subject Race Hits Total Stops White Black Latinx American Indian / Alaskan Native Asian Unknown / Other / Multiracial 7408 4869 732 515 421 1118 11823 7145 1131 751 638 1817 Total 15063 23305 Hit Rate 0.627 0.681 0.647 0.686 0.660 0.615 Table 3 depicts counts of Terry stops and hits by contact type—that is, whether the stop was the result of a Dispatched call, an On View by the officer, or an Unknown call initiation. Hit rates are slightly lower for on view stops than Dispatched stops. More importantly, we see that the vast majority of stops which did not result in hits are classified as Unknown contact type. This suggests that officers only record Contact Type consistently for stops resulting in a hit. Consequently, it is impossible to evaluate how Contact Type relates to hit rates–this measure cannot be used in the models below. This is a notable limitation as hit rates likely differ between On View and Dispatched stops. Table 3. Terry Stops by Contact Type Hits Total Stops Dispatched On View Unknown 10800 3416 847 11380 3655 8270 Total 15063 23305 Hit Rate 0.949 0.935 0.102 Table 4 tabulates the frequency of Terry stops by disposition. Note, for the purpose of this report, a Terry stop is defined as a hit when it results in any outcome other than field contact which indicates no offense was recorded. Arrest is also used for a strict measure of hits for comparison. Approximately a third of all stops result only in a field contact being recorded, meaning a hit is seen in roughly two-thirds of incidents. On a hit standard of arrest, the rate is 24%. Table 4. Terry Stops by Disposition Total Stops Arrest Referred for Prosecution Citation / Infraction Offense Report Field Contact 5651 480 87 8845 8242 Total 23305 % of Stops 0.242 0.021 0.004 0.380 0.354 8 Full Report Page Number 38 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 39 of 124 Table 5 provides a cross-tabulation of Subject Race and Disposition. Cells contain proportions of dispositions for stops within each racial/ethnic category. Note that this table indicates White subjects are less likely to face arrests than most minority groups and more likely to receive only a field contact—that is to be stopped without a sanction or recorded offense. Table 5. Terry Stop Dispositions by Race Arrest Referred for Prosecution Citation / Infraction Offense Report Field Contact White AIAN Asian Black Latinx UOM 0.230 0.022 0.003 0.371 0.373 0.277 0.025 0.005 0.378 0.314 0.274 0.011 0.009 0.365 0.340 0.264 0.020 0.004 0.394 0.319 0.235 0.018 0.005 0.389 0.353 0.217 0.015 0.005 0.379 0.385 Table 6 provides a cross-tabulation of Terry stops by contact type and whether the officer reported having a Subject Description prior to the stop. While descriptions are much more common for Dispatched calls, descriptions were reported for over twenty percent of On View stops. It is uncertain where descriptions are obtained in the course of On View calls or if these may be indicative of errors in recording Terry template data. Note the percentage of stops with a Subject Description for Unknown contact types sits between On View and Dispatched. This suggests Unknown contact type stops are comprised of a mix of Dispatched and On View calls. Table 6. Terry Stops by Contact Type and Subject Description Subject Descriptions (N) Total Stops On View Dispatched Unknown 985 9513 3972 3655 11380 8270 Total 14470 23305 Subject Descriptions (%) 0.269 0.836 0.480 Table 7 depicts counts of frisks and total stops and the resulting rate of frisks by Subject Race. Here we see White subjects are subject to the lowest rate of frisks while Asian, Black, and Latinx subjects experience similarly elevated frisk rates. Table 7. Stops and Weapon Frisks by Subject Race Frisks Total Stops White American Indian / Alaskan Native Asian Black Latinx Unknown / Other / Multiracial 2144 163 171 1850 292 385 11823 751 638 7145 1131 1817 Total 5005 23305 Frisk Rate 0.181 0.217 0.268 0.259 0.258 0.212 9 Full Report Page Number 39 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 40 of 124 Finally, Table 8 tabulates Frisk Hits–recovery of a weapon–and total Frisks and the associated Frisk Hit Rate. The hit rate is highest for White subjects and lowest for Unknown / Other / Multiracial. All minority subjects experience notably lower Frisk Hit rates than White subjects. Table 8. Weapon Frisks and Hits by Subject Race Hits Total Frisks White American Indian / Alaskan Native Asian Black Latinx Unknown / Other / Multiracial 410 27 25 244 44 46 2144 163 171 1850 292 385 Total 796 5005 Frisk Hit Rate 0.191 0.166 0.146 0.132 0.151 0.119 Methods The analyses in this report use fixed effects logistic regression models to estimate the relationship between the measures and outcomes described above. Logistic regression models are an appropriate method for modeling binary outcomes (the outcome either occurs or does not) such as whether or not a hit is observed in a Terry stop. This method models the effect of a linear combination of predictors on the logarithm of the odds of the outcome (hit) occurring. These models include fixed effects for beats. Fixed effects adjust for the beat-level average of the outcome, that is, they account for the base hit rate common to all stops in the beat. Using a fixed-effects approach addresses the statistical assumption of independence of errors between observations, which is likely violated if beats share similar police personnel, enforcement strategies, and/or resident populations. Fixed effects models are also a more conservative approach than random effects models, which assume idiosyncratic variation within groups (beats) is uncorrelated with the incident-level covariates included in the model. That assumption is unlikely to hold if differences between beats are attributable to unmeasured differences in officers, officer behavior, or local context. Violation of this assumption would result in incorrect and misleading estimates. Note the original proposal for this analysis featured a structural equation modeling (SEM) approach to estimate differential treatment in the decision to stop; see Appendix 1 for a discussion of the SEM approach. Model estimates displayed in tables 9 and 10 are in odds ratios, and should be interpreted as the multiplicative change in the odds of the outcome for a one unit change in the measure. Odds are defined as the probability P r(Outcome) of the event occurring divided by the probability of it not occurring: Odds = 1−P r(Outcome) . For example, an odds ratio of 1.2 indicates the odds of observing the outcome increases by 20% for each unit increase of the measure. All models control for month of year and hour of day. For all continuous measures (e.g. Percent Other or Out of Place), generalized additive models (GAM) with splines were used to determine the appropriate formula to relate the measure to the outcome. GAMs are a common method for evaluating the shape of the relationship that relates a predictor to an outcome. All key predictors were linearly related to the outcomes—that is they required no transformation. Nonlinear transformations derived from the GAM were used for the month (logarithm in hit models; quadratic in frisk models) and hour controls (cubic in hit models; quadratic in frisk models). Standard errors were generated through bootstrapping in tabular results and simulation in plots were used to address potential violations of error assumptions in the model. These standard errors did not differ notably from conventional estimates. It is important to note that one should not interpret model findings in terms of some hypothetical population the data were sampled from, as confidence intervals might imply. Consequently, 95% confidence intervals are depicted in the output provided here to assess the relative precision of model estimates, but should not be treated as a “hard” criterion for separating relevant from irrelevant results. Note the inclusion of fixed effects impacts the interpretation of estimates. The fixed effect estimate for a beat represents the average level of the outcome (e.g. probability of Any Hit) for all stops occurring in that 10 Full Report Page Number 40 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 41 of 124 beat. Estimates in the model are interpreted as differences from this mean within the beat. For example, this means that the estimates from Local Context measures, such as Percent Latinx, are relative to the beat’s average Percent Latinx. An odds ratio for Any Hit of 2.0 for Percent Latinx would indicates that the odds of a stop resulting in Any Hit are twice as high in a block groups which have a 1 percentage point larger Latinx population than the average block group in that beat. If every block group in a beat has the same level of Latinx, regardless of its value, it contributes nothing to the hit rate, because no block group is higher or lower than the average for the beat. All possible effects of the overall beat-level population composition are thus contained in the fixed effect estimate. The presence of fixed effects also explains the omission of crime rates, as fixed effects inherently control for the underlying beat-level crime rate. Figure 1 is a simplified diagram of the discretionary process for Terry stops which helps visualize what is being estimated in each model. The decision to stop is a product of pre-stop factors such as whether the race of subjects. Individuals who were not stopped remain unrecorded in this data, therefore the relationship between subject race and the decision to stop cannot be directly measured. Instead, I first model the differences in discretionary disposition of subjects—what officers decide to do following the stop. In this case, I consider whether any offense was recorded (Any Hit) or whether an arrest (Arrest Hit) occurred. If the likelihood of officers recording and offense or arresting subjects is conditional only on evidence a crime occurred, differences in these hit rates by race are indicative of lower thresholds to stop (and thus bias). If, however, there is bias in the discretionary assignment of dispositions, these hit rates will not accurately capture bias in the decision to stop. The decision to frisk during a Terry stop is similarly discretionary. However, there is no officer discretion involved in whether a weapon is found following the frisk, under the assumptions that found weapons are recorded equally across subject race categories. Because no discretion is involved here, it provides an alternate hit rate measure less likely to be contaminated by bias in stop resolutions. The disadvantage of weapon recovery hit rates is that findings are limited to the narrower case of frisks and found weapons are not necessarily indicative of a crime—it is possible the subject legally possesses the weapon. Regardless of this, if some subjects are more likely to be frisked but weapons are less likely to be found, it indicates disproportionate application of frisks to those subjects. Hit Probability Models The first model in Table 9—Any Hit—models the likelihood that a Terry stop results in a hit, as defined above. The likelihood of a hit is predicted by all measures shown in Table 1. This model also includes an interaction term between Subject Description and Out of Place. This interaction is included because individuals who are out of place may be easier to match to a provided description, and, conversely, a provided description may result in more false positives when it matches many people in an area. Note that due to the fact that Seattle is a predominantly white city, most racial minorities who are stopped are likely to be relatively high on this measurement. For Subject Race, estimates are differences from the baseline of White subject. Subject Gender is the difference from female subjects. Subject Age is compared to 1 to 17. The second model, Arrest Hit, models the likelihood that a stop results in an arrest by the officer. This is a more stringent definition of a hit as noted above and is provided as a comparison to the Any Hit model. After adjusting for all additional measures and controls, the likelihood of Any Hit for all minority subjects is modestly higher than for whites (except for individuals who are Unknown / Other / Multiracial). Hit 11 Full Report Page Number 41 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 42 of 124 rates are comparable across all age groups above 18 years of age, but much lower than the 1 to 17 year old reference group. Male subjects experience moderately lower hit rates while those with Unknown reported gender are not statistically distinguishable from the female reference category, though this is likely due to small numbers. Stops reported to have a Subject Known to the officer have substantially higher hit rates. All elements of stop context are associated with hit rates: Active Body Cam and Car Cam and additional Officers Present predict hits. Due to the presence of interaction terms, the base effect of Out of Place represents the association between being out of place with a hit when no Subject Description is provided. This effect of Out of Place is relatively strong. For officer characteristics, CIT Trained officers experience higher hit rates. Officer Age and Officer Experience are related to hits in opposite directions, indicating hit rates are lower for older but less experienced officers compared to officers who are younger but more experienced. Male officers display modestly higher hit rates. Past Officer Military experience is associated with lower hit rates. Minority officer race is generally associated with lower hit rates. For local context, only Proportion Latinx is associated with hit rates and the relationship is positive–this effect is more modest than the odds ratio suggests, as the Latinx population in Seattle is small and does not have large variation within beats. Lastly, the interaction term between Out of Place and Subject Description is positive. These indicate that hit rates are higher for Out of Place subjects when the officer reports having a Subject Description. The second model, Arrest Hit, displays largely similar results as the Any Hit model, though estimates are typically smaller. Subject racial categories, for instance, display significantly higher hit rates only for Asian and American Indian / Alaskan Native subjects as compared to the White reference group. Age estimates are inverted from the first model here: All age categories display higher likelihood of an arrest than juveniles, which is indicative of their different status. Note no interaction was found between subject age and race or other primary measures of interest, so this differential treatment appears similar across different contexts and other characteristics; all youth are less likely to be arrested. Male gender does not exhibit lower arrest hit rates. Population composition, mainly Proportion Other, has a moderately stronger relationship with Arrest Hit than Any Hit, however this effect is substantively weak given the small variation in these values within beats. Lastly, the interaction terms remain positively related but are weaker and less statistically relevant than in the Any Hit model. 12 Full Report Page Number 42 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 43 of 124 Table 9. Hit Probability Models Odds ratios from fixed effects logistic regression Any Hit1 Estimate Arrest Hit2 95% CI Estimate 95% CI Subject Characteristics Subject Subject Subject Subject Subject Subject Subject Subject Subject Subject Subject Subject Subject Race American Indian / Alaskan Native Race Asian Race Black Race Latinx Race Unknown / Other / Multiracial Age 18 - 25 Age 26 - 35 Age 36 - 45 Age 46 - 55 Age 56+ Age Unknown Gender Male Gender Unknown 1.481 1.333 1.191 1.223 1.043 0.627 0.602 0.563 0.558 0.531 0.512 0.854 0.840 (1.195, (1.085, (1.028, (1.053, (0.894, (0.519, (0.508, (0.478, (0.466, (0.435, (0.394, (0.796, (0.609, 1.799) 1.59) 1.351) 1.473) 1.203) 0.733) 0.692) 0.663) 0.661) 0.657) 0.653) 0.911) 1.141) 1.358 1.303 1.132 1.112 0.979 1.330 1.432 1.395 1.154 1.247 1.053 1.038 1.131 (1.104, 1.675) (1.057, 1.621) (0.97, 1.309) (0.912, 1.38) (0.816, 1.189) (1.108, 1.574) (1.218, 1.666) (1.169, 1.668) (0.968, 1.425) (1.007, 1.557) (0.783, 1.388) (0.958, 1.114) (0.772, 1.574) 1.361 1.232 (1.18, 1.554) (1.124, 1.337) 1.147 1.327 (0.978, 1.337) (1.204, 1.474) 1.315 1.475 1.436 0.591 (1.236, (1.326, (1.411, (0.457, 1.394) 1.643) 1.462) 0.757) 1.248 0.980 1.277 0.654 (1.173, (0.866, (1.252, (0.492, 1.327) 1.096) 1.3) 0.85) 1.315 0.988 1.011 1.274 0.761 0.715 1.044 0.806 0.634 0.977 (1.234, 1.397) (0.983, 0.993) (1.005, 1.017) (1.163, 1.375) (0.71, 0.822) (0.503, 1.024) (0.891, 1.219) (0.693, 0.93) (0.559, 0.719) (0.88, 1.074) 1.385 0.972 1.010 1.180 0.898 0.889 0.811 0.816 0.843 0.871 (1.295, (0.966, (1.004, (1.065, (0.838, (0.568, (0.671, (0.683, (0.721, (0.777, 1.475) 0.978) 1.018) 1.314) 0.982) 1.231) 0.931) 0.966) 0.961) 0.987) 1.143 1.330 2.045 0.765 (0.794, 1.615) (0.91, 1.956) (1.09, 3.649) (0.413, 1.445) 0.986 1.262 1.693 2.900 (0.676, (0.799, (0.946, (1.446, 1.477) 1.864) 3.141) 5.474) 1.902 (1.582, 2.306) 1.482 (1.193, 1.82) Officer Knowledge Subject Description Subject Known Stop Context Body Cam Car Cam Officers Present Out of Place Officer Characteristics CIT Trained Officer Age Officer Experience Officer Gender Male Officer Military Officer Race American Indian / Alaskan Native Officer Race Asian Officer Race Black Officer Race Latinx Officer Race Unknown / Other / Multiracial Local Context Proportion Proportion Proportion Proportion Asian Black Latinx Other Interaction Out of Place x Subject Description 1 2 N = 23,305, Log Likelihood = -13,538 (df=89), beat fixed effects and hour / month coefficients omitted. N = 23,305, Log Likelihood = -12,117 (df=89), beat fixed effects and hour / month coefficients omitted. 13 Full Report Page Number 43 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 44 of 124 Below, selected results are also depicted below graphically for ease of interpretation. Figure 1 depicts the probability a stop is a hit based on Subject Race and Subject Description. These results indicate subject descriptions make a larger difference for the Any Hit outcome than Arrest Hit. Differences in hit rate by Subject Race are similar between both outcomes. These estimates include all uncertainty from the model, therefore differences may still be statistically significant even if confidence intervals overlap. Refer to the tabled and written results above for within-measure comparisons such as differences between subject racial categories. Figure 2. Probability of Hit or Arrest By subject race and description Arrest Hit 0.8 0.8 0.6 0.6 Description Provided NO O M YES U As ia n Bl ac k La tin x AI W hi M O U AI hi W AN 0.2 te 0.2 As ia n Bl ac k La tin x 0.4 AN 0.4 te Hit Probability Any Hit Figure 3 depicts the probability a stop is a hit based on Out of Place and Subject Description. This figure reveals the degree to which the probability of Any Hit declines as subjects without a Description Provided are increasingly Out of Place. Because Seattle’s population is largely white, minority subjects will typically be Out of Place and may be subject to undue scrutiny. Note the relationship disappears or reverses for stops with a Description Provided. This indicates a combination of an Out of Place subject and a useful description may result in more productive stops (i.e. more hits). Alternatively, this relationship would also be observed if, when provided a description Out of Place individuals are unduly subject to more punitive dispositions. 14 Full Report Page Number 44 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 45 of 124 Figure 3. Probability of Hit or Arrest By out of place and subject description Arrest Hit 0.8 0.8 0.6 0.6 Description Provided NO Out of Place 0 YES 1. 0. 0. 1. 0. 0. 5 0.2 0 0.2 0 0.4 5 0.4 0 Hit Probability Any Hit Out of Place Figure 4 depicts the mean probability a stop is a hit within each beat and precinct compared to the city-wide average. Values above the solid line indicate higher hit probabilities, while values below are lower hit probabilities. Bars are 95% confidence intervals. The dashed red line is each precinct’s average hit probability. These estimates are derived from the model so they are adjusted for all measures including population composition. 15 Full Report Page Number 45 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 46 of 124 Figure 4. Difference in Hit Rates from City Average By precinct and beat Any Hit East 0.1 North L1 B2 L2 C3 South L3 B1 U3 E1 N1 E2 B3 R2 E3 G1 C2 C1 0.0 O2 O1 N2 U1 G3 Probability Difference from Mean O3 U2 G2 J1 J3 J2 N3 R1 R3 S2 −0.1 S3 Southwest S1 West 0.1 W3 M3 W2 0.0 K2 Q3 K1 Q2 M1 D2 D1 F1 M2 Q1 W1 K3 D3 F2 −0.1 F3 Arrest Hit East North South E1 0.1 E3 C3 G1 C1 L1 C2 0.0 R2 L2 E2 B2 J3 G3 G2 Probability Difference from Mean R1 N1 L3 B1 B3 J2 U3 N2 O1 N3 O2 S3 S2 O3 J1 U1 S1 U2 R3 −0.1 Southwest West 0.1 W3 M1 Q2 D2 M3 M2 D1 0.0 D3 K3 Q1 W1 −0.1 K1 F1 Q3 K2 W2 F2 F3 16 Full Report Page Number 46 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 47 of 124 Frisk Models Table 10 depicts models for frisks. The first model—Frisk—models the likelihood that a Terry stop includes a frisk for a weapon. This may be seen as modeling the relationship between these measures and propensity to frisk a stopped subject. The second model, Weapon Found, models the likelihood that a stop with a frisk results in the recovery of a weapon. This is an alternate measure of hits which may be more objective than the Any Hit or Arrest Hit outcomes because it is dependent on a weapon actually being present—there is less officer discretion involved in this outcome (see Goel et al. 2016). The frisk models use the same measures as the hit models above except they do not feature the interaction term with Out of Place. Frisks and recoveries of weapons occur after a stop has been initiated where subject-neighborhood interactions are less likely to be influential. Additional models (not shown) verify this interaction is not present for frisk outcomes. Officer race was also removed as it was non-predictive of initiating a frisk or recovering a weapon. All individuals in racial groups other than white experience elevated likelihoods of Frisk during stops, except Unknown / Other / Multiracial. Subject age is not notably related to frisks, while subjects of Male and Unknown gender are much more likely to be frisked than female subjects. A Subject Description but not Subject Known predicts being frisked. In Stop Context, active Car Cam and additional Officers Present predict frisks. Among officer characteristics, only CIT Trained, Officer Gender Male, and Officer Experience weakly positively predict frisks. Local Context is not strongly related to likelihood of frisk. In the model of likelihood of a Frisk Hit (i.e. a weapon is recovered), all racial categories exhibit lower hit likelihoods than White subjects. The relatively small number of frisks for most racial categories reduces statistical power in this case, however, so that only the Black and Unknown / Other / Multiracial categories display confidently lower hit rates than the White reference category. All age groups above the 1 to 17 year reference group exhibit higher hit rates. Weapons are recovered more often from Male but not Unknown gender subjects. Both Subject Description and Subject Known are associated with higher hit rates for frisks. Stops with an active Car Cam and More Officers present are associated with hits, but other Stop Context measures are not notably predictive. No Officer Characteristics are associated with increased weapon recovery. Local Context is also not predictive of weapon recovery with any statistical confidence. 17 Full Report Page Number 47 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 48 of 124 Table 10. Frisk Probability Models Odds ratios from fixed effects logistic regression Frisk1 Estimate Frisk Hit2 95% CI Estimate 95% CI Subject Characteristics Subject Subject Subject Subject Subject Subject Subject Subject Subject Subject Subject Subject Subject Race American Indian / Alaskan Native Race Asian Race Black Race Latinx Race Unknown / Other / Multiracial Age 18 - 25 Age 26 - 35 Age 36 - 45 Age 46 - 55 Age 56+ Age Unknown Gender Male Gender Unknown 1.212 1.302 1.121 1.246 0.996 1.115 1.085 1.036 0.944 0.946 1.349 2.955 2.227 (0.958, 1.473) (1.03, 1.598) (0.958, 1.289) (1.01, 1.561) (0.832, 1.186) (0.95, 1.302) (0.925, 1.274) (0.88, 1.202) (0.789, 1.137) (0.776, 1.156) (1.04, 1.804) (2.682, 3.257) (1.557, 3.103) 0.799 0.868 0.612 0.705 0.632 2.222 2.568 2.432 2.636 2.993 2.631 1.810 0.893 (0.421, 1.332) (0.519, 1.406) (0.457, 0.866) (0.421, 1.236) (0.401, 1.069) (1.42, 3.693) (1.653, 4.319) (1.543, 4.396) (1.562, 4.611) (1.747, 4.987) (1.31, 5.144) (1.396, 2.519) (0.302, 2.504) 1.686 1.049 (1.561, 1.808) (0.918, 1.169) 1.390 1.294 (1.169, 1.685) (0.971, 1.671) 1.003 1.427 1.369 0.996 (0.927, (1.241, (1.342, (0.787, 1.071) 1.653) 1.398) 1.305) 1.122 1.497 1.083 1.243 (0.95, 1.3) (1.07, 2.201) (1.029, 1.129) (0.634, 2.219) 1.118 0.992 1.011 1.229 0.950 (1.038, (0.986, (1.003, (1.107, (0.864, 1.204) 0.997) 1.02) 1.368) 1.038) 0.993 1.003 0.995 0.988 0.891 (0.832, 1.184) (0.986, 1.02) (0.978, 1.012) (0.759, 1.301) (0.74, 1.096) 0.752 0.959 1.205 1.273 (0.492, 1.104) (0.6, 1.447) (0.618, 2.175) (0.669, 2.562) 0.570 0.933 1.358 2.583 (0.212, (0.322, (0.267, (0.515, Officer Knowledge Subject Description Subject Known Stop Context Body Cam Car Cam Officers Present Out of Place Officer Characteristics CIT Trained Officer Age Officer Experience Officer Gender Male Officer Military Local Context Proportion Proportion Proportion Proportion 1 2 Asian Black Latinx Other 1.491) 2.982) 7.484) 14.051) N = 23,305, Log Likelihood = -10,748 (df=83), beat fixed effects and hour / month coefficients omitted. N = 5,005, Log Likelihood = -2,086 (df=83), beat fixed effects and hour / month coefficients omitted. Figure 5 depicts Frisk and Frisk Hit probabilities by subject race. Again, these estimates include all uncertainty from the model—differences may still be statistically significant even if confidence intervals overlap. Overlap is substantial in Frisk Hit due to high uncertainty. The reader should refer to the results above for within-measure comparisons, such as between racial categories. The comparison between Frisk and Frisk Hit probability within each racial category is also notable: Higher likelihood of Frisk for a group is typically accompanied by lower Frisk Hit probability. 18 Full Report Page Number 48 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 49 of 124 Figure 5. Probability of Frisk or Frisk Hit By subject race 0.3 Probability 0.2 Outcome Frisk Frisk Hit 0.1 M O U in x La t ac k Bl As ia n AN AI W hi te 0.0 Figure 6 depicts Frisk and Frisk Hit rates by precinct and beat. Again, these are deviations from the grand mean of each outcome. Noteworthy here is the much higher frisk rate in the South precint, which is not accompanied by a substantially lower hit rate. While frisks are substantially more common in this precinct, weapon recovery rates per frisk are comparable to other precincts. 19 Full Report Page Number 49 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 50 of 124 Figure 6. Difference in Frisk / Frisk Hit Rates from City Average By precinct and beat Frisk Rate East North South 0.2 O3 R3 R2 R1 S2 S3 O1 S1 0.1 O2 L1 Probability Difference from Mean G3 G2 G1 0.0 B2 C3 E3 J3 C2 E2 E1 J1 B1 N1 N2 C1 U2 U3 U1 J2 L3 B3 N3 L2 −0.1 Southwest West 0.2 0.1 F1 K3 W1 0.0 W3 K2 K1 Q3 Q2 D1 F2 −0.1 D2 D3 M1 M3 W2 F3 M2 Q1 Frisk Hit Rate East North South 0.3 0.2 0.1 C3 U2 B1 U3 L2 N3 Probability Difference from Mean G3 0.0 E1 O2 N1 L1 R1 B2 N2 E2 G2 S3 O1 O3 U1 B3 E3 J3 C2 C1 J1 R3 J2 −0.1 R2 S2 S1 G1 L3 −0.2 Southwest West 0.3 0.2 M1 K1 0.1 D2 K2 F2 M3 D1 M2 Q3 W1 0.0 Q2 W3 F3 F1 W2 Q1 K3 D3 −0.1 −0.2 20 Full Report Page Number 50 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 51 of 124 Discussion The primary conclusion of this analysis is that there is mixed evidence indicating minority subjects are subject to lower thresholds for Terry stops than white subjects. Overall hit rates for non-white racial categories were generally similar or higher than whites, both in raw numbers and once conditioned on all available measures of subjects, officers, stop context, and local context. For the decision to conduct a stop, the most consequential element of subject race appears to be the difference between subject race and local population composition. That is, subjects who are members of a racial group that comprises a smaller proportion of the immediate area are more likely to be subject to stop. If these subjects are generally participating in crime, this may indicate being out of place makes it easier to discern them from the background population (i.e. it increases true hit rate). Conversely, it could be due to officers choosing more punitive dispositions for individuals who are out of place when they are provided more justification by subject description (i.e. it increases differential disposition). Conversely, with regard to frisks, there is evidence that minority subjects are disproportionately subject to weapon frisks which contributes to lower hit rates. These differences were not attributable to variation in the underlying population composition, reported crime rate of block groups or beats, or other available measures of subjects, officers, and stops. This finding casts doubt on the assumption underlying standard hit rates that dispositions are unrelated to race for similarly situated subjects. To reiterate the statement in the introduction, in the Terry stops under examination, officers may exert discretion at three primary points: (1) the decision to stop in the first place, which we cannot observe directly, (2) the decision to frisk during a stop, and (3) the decision of a disposition which resolves the stop. Officers have little discretion in recovering a weapon from a frisk. The hit rate for stops is a function of the decision to stop (1) and the chosen disposition (3): in the aggregatem, it is just the number of offenses (numerator) divided by the number of stops (denominator). If subject race does not influence the dispositions of stops (i.e. the number of charges recommended), the hit rate is an indicator of the validity of the decision to stop individuals. More stops without more hits indicates differential treatment. If subject race influences dispositions, particularly if non-white subjects receive more punitive dispositions, the observed hit rate will be artificially raised for those groups, but only where officers possess discretion in disposition. That is, the hit rate will be higher because more charges recommended, but this higher hit rate does not reflect elevated participation in crime. While occurring only after an initial stop, the decision to frisk is an alternate indicator of suspicion which may (unduly) vary by subject race, however the result of the frisk is recovery (or not) of a weapon rather than a discretionary disposition. Differential treatment which inflates standard hit rates will not impact frisk hit rates, because discretion factors only into the choice to frisk but not whether a weapon is found. Frisk hit rates thus provide a point of comparison against dispositon-based hit rates. If overall hit rates were relatively uniform across subject racial categories and similar to those found in frisk hit rates, it would be convincing evidence against differential treatment. This was not found to be the case. The model outcomes in question may be ordered by discretion available to officers, with Any Hit subject to the most discretion and Frisk Hit the least. Findings indicate that hit rates for minority subjects decline as officer discretion decreases, being highest for Any Hit, declining slightly with Arrest Hit, then dropping substantially below the hit rates for whites in the Frisk Hit model. See Appendix 3 for a graphical depiction of this relationship. Likewise, the hit rates for minority officers are lowest in the Any Hit model but the difference disappears completely in the Frisk Hit model. This is most consistent with an underlying reality in which minority subjects may or may not stopped more often than White subjects overall but, when similarly situated, are more likely to be subject to sanctions and frisks. Similarly, the results are indicative that minority officers may be moderately less likely to apply sanctions where they have discretion—producing a lower hit rate—but are identical to White officers when discretion is low–producing the same frisk hit rate and weapon recovery. Note these findings are unlikely to be the result only of increased minority participation in crime as Frisk Hit rates are much lower for minorities. For this relationship to be explained by differential crime participation it would need to be the case that minorities were significantly less likely than White subjects to carry weapons but significantly more likely to be participating in criminal offenses—in both cases holding constant all factors in the models including geographical location. Note that a weapon recovery is not in and of itself indicative of subject participation in a crime. The key element that makes weapon recovery useful is that it is an 21 Full Report Page Number 51 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 52 of 124 objective indicator of a productive result from officer scrutiny (i.e. a frisk). When frisk hit rates are low for a subject racial group, it more convincingly signals they are systematically subject to additional scrutiny, because officer discretion can only impact the number of stops (denominator) not the number of weapons to find (numerator). This is analagous to the relationship between officer scrutiny which produces Terry stops (denominator), but absent the possibility that dispositions are chosen differently by subject race conditional on the same evidence (numerator). These frisk results are also similar to those of Goel et al. (2016) in New York City, which increases confidence in these findings. These results also suggest minority officers are less punitive in Terry stops. For officer race results to hold without minority officers being less punitive, these officers would need to be notably less effective when making Terry stops (lower hit rate) than White officers while being identically effective in conducting frisks (same frisk hit rate). Finally, there appears to be substantial variation across beats and precincts in hit rates, frisk rates, and weapon recovery rates from frisks. Fixed effects models do not permit decomposing this variation into potential causes. Possible sources of these variations could include differences in the quantity and type of crime in the beat, systematic differences in officer behavior, and differential enforcement strategies. For example, it would be expected that a policy of increased frisks in areas with gun violence would result in lower frisk hit rates, but overall counts of weapon recoveries might be higher—a desirable outcome. An auxiliary analysis suggests a mild negative correlation (−0.14) between beat-level frisk hit rates and the percentage of black residents in associated block groups. See Appendix 4 for a maps of frisk hit rates across beats. These differences across beats may contribute to overall variation in race-specific hit rates as seen in Gelman et al. (2007). Future analysis of Terry stops by SPD could be greatly enhanced by altering data collection practices to address limitations in these analyses. First, if feasible, all Terry stops regardless of disposition should be fully documented and linked to CAD entries. Unproductive stops–which are more likely to be unwarranted—are more problematic but also less likely to be well documented. At present, it is not even possible to discern whether unproductive stops were initiated by dispatch or officer on views. This is a significant limitation as past research suggests differential treatment is more pronounced in officer-initiated incidents. Mandating recording of key fields and linking all stops to CAD data would produce more complete records for future analyses. Second, Terry templates may be more useful for analytics if text fields were supplemented by checkboxes to indicate officer motivations for the stops. Written Terry narratives are necessary for record keeping but cumbersome for large scale analytics. Third, more consistent unification of Terry, CAD, and RMS data would permit deeper analysis of the results of stops. At present it was not possible to reliably determine the charges or other results associated with forms of sanction. A full consideration of differential impacts on subjects is impossible without data that describe each point in the policing process (Knox et al. 2019). In summary, while not definitive, these results are consistent with differential treatment in stop disposition and the decision to frisk. Hit rates for minority subjects that are high when discretion is also high but low when discretion is minimal are suggestive of differential treatment in dispositions. That is, it is likely dispositions are more punitive for similarly situated minority subjects, but we cannot evaluate the magnitude of this difference as it may be correlated with differences (if present) in thresholds to stop in the first place. In short, differential treatment in dispositions prevents accurate measurement of differential treatment in the decision to stop. Conversely, evidence is relatively strong that thresholds for frisking minority subjects—particularly black and unknown / other / multiracial individuals—are lower than for white subjects. This is observed as an elevated probability of frisk and a commensurate lower probability of weapon recovery. Because there is little or no discretion in weapon recovery, the source of differences by racial category likely lie in the discretionary decision to stop and/or frisk. 22 Full Report Page Number 52 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 53 of 124 References Albrecht, Steven and John Morrison. 1992. Contact & Cover: Two-Officer Suspect Control. Springfield, IL: Charles C Thomas Pub Ltd. Alpert, Geoffrey P., Roger G. Dunham, and Michael R. Smith. 2007. “Investigating Racial Profiling by the Miami-Dade Police Department: A Multimethod Approach.” Criminology & Public Policy 6(1): 25-56. Baumgartner, Frank R., Derek A. Epp, Kelsey Shoub, and Bayard Love. 2017. “Targeting Young Men of Color for Search and Arrest During Traffic Stops: Evidence from North Carolina, 2002–2013.” Politics, Groups, and Identities 5(1): 107-131. Brown, Robert A. 2005. “Black, White, and Unequal: Examining Situational Determinants of Arrest Decisions from Police-Suspect Encounters.” Criminal Justice Studies 18(1): 51-68. Brunson, Rod K., and Ronald Weitzer. 2009. “Police Relations with Black and White Youths in Different Urban Neighborhoods.” Urban Affairs Review 44(6): 858-885. Chanin, Joshua and Reynaldo Rojo-Mendoza. 2019. “Does Gender Matter? Using Social Equity, Diversity, and Bureaucratic Representation to Examine Police-Pedestrian Encounters in Seattle, Washington.” Administrative Theory & Practice 1-20. Close, Billy R., and Patrick Leon Mason. 2007. “Searching for Efficient Enforcement: Officer Characteristics and Racially Biased Policing.” Review of Law & Economics 3(2): 263-321. Crutchfield, Robert D., Martie L. Skinner, Kevin P. Haggerty, Anne McGlynn, and Richard F. Catalano. 2012. “Racial Disparity in Police Contacts.” Race and Justice 2(3): 179-202. Engel, Robin Shepard, and Eric Silver. 2001. “Policing Mentally Disordered Suspects: A Reexamination of the Criminalization Hypothesis.” Criminology 39(2): 225-252. Fagan, Jeffrey, Anthony A. Braga, Rod K. Brunson, and April Pattavina. 2016. “Stops and Stares: Street Stops, Surveillance, and Race in the New Policing.” Fordham Urban Law Journal 43: 539-614. Fagan, Jeffrey, and Garth Davies. 2000. “Street Stops and Broken Windows: Terry, Race, and Disorder in New York City.” Fordham Urban Law Journal 28: 457-504. Floyd v. City of N.Y., 959 F. Supp. 2d 540. Garner, Joel H., Christopher D. Maxwell, and Cedrick G. Heraux. 2002. “Characteristics Associated with the Prevalence and Severity of Force Used by the Police.” Justice Quarterly 19(4): 705-746. Gaston, Shytierra. 2019. “Enforcing Race: A Neighborhood-level Explanation of Black–White Differences in Drug Arrests.” Crime & Delinquency 65(4): 499-526. Gelman, Andrew, Jeffrey Fagan, and Alex Kiss. 2007. “An analysis of the New York City Police Department’s “Stop-and-Frisk” Policy in the Context of Claims of Racial Bias.” Journal of the American Statistical Association 102(479): 813-823. Goel, Sharad, Justin M. Rao, and Ravi Shroff. 2016. “Precinct or Prejudice? Understanding Racial Disparities in New York City’s Stop-and-Frisk Policy.” The Annals of Applied Statistics 10(1): 365-394. Gordon, Daanika. 2019. “The Police as Place-Consolidators: The Organizational Amplification of Urban Inequality.” Law & Social Inquiry 1-27. Grossmith, Lynne, Catherine Owens, Will Finn, David Mann, Tom Davies and Laura Baika. 2015. Police, Camera, Evidence: London’s Cluster Randomised Controlled Trial of Body Worn Video. London, UK: College of Policing Limited and Mayor’s Office for Policing and Crime. Retrieved October 1, 2019 (https: //whatworks.college.police.uk/Research/Documents/Police_Camera_Evidence.pdf). Higgins, George E., Gennaro F. Vito, and William F. Walsh. 2008. “Searches: An Understudied Area of Racial Profiling.” Journal of Ethnicity in Criminal Justice 6(1): 23-39. 23 Full Report Page Number 53 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 54 of 124 Kahn, Jonathan. 2019. “The 911 Covenant: Policing Black Bodies in White Spaces and the Limits of Implicit Bias as a Tool of Racial Justice.” Stanford Journal of Civil Rights & Civil Liberties 15(1): 1-41. Knox, Dean, Will Lowe, and Jonathan Mummolo. 2019. “Administrative Records Mask Racially Biased Policing.” Available at SSRN: https://ssrn.com/abstract=3336338. Kochel, Tammy Rinehart, David B. Wilson, and Stephen D. Mastrofski. 2011. “Effect of Suspect Race on Officers’ Arrest Decisions.” Criminology 49(2): 473-512. Lanfear, Charles C., Lindsey R. Beach, and Timothy A. Thomas. 2018. “Formal Social Control in Changing Neighborhoods: Racial Implication of Neighborhood Context on Reactive Policing.” City & Community 17(4): 1075-1099. Lauritsen, Janet L., Karen Heimer, and James P. Lynch. 2009. “Trends in the Gender Gap in Violent Offending: New Evidence from the National Crime Victimization Survey.” Criminology 47(2): 361-399. Levchak, Philip J. 2017. “Do Precinct Characteristics Influence Stop-and-Frisk in New York City? A Multi-Level Analysis of Post-Stop Outcomes.” Justice Quarterly 34(3): 377-406. Lytle, Daniel J. 2014. “The Effects of Suspect Characteristics on Arrest: A Meta-Analysis.” Journal of Criminal Justice 42(6): 589-597. Meehan, Albert J., and Michael C. Ponder. 2002. “Race and Place: The Ecology of Racial Profiling African American Motorists.” Justice Quarterly 19(3): 399-430. Mitchell, Ojmarrh, and Michael S. Caudy. 2015. “Examining Racial Disparities in Drug Arrests.” Justice Quarterly 32(2): 288-313. Novak, Kenneth J., and Mitchell B. Chamlin. 2012. “Racial Threat, Suspicion, and Police Behavior: The Impact of Race and Place in Traffic Enforcement.” Crime & Delinquency 58(2): 275-300. Ousey, Graham C. and Matthew R. Lee. 2008. “Racial Disparity in Formal Social Control: An Investigation of Alternative Explanations of Arrest Rate Inequality.” Journal of Research in Crime and Delinquency 45(3): 322-355. Peterson, Bryce E., Lilly Yu, and Nancy La Vigne. 2018. The Milwaukee Police Department’s BodyWorn Camera Program: Evaluation Findings and Key Takeaways. Washington, DC: Urban Institute. Retrieved October 1, 2019 (https://urban.org/sites/default/files/publication/98461/the_milwaukee_police_ departments_body_worn_camera_program_1.pdf). Ready, Justin T., and Jacob T.N. Young. 2015. “The Impact of On-officer Video Cameras on Police–citizen Contacts: Findings from a Controlled Experiment in Mesa, AZ.” Journal of Experimental Criminology 11(3): 445-458. Sampson, Robert J. and Stephen W. Raudenbush. 2004. “Seeing Disorder: Neighborhood Stigma and the Social Construction of “Broken Windows.” Social Psychology Quarterly 67(4): 319-342. Seattle Police Department. 2019. “Crisis Response Team.” Seattle, WA: City of Seattle. Retrieved October 5, 2019 (http://www.seattle.gov/police/about-us/crisis-response-team). Smith, Douglas A. 1986. “The Neighborhood Context of Police Behavior.” Crime and Justice 8: 313–341. Smith, Douglas A. and Christy A. Visher. 1981. “Street-level Justice: Situational Determinants of Police Arrest Decisions.” Social Problems 29(2): 167-177. Smith, Michael R. and Matthew Petrocelli. 2001. “Racial Profiling? A Multivariate Analysis of Police Traffic Stop Data.” Police Quarterly 4(1): 4-27. Stoughton, Seth W. “Terry v. Ohio and the (Un)Forgettable Frisk.” Ohio State Journal of Criminal Law 15(1): 19-33. 24 Full Report Page Number 54 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 55 of 124 Stewart, Eric A., Eric P. Baumer, Rod K. Brunson, and Ronald L. Simons. 2009. “Neighborhood Racial Context and Perceptions of Police-based Racial Discrimination Among Black Youth.” Criminology 47(3): 847-887. Terry v. Ohio, 392 U.S. 1 (1968). Tillyer, Rob. 2014. “Opening the Black Box of Officer Decision Making: An Examination of Race, Criminal History, and Discretionary Searches.” Justice Quarterly 31(6): 961-985. Tyler, Tom R., Jonathan Jackson, and Avital Mentovich. 2015. “The Consequences of Being an Object of Suspicion: Potential Pitfalls of Proactive Policing.” Journal of Empirical Legal Studies 12(4): 602-636. Varano, Sean P., Joseph A. Schafer, Jeffery Michael Cancino, and Marc L. Swatt. 2009. “Constructing Crime: Neighborhood Characteristics and Police Recording Behavior.” Journal of Criminal Justice 37: 553-563. 25 Full Report Page Number 55 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 56 of 124 Appendices Appendix 1: Structual Equation Model Approach The original proposal for this project featured an SEM approach. The purpose of that method was to separate the use of relevant information from sources of bias in the officer decision making process. After conducting preliminary analyses on the data received, this approach was deemed infeasible. There were two primary issues preventing use of SEM. First, due to prohibitive labor required to process Terry stop text narratives, important measures of officer information were not available in the final data. Separating officer information from bias would be unreliable without these data, and without this component the primary motivation for using SEM is removed. Second, a prohibitively complex SEM model would be required to address the substantial beat-level heterogeneity unattributable to available measures. In the absence of latent variables (e.g. officer information), the fixed effects logistic regression approach offers an equivalent but easier to estimate model to the SEM approach. These problems were anticipated in the initial proposal as possible reasons a hierarchical model (e.g. fixed effects) might be used instead of SEM. Future work may usefully pursue the original SEM approach laid out in the proposal, but it would require extensive mining of officer narratives from Terry templates. While labor intensive, an algorithmic approach to mining narratives is feasible, likely to make a contribution to knowledge on its own, and would facilitate more detailed analyses of Terry stops. Appendix 2: Omitted Measures Appendix Table 1. Omitted Measures Concept Measure Local Context Population density; index of disadvantage (unemployment, public assistance, and poverty); counts of all offenses; counts of violent offenses; proportion of offenses which are violent; counts of arrests; racial composition of those counts; congruence between local racial composition of offenders and subject race Duration officer observed subject before stop Subject’s perceived build and height Officer rank; officer certifications Spatial lags Stop Context Subject Characteristics Officer Characteristics Statistical Controls 26 Full Report Page Number 56 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 57 of 124 Appendix 3: Hit Probability by Discretion Appendix Figure 2 depicts differences in hit rates by subject race from the overall mean for each hit rate. This permits comparison of hit rates between subject race categories across each type of hit rate. This plot reveals that as discretion decreases from left to right, the hit rates for minority subjects generally declines. Conversely, the hit rate for white subjects increases. We would expect to see this pattern if minority subjects are subject to differential treatment in disposition, that is, if it is more likely for officers to recommend charges for non-white subjects compared to white subjects in otherwise similar situations. Appendix Figure 1. Hit Probability by Discretion By subject race Mean Probability Difference 0.04 0.02 0.00 −0.02 Any Hit Arrest Hit Frisk Hit Decreasing Discretion → White Asian Latinx American Indian / Alaskan Native Black Unknown / Other / Multiracial Appendix 4: Frisk Hit Rate by Beat Appendix Figure 3 depicts frisk hit rates across SPD beats. 27 Full Report Page Number 57 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 58 of 124 Appendix Figure 2. Frisk Hit Rate by Beat frisk_hit_rate 0.4 0.3 0.2 0.1 28 Full Report Page Number 58 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 59 of 124 DISPARITY REVIEW – PART II APPENDIX B INDEX OF 62 RANDOMLY SELECTED CASES Cases Screened and Selected for External Review Item # Type Type of Call 1 Frisk Shots Fired / Suspicious Circumstance 2 Frisk Traffic 3 Frisk Shots Fired General Description Individuals in and around a car, parked in supermarket parking lot at night, stopped by patrol officers. Asked to exit car and frisked. Officers looking for shooting suspect. Man stopped for running stop sign. Issue with bill of sale. Frisk for visible knife. Complainant called for shots fired in alley. Officers stop and frisk two subjects Received Time of Day Race/ Ethnicity of Subject(s) Outcome Length of Stop (if applicable) On-view Night Black All released; subject not found >15 min 6 On-view Daytime Hispanic Traffic Warning issued 5 - 10 min 5 Dispatch Daytime Hispanic All released; no firearm found 10 - 20 min 12 SEATTLE POLICE DEPARTMENT # of Officers Present Factors considered for not including in group review sample 59 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 60 of 124 DISPARITY REVIEW – PART II matching general description. 4 Frisk Narcotics 5 Frisk Trespass 6 Point Shoplifting / Robbery / Crisis 7 Point Suicide Young adult stopped after seeing likely make a drug deal through a car window and stuffing items into backpack. Officer was on patrol and noticed a man trying to pry open a locked fence to a lot. Officer frisked for safety as burglar tools often can be used as weapons. Baggy clothing with bulging pockets. Man alleged to have stolen an ice axe. Officers follow through streets trying to get him to drop it. Subject called reporting suicidal and armed with a knife. Officers respond and get On-view Daytime White Released 10 - 20 min 2 On-view Night Asian Arrested 5 - 10 min 3 Dispatch Daytime White Arrested NA 8 Not reviewed due to time constraints Dispatch Daytime Hispanic ITA NA 6 Not reviewed due to privacy concerns/policy SEATTLE POLICE DEPARTMENT 60 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 61 of 124 DISPARITY REVIEW – PART II him to drop the knife. 8 Point Hit & Run / Fleeing 9 Point Weapon Threat 10 Point Stabbing Subject engaged in dangerous driving and then fled in vehicle. Subject quickly exited vehicle on foot. Later seen again and officer and community members apprehend. Officer was dispatched to ca call of a man making a threat with a gun. Officer saw described car and man get out matching the subject description. Suspect ignored commands Officers were called to a report of a man swinging a knife at another at the food bank. When located they commanded him to get to the ground. On-view Daytime Hispanic Arrested NA 13 Dispatch Daytime American Indian / Alaskan Native Arrested NA 4 Dispatch Daytime Asian Arrested NA 15 SEATTLE POLICE DEPARTMENT Not reviewed following CPC advice 61 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 62 of 124 DISPARITY REVIEW – PART II Cases Screened and Removed From Sample as Not Suitable for External Review 1 2 3 Frisk Frisk Frisk Man with Gun Traffic Shoplift Call of male brandishing a gun in parking lot. Witnesses reported he was intoxicated and brandishing the firearm. Expired tags noted. When police car turned around, subject parked and walked away. When officer told him to stop and walk back, he stuck his hands in pockets, he did not follow command to keep hands out of pockets. Store had called reporting a shoplift. Subject was the person observed with goods, but turned out subject had goods before entering the Dispatch Dispatch Dispatch Night Daytime Night Black Arrested on warrant > 20 min 6 Call of man with gun - high risk stop; age of case Asian Paraphernalia, knife, sword, and contraband surrendered. Warning issued. 10 - 20 min 4 Difficulty locating ICV; age of case 2 Subject asked to be frisked so he could put hands in pockets Hispanic Released SEATTLE POLICE DEPARTMENT 5 - 10 min 62 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 63 of 124 DISPARITY REVIEW – PART II store. The store security was mistaken. 4 5 6 Frisk Disturbance Weapon Frisk Disturbance Domestic Frisk Disturbance Call of male swinging a large knife at a bus stop. On arrival no knife present and subject said he did not have one. Frisk for safety. Subject said earlier he had a stick. Multiple calls of two subjects fighting, one with a knife. Officers arrived to two bloody subjects. Officer was flagged down by subject, who claimed other subject had taken money from her and wanted him removed from the doorway where she was sleeping. Dispatch Dispatch On-view Daytime Black Released Daytime American Indian / Alaskan Native Crisis template; released Night Black No action SEATTLE POLICE DEPARTMENT 10 - 20 min > 20 min 5 - 10 min 3 Call and witness statement that they saw the man from a close distance threatening someone with a knife 7 Call of man with knife -both subjects bloody 2 Unclear in text if frisk actually happened; age of case; difficulty in ICV 63 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 64 of 124 DISPARITY REVIEW – PART II 7 Frisk Car Prowl 8 Frisk Suspicious / Assist Public Officer was called to an in-progress car prowl. Subject's location and clothing matched description. Frisked due to experience that prowlers often have heavy tools that could be weapons. Caller positively identified subject. Subject had a purse containing likely stolen property. Backpack had prowling tools. Called to a man slumped over his steering wheel in a park. When officers arrived the car quickly drove away. Followed vehicle and driver fled from vehicle. Suspicious nature of who owns the vehicle. Driver fled because of DWLS and DOC warrant -- as stated by his Dispatch Night White Arrested on PC > 20 min 1 Age of case; difficulty with ICV Dispatch Daytime White Released 10 - 20 min 2 Age of case; difficulty with ICV SEATTLE POLICE DEPARTMENT 64 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 65 of 124 DISPARITY REVIEW – PART II passenger. Officer could prove no crime and they were all released. When interviewing passenger had to keep telling her to keep hands out of pockets, so finally frisked. 9 Frisk Burglary 10 Frisk DV Assault 11 Frisk Warrant Officers called to report of a man sleeping in a business he was not supposed to have access to. While handcuffed subject stated he had a knife so he was frisked. DV issue combined with substance/crisis needs. Frisked due to allegation of assault against him. Bike officers were looking for a warrant suspect. Matched description of suspect last seen running from area - and was Dispatch Daytime White Warned 5 - 10 min 3 Subject disclosed possessing a knife. Dispatch Daytime White Arrested > 20 min 4 Due to nature of call, frisk was going to occur. On-view Night White Arrested on warrant 5 - 10 min 7 Video issue SEATTLE POLICE DEPARTMENT 65 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 66 of 124 DISPARITY REVIEW – PART II sweating. Knife located. 12 Frisk Stolen Auto 13 Frisk Assault 14 Frisk Assault Patrol vehicle ran plate and it came back as stolen. Vehicle stopped before patrol officer activated lights and passenger got out on his own and waited on sidewalk. Subject matched description of a man armed with a very large knife. Officers were responding to multiple 911 calls about a male striking two females on the street. Responding to fight in progress. Call stated one subject had a knife out. Victim approached officers saying subject had assaulted him. When asked subjected On-view Night White Arrested > 20 min 9 Seems as frisk was incident to arrest Dispatch Daytime White Arrested 10 - 20 min 6 Report seemed linked to incorrect Terry 4 Nature of call and subject saying he had a knife. Dispatch Daytime White Arrested SEATTLE POLICE DEPARTMENT 5 - 10 min 66 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 67 of 124 DISPARITY REVIEW – PART II admitted to having a knife, so a frisk followed. 15 16 Frisk Frisk Disturbance DV Disturbance Weapon On patrol observed couple in a heated discussion. A few moments later witnesses flagged officer down to say male had shoved the female. While being questioned male said he had a sword in his pants. So, a frisk followed. Dispatched to call of two subjects fighting, with one pulling a knife on the other. On arrival witness pointed to suspect with knife. Investigation revealed the misunderstanding and neither wanted to press charges. On-view Dispatch Night Daytime White White Arrested Released SEATTLE POLICE DEPARTMENT > 20 min > 20 min 5 Subjected admitted to having a sword in pants. 5 Witness pointed to subject and said he had threatened other subject with a knife. Subject admitted to displaying knife 67 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 68 of 124 DISPARITY REVIEW – PART II 17 Frisk Shoplift 18 Frisk Suspicious / Narcotics Frisk Stolen Property 19 20 Frisk Stolen Vehicle Subject was stopped during a shoplift from a business. Frisked during stop. Also wanted on a felony warrant. Call of intoxicated man going doorto-door asking to do odd jobs. One officer recognized the subject as a frequent burglar/car prowl subject. Based on prior knowledge frisked for safety as his burglary tools could be used as weapons. Officer got ALPR notice of a stolen vehicle. Officers initiated a highrisk vehicle stop. Conducted routine check of a license plate. Followed. Vehicle pulled over without activating lights. Waited for backing units. Complied with orders but was frisked for safety. Dispatch Night Black Arrest warrant 10 - 20 min 4 Frisk appears to be to check for other stolen property and incident to arrest. Dispatch Daytime Asian Arrested 0 - 5 Min 2 Subject known to officer with prior history. 12 High risk stolen vehicle stop. May not be a Terry 12 High risk stolen vehicle stop. May not be a Terry Dispatch On-view Daytime Night Asian Asian Arrested Arrested SEATTLE POLICE DEPARTMENT > 20 min > 20 min 68 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 69 of 124 DISPARITY REVIEW – PART II 21 Frisk Fight / Weapon 22 Frisk Stolen Vehicle 23 24 Point Point Officers dispatched to call of two men fighting, one with a knife. When contacted subject confirmed he had a knife in his pocket. Officers frisked and removed knife. Officer was on patrol and ran a routine license plate check. It returned as stolen. Officer stopped the vehicle and had all the occupants get out. Warrant Officers were serving a narcotics search warrant. Stolen Property Detectives were attempting to recover stolen property. Subject admitted to possessing stolen property but said he didn't steal it. Detectives later got a search warrant to search a room; Occupant Dispatch Night Asian Released 0 - 5 min 9 Call of man with a knife fighting another man. Man admitted to having knife. On-view Daytime Asian Arrested > 20 min 20 Felony stolen vehicle stop. White Arrested for various pills, currency, and equipment. NA 12 Warrant service White Arrested on various crimes and the warrant. On-view On-view Daytime Night SEATTLE POLICE DEPARTMENT NA Op / Warrant Service / Traffic Stop / Age of case 69 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 70 of 124 DISPARITY REVIEW – PART II left in a vehicle and officers later had to force it to stop 25 26 27 Point Point Point Crisis Officer dispatched to crisis call at boat ramp. Subject relayed to dispatch he was suicidal and had a gun. Shots Fired Officers dispatched to shots fired call. 810 shots. Male running north. Dispatch Disturbance Officers looking for a warrant suspect heard a faint scream of woman asking for the police. Officers looked to sound and saw a group of people fighting. Officers tried to intervene in fight and they started to get assaulted. On-view Dispatch White Medical / ITA NA Presence of gun / threat of harm Night White Arrested on agg assault; doc violation; multiple weapon possession. NA Shots fired call; witness Night White Arrested for agg assault NA Daytime SEATTLE POLICE DEPARTMENT 12 Assault on officer. 70 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 71 of 124 DISPARITY REVIEW – PART II 28 Point Warrant 29 Point Trespass 30 Point Stolen Vehicle Officers were working to apprehend subject on a narcotic warrant. Located subject in his vehicle. Officers were dispatched to a 911 call. Caller said people in his house he was working on renovating. Officers could see people inside. They would not let the officers in. Finally, the people inside agreed to come out. Then heard another person in the attic. Officers on bike patrol heard a broadcast for a stolen vehicle and saw it. Pursuit occurred but vehicle ultimately was not stopped. On-view Night White/Hispanic Arrested on warrant; additional charges on weapons and ammo NA 13 Warrant service; vehicle stop Dispatch Daytime White Arrested on various charges. NA 8 Circumstances of getting people out of a house they are not allowed in. Age of case. Dispatch Daytime White Eluded NA 4 Pursuit / High Risk stop SEATTLE POLICE DEPARTMENT 71 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 72 of 124 DISPARITY REVIEW – PART II 31 Point Assault - Gun 32 Point Shots Fired Officers were dispatched to a call of a man having pointed gun at caller. During area check officers located man matching the description, including a unique bicycle. Asked to put hands up multiple times. Told he matched description of man with a gun. Shots fired call. Caller said his friend who rents the room in the back may have fired a gun in the yard. Saw friend carrying the gun. Subject did not respond to officers. Subject tried to enter several homes. Continued to advance on officers refusing orders. Dispatch Daytime White Released (had an outstanding warrant) Dispatch Night Black Arrested on several charges. SEATTLE POLICE DEPARTMENT NA 8 Call of man with a gun. NA 13 Shots fired call; witness saw man with gun. 72 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 73 of 124 DISPARITY REVIEW – PART II Point Narcotics Warrant 34 Point Felony Drug Stop 35 Point Drug Buy Op 36 Point Shots Fired 33 37 Point Collision/Gun SW and MCTF serving narcotics warrant. Firearms pointed as cover on door of motor home. Officers told subject was in parking lot with a fresh supply of narcotics and dealing to 4 to 5 people. Subject known to officers to be assaultive to officers. ACT team working with CW to purchase narcotics. Officer heard sound of gunshot downtown. Saw two individuals, one with what was clearly a firearm in his hand. Apparent collision where occupants said another vehicle had a gun and had pointed it at them. Officers kept guns on that vehicle waiting for backup. On-view Daytime Hispanic Multiple arrests including stolen shot gun. NA 6 Warrant service NA 6 Felony stop; known assaultive subject On-view Daytime Black Arrested for multiple charges. On-view Night Hispanic Arrested on multiple charges. NA 15 Drug buy op On-view Night Black Arrested. Gun did not belong to him. NA 8 Shot fired call; officer saw gun. 10 High-risk vehicle stop; victim reported gun threat On-view Night Asian Arrested SEATTLE POLICE DEPARTMENT NA 73 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 74 of 124 DISPARITY REVIEW – PART II 38 Point Shooting 39 Point Narcotics Warrant 40 Point Carjacking 41 Point Robbery Radio call to two males shooting at each other. Gun used in taking suspect believed to have a gun into custody. Officers were serving a narcotics search warrant on an individual in a car. Caller said daughter had been carjacked at gunpoint. Observed subjects ran from suspected vehicle, while others drove away at a highrate of speed. K9 and Guardian One used to help locate subjects. Confusing call alleging being robbed at gunpoint; made to strip and lay on ground; found, followed, and stopped stolen vehicle. Dispatch Daytime Black Suspect outstanding NA 30 Active shooting incident. On-view Daytime Black Arrested NA 11 Narcotics search warrant. Dispatch Night Black Arrested NA 16 High-risk stop and search; carjacking at gunpoint Dispatch Night Black Arrested NA 13 High-risk vehicle stop SEATTLE POLICE DEPARTMENT 74 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 75 of 124 DISPARITY REVIEW – PART II 42 43 Point Point AssaultWeapon DUI 44 Point DV Assault 45 Point Shots Fired Point Child Kidnapping 46 Call for man in stairwell threatening woman with a knife and rifle. Officers confirmed both weapons. Call for vehicle stopped in middle of the road with driver slumped over. Officer dispatched to a DV assault in progress. Exigent circumstances, concerned if someone was injured or dead inside. Officer dispatched to shots fired call. Officers joined a dispatched call for a fleeing AMBER suspect vehicle. Vehicle eluded officers and rammed patrol vehicles with child in car. Subject used child as a shield. Dispatch Dispatch Night Night Hispanic Black Arrested Arrested NA NA 25 Man with visible knife and gun. 12 Nature of call -approaching a car stopped in middle of the road. Dispatch Daytime White Arrested NA 5 DV call -- officer unsure if someone was in danger inside. Exigent circumstances. Dispatch Night Black Suspect outstanding NA 10 Shots fired call 30+ AMBER alert; fleeing car; rammed vehicles Dispatch Daytime Black Arrested SEATTLE POLICE DEPARTMENT NA 75 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 76 of 124 DISPARITY REVIEW – PART II 47 48 49 Point Point Point Shooting Eluding Robbery / Gun Officers were looking for juveniles who had caused a disturbance in a mall. Saw a suspect vehicle but didn't have enough to stop. Sitting next to it at a light, heard gunshots and saw two males running from the mall. Both were actively firing their guns. Officer recognized car involved in a shooting. Vehicle attempted to flee. Vehicle collided with another vehicle. Driver bailed and exited on foot. Officers dispatched to call of a woman who had been assaulted and robbed and reported a gun. While investigating, saw a male jump up from bushes and Dispatch On-view Dispatch Night Daytime Night Black Black American Indian / Alaskan Native Arrested Arrested Arrested SEATTLE POLICE DEPARTMENT NA NA NA 30+ Subjects actively firing weapons. 20+ Eluding vehicle. Crashed into another vehicle. Fled from vehicle. Vehicle was known to be involved in multiple shootings. 10 Fleeing individual believed to have a gun for the robbery. Not reviewed publicly as it was a juvenile. 76 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 77 of 124 DISPARITY REVIEW – PART II run, and did not respond to commands to stop. Point Stolen Vehicle 51 Point BurglaryOccupied 52 Point Assist Other Agency 50 On patrol officer saw a stopped car in an area known for drug dealing. Running the plate revealed it as stolen. Decided to initiate a felony stop. Man trying to force entry into house. Vehicle associated with recent shootings. Felony stop of vehicle. Multi-agency, multi-warrant service on narcotics. On-view Night Black Arrested NA 4 Felony stop for stolen vehicle. Dispatch Daytime Asian Arrested NA 8 Felony stop for vehicle involved in shootings. On-view Night Asian Arrested NA 15 Narcotics search warrant. SEATTLE POLICE DEPARTMENT 77 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 78 of 124 DISPARITY REVIEW – PART II APPENDIX C COMMUNITY “INVITE” Community Partners and Leaders, The Community Police Commission in partnership with the Seattle Police Department respectfully requests your assistance in gathering Community voices to further the conversation of police accountability in Seattle. In April of 2019 the Seattle Police Department published “Disparity Review-Part 1,” a report that detailed racial disparity captured in SPD data. The findings included that People of Color were more likely to be frisked, but less likely to be found with a weapon; and People of Color were more likely to have a firearm pointed at them. Where the first part of the Disparity Review was driven by data, SPD identified conversations with Community as an immediate and necessary step to begin the second part of the Review. The CPC was asked to assist in gathering Community and creating spaces that support sharing this crucial feedback. In the evenings of September 24th through the 27th, The CPC in partnership with SPD, will host a series of forums to uplift the voices of the Communities who experience these disparities. The forums will be racially caucused to promote observations that maybe specific to the Community of the participants. Participants will review materials, both written and audio-visual, from randomly selected cases that reflect the disparity data. This is undoubtedly a labor-intensive task, both critically and emotionally, so participants will be compensated for their time and expertise. Anyone who feels that this subject matter, materials, or working closely with members of SPD will be triggering or particularly emotionally distressful should feel empowered to decline participation. These conversations are necessary to protect the wellness of the impacted Communities, and so should the persons who makeup those Communities feel free to protect their wellness. Our hope is to gather 6 to 8 participants for every forum. Your assistance in sharing this information, and recommending people from your community to participate, would be greatly appreciated. More details about the events are available on the attached flyer. • • • • September 24, 2019: African American Forum 6:30PM- 8:30PM September 25, 2019: Asian/ Southeast Asian/ Pacific Islander Forum 6:30PM8:30PM September 26, 2019: Native American Forum 6:30PM- 8:30PM September 27, 2019: Latinx Forum 6:30PM- 8:30PM SEATTLE POLICE DEPARTMENT 78 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 79 of 124 DISPARITY REVIEW – PART II APPENDIX D INTERNAL REVIEW COMMENT SUMMARIES Incident #1 involved a 911 call from an eye-witness who reported hearing a gun shot and seeing two individuals in his alley with a black handgun. The responding officer arrived to take his statement, while other officers canvassed the area looking for individuals who met the initial description of the subjects. While patrolling, officers located two individuals they thought matched the subjects’ descriptions. When they were located, officers instructed them to stop from their patrol vehicles, and several exited their vehicles with their firearms drawn and pointed. A stop and frisk commended on both subjects. Later, the officer with the caller, transported the caller to the location and he confirmed these individuals were not the two people he saw in his alley. The frisk of the two subjects resulted in the discovery of illegal drugs but no weapons, and both subjects were released from the scene. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the call being for a reported shots-fired call. They all felt that because a handgun was likely involved, they would have been on heightened alert. SPD raters note that placing the subjects in handcuffs for a safety frisk was consistent with training. There was one comment that one subject was not cooperating and continued to put his hands in his pockets, further raising concerns. All raters noted that the involved officers remained calm throughout the encounter and explained what they were doing and why. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the job the officers did explaining what was happening so the subjects did not fear that something else was going to happen. One rater felt that initially given the scale of the response, they would have felt unsafe – the response would make anyone feel that something serious had occurred. The raters noted that the officers’ efforts at calmly explaining what was happening would make them think the subjects would feel less at risk. One rater noted that it was understandable that the subjects might have felt that they were stopped because they were near a methadone clinic. SEATTLE POLICE DEPARTMENT 79 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 80 of 124 DISPARITY REVIEW – PART II Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused the likelihood they would have assumed something serious and potentially dangerous was occurring, given the number of officers responding. The presence of officers exiting their vehicles with commands of hands up and drawing their firearms, likely further contributed to a sense of volatility. Raters also were in agreement that the way in which the officers quickly got the situation under control and ratcheted down the tension, would have alleviated a lot of the concern. The general calmness of the officers and general compliance of the subjects also would have made them feel more safe. One rater notes that many community members captured on the video proceeded with their daily activities without much observable concern. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on the job the officers did explaining why they were stopped…specifically, that they matched the description of subjects thought to have recently fired a gun in the area. Furthermore, the officers explained they were only looking for a gun, and did not worry about the drugs/paraphernalia they located. As soon as they got confirmation they were not verified by the caller and discovered no weapons, they were immediately release. Together, these factors led the SPD raters to feel that the subjects likely felt they were treated fairly and professionally. Overall Perception of the Incident The reviewers focused on the scale of the response for two individuals in the middle of a busy health care and retail corridor and how that would be perceived by the community. If the community members did not know that they were responding to a shots-fired call, that would explain a lot of the response, thought there was still some concern about the number of officers, cars, and guns. The other comments focused on how calm the officers were and how they were clearly communicating to the subjects why they were stopped and what they could consent to, or not consent to. Incident #2 involved a traffic stop when the officer observed the vehicle fail to stop at a stop sign and then dangerously weave through traffic in an area busy with vehicular and pedestrian traffic. When the automobile finally came to a stop the driver exited the car and ran off. A few moments later, while detaining the passengers in the car (waiting for additional officers), the driver ran past the car. The officer noticed the driver, and chased after. The officer’s weapon was drawn due to the nature of the stop. As the officer chased the driver, he asked for pedestrians to help stop the driver, which they did. He was taken into custody. This incident occurred during daylight hours. SEATTLE POLICE DEPARTMENT 80 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 81 of 124 DISPARITY REVIEW – PART II Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the concern for why the driver failed to stop and then fled. All raters understood why the officer had his firearm drawn. The driver had placed drivers and pedestrians at risk. Given the driver’s abnormal behavior to avoid being stopped, a reasonable officer would have serious concerns about what else the driver was willing to do. One rater noted that in the video you can see the driver reach in the back of the car before fleeing – providing an opportunity to grab a weapon. This same rater noted they would not have felt unsafe during the incident. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the expectation that the subject must have anticipated being injured as his actions would require some level of force to apprehend him. One rater noted the subject was not likely to know the officer had his firearm drawn. The rater noted that given this, and given that the officer was not using derogatory or threatening language – suggesting the officer was in control and not likely to resort to violence. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on how they might have felt seeing an officer running down the street with their gun pointed yelling for people to grab someone. One officer noted that the community members in the video did not appear to be particularly concerned. This rater hypothesized that the officer wouldn’t have asked community members to help stop the individual if there was any significant risk. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on how the officer never denigrated the subject. The officer remained calm despite the subject putting the officer, the community, and themselves at risk. One rather highlighted the calm nature of the officer’s interaction with the subject once they were apprehended. This rater noted that nothing about the style of the officer’s interaction with subject would lead them to believe that they were being treated unfairly or were at any risk SEATTLE POLICE DEPARTMENT 81 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 82 of 124 DISPARITY REVIEW – PART II Overall Perception of the Incident It would be unclear to the general community why there were so many officers. They would not know that one officer technically “wasn’t there,” because he was a field training officer. So, having three officers respond to a traffic stop where the subject seemed to be following all orders and was calm, would lead some to question the response. If they were close enough to hear, they would know that the officers were explaining every step and behaving in a calm and professional manner. Incident #3 involved a pursuit of a vehicle believed to be involved in an armed carjacking. Two subjects run away from the car while two others got back in and dangerously fled the scene. The subjects eventually bailed and the officer called in K9 to assist with the track. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the fact that the officers were responding to a reported armed carjacking. One rater notes that the decision to point the firearm was in line with training and policy. The officers had every reason to believe that the fleeing subjects were willing to do anything to avoid apprehension. The time of day and its darkness also was cited as a factor in the officer’s decision making. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the ability of the officers to calmly communicate what was happening and why. One rater stressed that the officers did a good job of delivering one message so that there was no confusion, which can lead to anxiety. One rater noted that if the subjects were involved in the alleged carjacking, they had to be aware the officers would be concerned for their safety and might be prepared to use significant force. The officers gave the subject the chance to safely cooperate. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on how everyone was at-risk given the number of officers involved in locating someone believed to have a firearm. Multiple firearms were drawn and the K9 was deployed. A random community member would likely have no other conclusion than that something very serious was happening. One officer noted how the community would have seen the officers treat the subject with a high level of professionalism and respect. This rater also noted that the officers used their resources to establish containment to ensure SEATTLE POLICE DEPARTMENT 82 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 83 of 124 DISPARITY REVIEW – PART II everyone’s safety. The community members would have noted that the officers were not solely focused on arresting the subject, but on keeping everyone safe. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on how professional the communication was from the officers. They even stressed how the officers explained their concern for the subject’s friend’s safety. One rater noted the officers never used demeaning language and did all they could to apprehend them without injuring them. One rater noted that officers must have been clear in their instructions as the subject followed every step of their instructions. Overall Perception of Incident There were general perceptions that if someone didn’t know what was occurring, it would look like a robust response. It would be feasible to assume that the officers had a reason for such a large response and were not randomly targeting the individuals. It was noted that in some instances during the incident the officers explained to community members what was happening. It also was noted that the community would have an overall positive view of the incident given the professionalism with the subjects were treated, especially following their very dangerous behavior. Incident #4 involved a call from a subject who said he had a knife and wanted to hurt himself. When the officer team arrived at the apartment they had to wait for the subject to open the door. When he opened the door, he had a knife in his hand. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the subject having a knife and the possibility that he might attempt to have the officers harm him. The presence of the knife had to raise the threat level. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the officers attempts to keep the situation calm. If the subject had the intention to hurt himself, he may not have had any concerns for his safety. SEATTLE POLICE DEPARTMENT 83 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 84 of 124 DISPARITY REVIEW – PART II Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on the officers communicated that they wanted the subject to be safe. The officers, in a residential apartment space, had to be sure the subject, the officers, and the neighbors were safe. This meant being able to subdue any potential violence. This also meant trying to keep things calm and being clear that they were there to help the subject and were trying to end the event with no one getting hurt. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on the officers seeming to express genuine concern for the subject’s well-being. Even after the danger had passed, the officers were making sure the subject had everything he needed to before they took him to a hospital. They seemed to genuinely want to know what was leading him to wanting to hurt himself that day. The officers treated him like he was just a regular person who needed help, once the danger from the knife was subdued. Overall Perception of Incident The reviewers generally felt that a community member who saw the entire incident or knew what was happening would understand why they had their firearms pointed at the individual with a knife and threatening to harm themselves with it. Before the door was open, it might be unclear why the officers already had their guns drawn. It was noted that the officers were consistently calm and professional. If a community member saw the resolution of the incident they would satisfied with the manner in which the officers cared for this community member. Incident #5 involved an incident where patrol officers were assisting detectives in locating and handling a car believed to contain a suspect for which they had probable cause to arrest for a shooting. The vehicle was sitting in a grocery store parking lot and one person was standing outside of the car while some number of individuals are inside the car. It is dark outside. When the patrol vehicle pulls up with its lights on, the subjects do not attempt to flee. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the nature of the stop – following a reported gunpoint robbery. This led to officers following their safety training by having firearms deployed during the stop. One respondent felt that given the SEATTLE POLICE DEPARTMENT 84 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 85 of 124 DISPARITY REVIEW – PART II excellent tactics that the officer used, they would have felt very safe during the incident. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the overall tone and calmness of the officers. They gave clear instructions and did not escalate the situation. One rater noted that any subject might be concerned about the seriousness of the incident given the number of patrol vehicles, officers, firearms pointed, lights, sirens, etc. If the subjects were involved in the alleged behavior, they likely would be concerned about what level of force the officers might be prepared to use, given their knowledge that they likely still had a weapon with them. The subjects did appear to remain calm throughout the incident. One rater noted that the officers made sure the scene was well-lit, considering it was dark, so that there was no question about the subjects’ movements or actions. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on the calmness of the officers and how that would have communicated that while this was serious, it was not volatile. Community members might have felt a gun might be used, given the sheer number of weapons present, and maybe having a very general belief that traffic stops are very dangerous. Community members might have felt that the officers treated everyone in a calm and respectful manner. One rater also noted that the officers’ work at illuminating the scene likely could have made people feel more safe, than if this all had been happening in the dark. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on how the officers made sure to explain what was happening, answer questions, and even adjust the handcuffs when one subject noted that they hurt. The officers were clear and polite, and never used any demeaning or aggressive language. Overall Perception of Incident If a community member was witnessing this and had no knowledge of what preceded it or how it resolved, they might assume the individuals were stopped due to their appearance. The subjects were calmly hanging out in a very public parking lot and then officers approached them. Ultimately, they were all released. This resolution might lead to conclusions that they had the wrong person (which was the case) or were harassing them. It did seem like a large number of officers initially for the generally lowkey manner in which the rest of the incident unfolded and resolved. SEATTLE POLICE DEPARTMENT 85 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 86 of 124 DISPARITY REVIEW – PART II Incident #6 involved a call where the officers were working with detectives to locate a car that had been connected to multiple shootings across the previous days. When located the car started to actively elude the police with a high-speed chase, a collision with another vehicle, and a foot pursuit as the subjects fled from the vehicle. This was during the day time. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the connection to shootings, the high-speed chase, the potential harm to others through the chase and the crash, and the fleeing on foot. Together, all of these behaviors likely made the officers have a level of concern about what the subjects were willing to do to avoid apprehension. Having to call out K9, Guardian One, and so many officers, likely raised the level of tension on the incident. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on how the subjects all were generally calm when they were apprehended. One rater noted that given the severity of the alleged behaviors, the subjects might have been concerned about the officers’ level of concern for their safety and that of the community. One rater noted that the officers remained calm and never got overly loud or obviously agitated. The officers were very clear in the directions they were given once they apprehended the subjects. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on the scale of the response and how the community would not know what exactly was happening…though with some acknowledgement that it also would have potentially indicated that incident was very dangerous. One rater also noted that a community member would have noted how calm the officers were and how professionally they treated the subjects. One rater noted that the community might have been very concerned about what the subjects were willing to do given the response. Another rater noted that until communication would have been shared it would have been really easy for the community to jump to conclusions, and that the officers did an excellent job of reassuring the community that generally everything was under control. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on how calmly and professionally the officers treated the subjects once they were apprehended. A rater noted that the apprehension was forceful, but given the circumstances of the incident, totally within policy and understandable. Another rater noted how quickly the officers ensured that medical treatment was provided to multiple subjects. SEATTLE POLICE DEPARTMENT 86 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 87 of 124 DISPARITY REVIEW – PART II Overall Perception of the Incident One reviewed noted that this was an exceptionally large response and even if a community member had knowledge of what was happening, they might question the size of it. The reviewers did note that in all of the footage the officers seemed to behave professionally and calmly. One reviewer noted that given the lengthy, dangerous pursuit, that any community member who was aware of that would understand the officers’ actions. Incident #7 involved a Terry stop where the officer observed the subject acting suspiciously in how they leaned into a car window in an area known for high levels of drug dealing. The officer further noted they saw the subject shove some items into their backpack once they saw the officers. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on how the subject generally was following the officers’ orders from the beginning. One rater noted that the daylight also contributed to this stop not causing a high level of concern. One rater did note that the officers performed a frisk due to the behavior they had seen – where it wasn’t clear what the subject was attempting to conceal. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on how the officers communicated very clearly and calmly what the needed the subject to do. A rater noted that initially the subject seems to be very nervous. The officers explained why he had been stopped, while remaining calm and polite. This rater also noted that the lack of a large response also kept the level of tension down. A rater noted that while no one likes being stopped by the police, the officer was clear and calm and the subject never appeared to feel unsafe. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on how quickly the officers had the scene under control and removed anything that could have been dangerous from the reach of the subject. The raters also focused on how much time the officers were taking to explain why they were doing what they were doing. Another rater noted that most community SEATTLE POLICE DEPARTMENT 87 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 88 of 124 DISPARITY REVIEW – PART II members in the area would have assumed that the stop had something to do with a drug stop, and that the low-level of the response would mean it was not an overtly dangerous interaction. One rater noted that the officer just stopping in the middle of the street and jumping out might have led to some community members being concerned it was an emergency matter. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on how the officers were direct but firm. One rater notes that the officers could have arrested the subject, but did not, because of how cooperative he had been. Another rater noted the subject might have assumed he was being targeted because of the area he was walking in and his known association with drug activity. The officers communicated their main concern was his warrant for theft and trying to understand why he seemed to have so much property that likely wasn’t his. Overall Perception of Incident Reviewers felt that local community members are aware of the drug dealing issues in the area and would understand why this person was stopped given their observable behavior. The officers were calm, professional, and polite in all their interactions with the individual. It was noted officers allowed the subject to sit down. The subject and the officers all remained calm which could lead a community member to conclude everything was going well. Incident #8 involved a Terry stop where the officers had been called to the scene of a fight where one person was alleged to have pulled a knife on the other. The officers quickly ascertained that neither subject was injured and both just wanted to resolve the incident. This was a daytime stop on a busy street. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on how cooperative both subjects were once the officers arrived. They were frisked based on the caller’s (on scene) statement that one of them had pulled a knife on the other. One rater noted that since neither individual is handcuffed the officers must have generally felt very safe. One rater noted this likely was not actually a Terry stop. Subject Sense of Safety SEATTLE POLICE DEPARTMENT 88 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 89 of 124 DISPARITY REVIEW – PART II When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on how quickly and professionally the officers took control of the scene. They kept the subjects separated to ensure everyone’s safety. The officers never drew their weapons, likely further contributing to a sense of general safety. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on how professionally and politely the officers were controlling the situation. They explained what was happening and why. One rater noted that since there were a large number of cars, a community member would not have known it was a call for someone fighting with a knife. One rater noted if a community member saw how the incident ended, they would be confused why there was such a large response and no apparent official action. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on how the officers listened to both subjects and were polite and professional with both. The subject with the knife seemed to completely understand why the officers were there and what they were doing, since he knew he had pulled a knife on someone. Overall Perception of Incident The reviewers noted that the witness was there on scene and approached officers to explain why he had called. The officers were calm and professional and the subjects all seemed to remain calm and not have an issue with the officers’ interactions. If community members knew a call had come in saying someone had a knife, they would understand the immediate nature of the response by officers and appreciate how quickly they de-escalated. Incident #9 involved a call from a retailer of a man shoplifting items from a store, including an ice axe. When the officers located the subject, he was brandishing the axe, walking through a heavy commercial/residential area, and not responding to officer directions. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the axe was a very dangerous and deadly weapon and the subject seemingly not responding SEATTLE POLICE DEPARTMENT 89 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 90 of 124 DISPARITY REVIEW – PART II to officers at all elevating concerns to feeling completely unsafe for themselves and others. One rater noted that the officers attempted de-escalation and once it was apparent the subject was nonresponsive, that is when they determined the threat level was high enough for action. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on how until the moment he was tackled, the subject likely felt safe and somewhat in control as the officers were keeping their distance and talking to him. The pointing of a firearm likely elevated his concern, if he was cognizant enough to realize that was happening. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on the concern about the displaying of a deadly weapon and the subject seemingly completely ignoring the officers. One rater noted that many community members would be more concerned about their safety than that of the individual brandishing the axe. They noted officers clearly did not want to hurt the subject. One rater noted the community members’ reactions were calm or disengaged, suggesting they were not that concerned. One rater made the point that some might have been concerned that the officers let the situation go on so long, putting them at risk. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on how much time officers gave the subject to comply and how generally calm and professional they were when following him. If the subject was coherent to understand what was happening, the officers made it clear they were going to give him time and space, and didn’t want to harm him. They used professional language. Overall Perception of the Incident Reviewers noted that anyone witnessing a subject walk down the street swinging an ice axe and not responding to the officers’ commands would understand their behavior and the nature of the resolution. One reviewer noted that some community members might even question why the officers allowed the subject, armed with a dangerous weapon, to walk around so much of the area before interceding. Incident #10 involved a call about a reported shoplift at a drug store. When the subject matching the description was located the officer stop his patrol car and asked the subject to stop. The officer asked the subject to put the goods in his hands down, and to keep his hands out of this pockets. As the subject had a hard time keeping his hands out of his pockets, the officer offered to frisk him if he wanted, so he could put his hands in his pockets. SEATTLE POLICE DEPARTMENT 90 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 91 of 124 DISPARITY REVIEW – PART II Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the subject’s initial behavior of repeatedly reaching into his pockets. Though no weapon was reported, his decision to not follow the officer’s directions concerned one rater. Once the subject was frisked and no weapons were found, the responder stated that the rest of the encounter felt very safe. Another rater noted the importance of a backup officer arriving and the effect it has on the scene in terms of ensuring safety. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on how the officer always was calm and never elevated the incident. One rater credited the officer with offering the subject a frisk if he wanted to put his hands in his pockets. One rater noted that the only agitation was the subject being frustrated that the officer didn’t believe him that he didn’t steal the product. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on how the officer and subject were both calm and never seemed concerned about their safety. A random community member likely would think the officer had a reason to stop and speak with this person, and that they were doing it in a very professional and polite way. One rater noted that some community members, if they didn’t know what the other officers were doing, would wonder why the detention was lasting so long. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on how politely and calmly the officers explained what was happening and why. One rater noted they were firm, but polite. The subject only seems annoyed that they don’t believe he did not steal anything, but seemed to be ok with how the stop was proceeding. Overall Perception of Incident The reviewers noted that if you knew the person had been in the store and the store had called to report a theft, you would understand why he was stopped. They noted the calmness and professionalism of the officer – citing his suggesting a way that the individual would be allowed to put SEATTLE POLICE DEPARTMENT 91 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 92 of 124 DISPARITY REVIEW – PART II his hands in his pockets since it was cold. If you didn’t know the preceding facts, you would wonder why this individual calmly waiting for the bus was approached by officers. It could lead to wondering if was because how he appeared. Reviewers noted that the fact that the misunderstanding started with the store management would be an important part for the community to understand. Incident #11 involved a call for an armed robbery. The car had been located and the patrol vehicle was part of the team that stopped the car. When the first patrol vehicle pulled up on the subject vehicle they waited for backup and then slowly ordered those in the car out of the car one at a time, slowly moving them back to the officers. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the safe distance at which the officers stopped the car, and how the subjects complied with the officers’ orders. Since it was a high-risk stop, their response was reasonable and within training. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused the calmness of the officers’ voice and the clear instructions. It reasonably made everyone involved feel like the situation was under control. There was a perception that the size of the response would make anyone concerned about how the incident would unfold. One comment was that if either subject was involved and knew that the officers thought they had a gun, they would understand the response, but maybe be concerned about how it would play out. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on the large number of guns displayed and the perception that would create that someone was likely to be shot. Another reviewer noted that professional nature in which the officers took the subjects into custody, and how clear their directions were, along with how calm they were throughout the incident. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on the how the officers explained exactly what they were doing. It was noted that they tried to make the handcuffs as SEATTLE POLICE DEPARTMENT 92 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 93 of 124 DISPARITY REVIEW – PART II comfortable as possible. Given the alleged behavior of the subjects, the officers professionalism likely prevented the subjects from experiencing in significant use of force. Overall Perception of the Incident There was the general perception that without knowing the nature of the call, a community member would think it was a strong response to a traffic stop. Even if that was the case they would note the professional nature with which everyone was treated. They noted the extreme care the officers took to not injure anyone. Incident #12 involved an incident where a patrol officer comes upon an individual, during a rainy night, appears to be trying to pry open a locked gate. The officer approaches the person and investigates if they were attempting to enter the area. The individual alleged to be living there in a tent with permission. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on how the officer approached the subject from behind without announcing their presence. There was some concern that this could have surprised the subject and caused a startled reaction. It was noted this was an obvious tactic to prevent the subject from having time to reach for a weapon. The officer’s calm demeanor – as well as telling a backing officer they could leave – suggested the officer felt relatively safe after doing a quick pat down. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the general nervousness of the subject. Some of this may have been a perception due to some non-English speaking communication issues. A reviewer noted that the subject may have been nervous/amped up because they were surprised by the officer’s approach from behind. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on the manner in which the officer approached the subject and the somewhat demeaning way they addressed the contents of the wallet. There was some concern about not giving the subject the time/ability to calm down and explain what was happening. SEATTLE POLICE DEPARTMENT 93 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 94 of 124 DISPARITY REVIEW – PART II Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on the way the officers regarded the contents of the subject’s wallet. It was noted that it was unlikely they would have made some of the comments to someone who spoke better English. Overall Perception of the Incident The raters noted that anyone who witnessed the entire interaction would note that the person was behaving suspiciously at the fence, which likely warranted the officer checking the situation out. There was generally agreement that approach of “surprising” the subject likely carried some risk that could have been avoided with different tactics. As noted elsewhere, there was concern in how the officer interacted with the individual, particularly around a style of communication that seemed to be connected to his not very clear English. Incident #13 involved an incident where officers were responding to call of a man having slashed a man at a food bank with a knife. When officers located the man on the street they commanded him to the ground at gunpoint. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the nature of the call – a man armed with a knife. When the officer located the subject, they complied with the orders and did not appear to immediately have access to a weapon. At first the officer did appear to be concerned about ensuring the safety of all in the area, but once the subject was compliant the tension of the situation calmed down. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the response of the subject who it appears takes a second to understand what is happening and then complies and goes prone on the ground. The officer’s initial demeanor may have made the subject concerned about how they were going to respond – especially if the subject did not just cut someone or was not aware of their behavior. Having an officer run at you with their gun drawn and pointed at you would cause anyone to be concerned for their safety. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on the officer running up the street yelling and pointing their firearm. If a community member was not aware that anything had happened in the area, they would be extremely concerned and frightened by SEATTLE POLICE DEPARTMENT 94 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 95 of 124 DISPARITY REVIEW – PART II anyone – including an officer – running by them with a gun pointed. The intensity of the interaction would likely lead a community member to conclude something serious had happened and this was not a random interaction. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on the way initial interaction. Once the officer had the subject handcuffed, and when back up officers were there, the scene calmed down. The initial interaction was very intense, but the officer never acted unprofessionally. Overall Perception of the Incident Anyone observing this incident would be startled by how it began, but would note that it resolved very calmly and with little to no force being used beyond the pointing of the firearm. The officers used normal, but loud, language. As soon as the subject was secure the scene took on a much different tone. If you did not know they were looking for someone with a knife, you would wonder why they ran at the man with a gun drawn. One comment noted that even if you knew that, it wasn’t apparent that the subject had a knife anywhere reachable and it may have been severe to run past civilians, on a sidewalk, with a gun ready to fire at the subject from more than 20 feet away. Incident #14 involved an incident where officers were responding to call of a man having threatened another man with a gun. When officers arrived on scene having located the suspected automobile, they saw the subject walking away slowly on the sidewalk holding a phone. They first cleared the car – which had a passenger, and then arrested the subject at gun-point, including a rifle. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the fact that officers were responded to a call of threats made with a gun. There were comments that the subject’s general behavior and demeanor didn’t present an obvious threat, but that the call had obviously made the officers concerned about their safety and that of the public given their belief the person had a gun. There were comments that in all of the incidents reviewed, the officers here seemed the most “amped” up, when it seemed to be a relatively calm scene other than the allegation of a gun. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the idea that until the officers approach the subject, he appears to have no idea that officers may be looking for him, as he is calmly walking down the road looking at his phone. Others commented SEATTLE POLICE DEPARTMENT 95 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 96 of 124 DISPARITY REVIEW – PART II that if the subject had engaged in the alleged behavior, the style of walking and nonchalantly looking at the phone may have been intentional to not call attention to himself, as he more than likely heard all of the sirens and saw the police car drive by before it stopped and recognized his vehicle. Once the officers apprehended the subject he is immediately compliant, though likely somewhat concerned given he sees a pistol and a rifle pointed at him from short range. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on if a community member knew that the call was for a threat with a gun. If someone did not know this, the officers seemingly randomly approaching this individual – who would appear Hispanic of Native – with guns drawn, including a rifle, would make a community member concerned for everyone’s safety – not knowing the cause of the event or the level of danger. If someone did know the nature of the call, they would still be concerned about the potential for the discharge of a firearm. The officers’ use of harsh language may negatively have influenced a community member’s perception of the event. Subject’s Sense of Fairness and Dignity When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on the language the officers used and the presence of the long gun. While both decisions were within policy and easily understandable to those familiar with law enforcement, the subject may have felt targeted and demeaned/threatened by the experience. The other comments noted that the officers did calmly explain what was happening once the situation was under control. At all times they appeared to treat him professionally with no significant uses of force. If he was not involved in the alleged behavior he would definitely be concerned about why such a significant action happened to him, and he may be specifically concerned about the language used. Overall Perception of the Incident Anyone observing this incident would be particularly focused on a rifle drawn in the middle of a street and pointed at a person at close range. They also likely would have been startled by the amount and volume of cursing used during the incident considering to a random passerby the subject was simply walking down the street, looking at his phone. If there was no awareness that the officers had identified the vehicle and the subject as the person alleged to have threatened someone with a gun, the action of the officers would seem disproportional. SEATTLE POLICE DEPARTMENT 96 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 97 of 124 DISPARITY REVIEW – PART II Incident #15 involved an incident where an officer stopped an individual for failing to stop at a stop sign. When the officer – who was with a field training officer – got the individual’s paperwork, he discovered that the driver’s name did not match the registration/bill of sale. He eventually had the individual get of out the car to talk, and during that time the field training officer quickly removed/frisked a visible pocket knife from the driver. Officer Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the officer, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the fact that the officer was apparently conducting multiple stops as part of his training and wasn’t particularly concerned about this apparent stop. He approached the car calmly and spoke professionally and politely to the driver. It was not until the officer noticed that the names did not match that he became a little concerned about. The field training officer was noted as raising the officer’s level of concern and suggesting calling in backup since the paperwork “didn’t feel right.” When the backup officers were on scene, everyone remained calm and professional, but it did seem to raise the tension given that there were now four officers on scene – even though the field training officer does not technically count. Subject Sense of Safety When the internal SPD raters were asked to express how they might have felt if they were the subjects, in terms of observable factors that would have affected their sense of safety, the comments generally focused on the fact that the driver appeared and stated that he was nervous as he had never been stopped before. He appeared not to feel at-risk or targeted, as he admitted to not seeing the sign and tried to explain the whole process of how he came to own the car. When the officers – now four of them – re-approach the car -- the driver’s anxiety is definitely raised given that there are now four officers and he likely assumed they thought the vehicle was stolen and what that might mean. Community Perception of Incident When the internal SPD raters were asked to express how they might have felt if they were a random community member observing the incident, in terms of observable factors that would have affected their sense of safety for themselves, the subject(s) and the officer(s), the comments generally focused on if a community member knew that the call was for a threat with a gun. If someone did not know this, the officers seemingly randomly approaching this individual – who would appear Hispanic of Native – with guns drawn, including a rifle, would make a community member concerned for everyone’s safety – not knowing the cause of the event or the level of danger. If someone did know the nature of the call, they would still be concerned about the potential for the discharge of a firearm. The officers’ use of harsh language may negatively have influenced a community member’s perception of the event. Subject’s Sense of Fairness and Dignity SEATTLE POLICE DEPARTMENT 97 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 98 of 124 DISPARITY REVIEW – PART II When the internal SPD raters were asked to express what their feelings about fairness and dignity would have been if they were the subject(s), the comments generally focused on the language the officers used and the presence of the long gun. While both decisions were within policy and easily understandable to those familiar with law enforcement, the subject may have felt targeted and demeaned/threatened by the experience. The other comments noted that the officers did calmly explain what was happening once the situation was under control. At all times they appeared to treat him professionally with no significant uses of force. If he was not involved in the alleged behavior he would definitely be concerned about why such a significant action happened to him, and he may be specifically concerned about the language used. Overall Perception of the Incident Anyone observing this incident would be particularly focused on a rifle drawn in the middle of a street and pointed at a person at close range. They also likely would have been startled by the amount and volume of cursing used during the incident considering to a random passerby the subject was simply walking down the street, looking at his phone. If there was no awareness that the officers had identified the vehicle and the subject as the person alleged to have threatened someone with a gun, the action of the officers would seem disproportional. SEATTLE POLICE DEPARTMENT 98 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 99 of 124 DISPARITY REVIEW – PART II APPENDIX E Summary of Community Session Feedback Notes from Community Roundtables – Disparity Review – Night One – African American/Black 7 community members 1. Incident – Traffic Stop with a Terry frisk One community member began the conversation by asking if it was normal for officers to just reach and take something from someone without explaining that they were going to do it (the officer reached and took a visible knife from the pocket of the individual involved in the traffic stop). The community member asked what would have happened if the individual had reacted to the officer’s sudden intrusion into his personal space – especially since as far as he knew the only reason he was stopped was for running a stop sign. SPD explained that preferred method would have been to explain that you were about to remove the knife so that everyone could feel safer. The intent being to prevent this from escalating if the individual continued to casually get his hands near his pockets and the knife. Another community member remarked on the number of officers responding and their general demeanor as being a little apprehensive. This was noted in regard to the nature of the stop, the calm behavior of the stopped individual, and the nature of the day – bright blue sky, in a semi-busy area of the city, where the officers were not out “on their own.” Another community member asked if it was legal for officers to just frisk someone when you have taken them out of a car for running a stoplight. A couple community members stated they were surprised that the actions taken in the case were legal, given the reason for the interaction. SPD explained that there are “frisk factors” (articulable reasonable suspicion) that together allow an officer to legally stop and frisk an individual. Another community member spoke about the inherent white privilege they saw in the video. Up until the officer was at the side of the car and had the person’s clear id – determining they were Hispanic – they felt the nature of the stop changed on this discovery. Another community member focused on how the Field Training Officer seemed to be pushing the officer to conclude something “wasn’t right” or “was wrong.” They noted that this may be the transmission of SEATTLE POLICE DEPARTMENT 99 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 100 of 124 DISPARITY REVIEW – PART II the FTO biases. Suggestions were that FTOs shouldn’t talk about feelings, but teach the officer what factors to make sure they consider. Another community member mentioned that the officer seemed to be calm and just treating it as a traffic stop, and was very polite and gentle with the individual. Then the FTO seemed to ramp up the situation. Another community member mentioned the issue of the officer immediately viewing the individual’s work tool – the knife – as a weapon. And, that by grabbing it and throwing it in the individual’s car, the officers was stripping the person of their humanity. Reducing their work/part of their value to a weapon. In relation to the officer tossing the individual’s knife into the car, and then another officer stating that “that’s ok that’s not what we are worried about,” when the individual explained why they had the knife, community members stated general confusion about why they bothered with the knife at all. SPD explained that removing the potential weapon – whether legal to possess or not – was to keep everyone safe so that the officers wouldn’t have to be concerned if the individual’s hand got near the knife again. There was a lot of discussion around the need for police to consider the history of policing, in terms of how individuals of color are experiencing even these more mundane interactions. Perception/feeling that when officers address the poor and disenfranchised officers take on a more warrior approach, while the guardian approach with white people and others with privilege. In general, community members stressed the need for officers to “explain, explain, explain” what they are doing. SPD totally agreed that more of the training needs to focus on this. The community members also urged that for real change to occur more officers need to look like and come from the neighborhoods they are policing. SEATTLE POLICE DEPARTMENT 100 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 101 of 124 DISPARITY REVIEW – PART II 2. Incident –Stop looking for Shooting Suspect – Safeway One community member began the conversation asking how the officers could have known for certain that the person they were looking for was in the car. That they should have been 100% certain to roll in the way that they did. Those young people were doing what so many do – comfortable in their community – and the police just rolled in and started yelling to put their hands on their heads, and took them out of the car and handcuffed some of them. SPD explained that the detectives – who appear later after the patrol cars have been used to initiate the stop (so that there is no confusion that it is the police, as the plain clothes detectives sometimes can cause confusion) – got up to the car as they knew exactly who they were looking for. Once they knew he wasn’t in the car, they started the process of letting everyone go. One community member noted that given this is in the middle of a community shopping area, along with the large police presence, you are immediately stigmatizing the young people as obviously being involved in something serious. One community member suggested that the detectives should have been ready to get up to the car as soon as the stop was initiated so they could determine if the person was in the car without the need of removing everyone. One community member noted that they seemed to only handcuff some of the individuals, noting that since they were in a car, there was no way of knowing who had the suspected gun on them. SPD explained that the detectives knew they were looking for male so they had no reason to handcuff the females. The community members were confused, since if the idea is about safety, and you knew the supposed shooter wasn’t in the car, but if he was anyone could have the gun, why only handcuff some. Seems to be that the handcuffing was for another reason. One community member noted that a large part of the action takes place outside of the ICV. SPD noted that this was before BWV, and that the camera in the other car also did not capture the area where the two other individuals are being addressed. One community member asked if there might be a reason why the officers moved out of view. SPD noted that this is why the BWV is so great; you capture the audio and video of every individual interaction. One community member noted additional concern about the method of approach. Did the patrol cards need to come in the way they did and leave their lights on. It was fairly obvious none of them were concerned about being discovered by the police. If the shooter had been there, would they have been sitting in the parking lot hanging out, and do we really think they would have a weapon on them. The nature of the interaction obviously lessened the dignity of those involved. What was SPD doing to restore the personal dignity of those individuals. SPD explained that most of the tactics were designed SEATTLE POLICE DEPARTMENT 101 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 102 of 124 DISPARITY REVIEW – PART II to prevent a pursuit where all involved – including the general community – are placed at heightened risk. There was a lot of discussion about the exchange where the officers explain that they are looking for a shooting suspect and these individuals matched the description. One of the involved young people notes that “they always seem to match the description.” The community members then had a conversation about how it is always that way. The phrase to provide a reason/legitimization for stopping people of color always seems to be “matching the description.” Several community members discussed the very nature of the stop. Given that the detectives/officers were looking for a suspect that had been involved in shots fired call(s) in the preceding days – this casts a wide net for a shooter who seems to resemble a large segment of frequently policed individuals. One community member noted that no one in the parking lot seems to be concerned about this happening – because it happens so routinely. The room had a discussion about whether the results of the stop mattered. That is, if the shooter was in the car would the reaction have been different. The overwhelming response was that, no, it would not matter. This still happens too frequently – assuming someone is a suspect because they are a young black male. This could have resulted in finding the suspect; it also could have resulted in someone being hurt if the young people had not been calm and generally responsive to the officers. SEATTLE POLICE DEPARTMENT 102 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 103 of 124 DISPARITY REVIEW – PART II Notes from Community Roundtables – Disparity Review – Night Two – Native/Indigenous/Latinx 9 community members 3. Incident –Shots Fired call; Stop with Firearm pointing and two Frisks The discussion began with a community member asking if it was standard/allowed to handcuff someone before they know why they are being arrested. SPD explained that this at that moment the individual was not technically under arrest, and the handcuffs were a safety measure given the nature of the call, and the heightened potential that a firearm was involved. Another community member noted that the location where this occurred – in front of a busy medical facility (one with a cancer treatment center as well) was a poor choice. The stress and trauma from a stop with multiple cars, lights flashing, and multiple officers pointing their firearms and yelling was not a good idea. One community member remarked on the size of the response. They counted approximately 15 officers and 6-8 cars on the scene. It was stated this was both unwarranted, escalating, and a waste of resources. SPD explained that the procedure, especially on a call with a potential weapon, is to have overwhelming deployment at the beginning of the call to minimize the likelihood that anyone will feel their last and only resort is deadly force to control a situation. Once the scene is rendered safe, it is the supervisor’s job to immediately start releasing officers from the scene as soon as possible. One community member noted that generally the female officers on the scene seemed to be more calm and not escalating the situation through their tone of voice, commands, or physical actions. The community member noted that having more female officers on the scene seemed like an important element given that factor Another community member asked if this stop would have proceeded differently if the subjects would have been darker skinned. There were concerns that given the video involved a light-skinned subject of reported Hispanic heritage, it was not highlighting the issues that occur in stops/interactions of those individuals. SEATTLE POLICE DEPARTMENT 103 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 104 of 124 DISPARITY REVIEW – PART II One community member asked if the dispatcher checks any online resources to see if individuals have reported other factors. This was asked in response to the fact that in this incident the officers are looking for someone based on the general description of one witness. SPD explained that dispatchers do not have time to review online resources, and, in most cases, do not have permission to access many of the social media systems due to privacy/surveillance restrictions. A large part of the remaining discussion was on the nature of the stop and whether the individuals who were stopped sufficiently matched the description offered by the witness. The witness, in general, described the clothing the two subjects, the gender of the two subjects, and the perceived race/ethnicity of the two subjects. When the officers find the two individuals they decide to stop, the gender of one is different and the only item of clothing matching the description is one subject’s shoes (which are a common color combination for a popular style of shoe). Another community member asked if there was a minimum amount of a “match” that the officer must have to stop a person. SPD explained that other than shoes it is easy to change clothes, so they often rely on shoes as part of the match. SPD stated there was no “scoring matrix” for when there was a “sufficient” match, but that, rather, it is about articulable reasonable suspicion that the person matches the description of someone who ,or is, about to/is/just committed a crime. A community member noted that the witness, who was concerned that the subject had a firearm, (or in other types of cases), may have been under stress and just taking that description as a reason to stop two people with guns drawn, seems to not be right. SPD stated that if a person thinks they were incorrectly/inappropriately stopped, the sergeant reviews them all and the subject can file a complaint. Another community member noted that the officers generally tried to explain what they were doing, but this was not until a little way into the stop. They stressed that explaining why and what would go a long way to keeping things from going badly or escalating. During a stop, and particularly if you know you did nothing wrong, your adrenaline is flowing, you are nervous, and maybe angry, especially if it has happened to you before. The instructions need to be given in way that is more than going through the steps. Officers need to think about how each interaction could make the person feel and work that into the explanation. SPD explained that they have been focusing more and more on making sure officers professionally explain what and why they are doing what they do. Acknowledged that more can be done – especially in the style of communication. Unintentional actions cause people to panic. A couple community members noted that it likely was confusing for the subjects given that multiple officers were giving commands. They noted that if a person got confused and did the wrong thing, that is how things go wrong. SEATTLE POLICE DEPARTMENT 104 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 105 of 124 DISPARITY REVIEW – PART II SPD explained that the training around that has changed; the new model of “contact and cover,” automatically places the primary officer in the role of talking to the subject(s) with any back-up officers securing the scene, protecting the officer and subject(s), and engaging any external individuals. During the course of this discussion, a community member raised the need for people who had been arrested by the police to be included in these discussions. Many in the room had family who had been, but few had personal had the experience. All agreed this was a good recommendation for future sessions. This was noted because if you have been and you know what all is involved, the beginning of the process can be even more stress inducing. One community member recommended that officers need to undergo sensitivity training. They noted the aggression coming from the voice and actions of one of the officers whose body camera we watched. The community member noted that the officer could have been far more level-headed and calm. Another community member noted that repeatedly telling someone who has been stopped – especially if they know they have done nothing wrong – to “calm down,” can be very enraging/escalating. This is particularly important in the face of behavioral health needs. Officers need to be held to a higher standard about how they talk/act and what they say. The public is not trained to deal with all of that stress. One community member also inquired about why there were so many guns drawn. Was it really necessary to have so many guns on two people, especially given it was in the middle of such a busy area. Several community members discussed that having more officers who look like the people they are policing and come from/know the neighborhoods they are policing can help with a lot of these issues. Some community members noted that it is asking a lot for their kids, brothers, sisters, etc., to place themselves “at risk” to help fix a department if the majority people in it can’t do the work. SEATTLE POLICE DEPARTMENT 105 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 106 of 124 DISPARITY REVIEW – PART II Notes from Community Roundtables – Disparity Review – Night Two – Native/Indigenous/Latinx 9 community members 4. Incident –Narcotics Activity with a Frisk This incident resulted in the group expressing that it had thought it would watch videos of individuals who looked expressly like them, and interactions that were more intense. One community member noted that they saw nothing wrong with this stop – because this is what happens when White or light-skinned “passing” individuals encounter the police. Others stated that part of the issue is that the officers are not from the areas they patrol and do not know or have the individualized trust of the people they serve. One community wanted to confirm that the department is not just doing all the new training with the tenured officers. They also asked if SPD is doing training around autism. Also, they wanted to know if SPD is engaging young people in schools. Another community member mentioned that they know from their experience and that of their friends and family, that if you aren’t light skinned, you will get the “Mexican treatment.” Another community member noted that in their personal experience, officers see native/indigenous people as drunk/crazy/armed and that leads to disparity in treatment. There needs to be cultural training. One community member noted that officers are not going to say or write something that is going to get them in trouble. They are going to find a way to explain what they need to. SEATTLE POLICE DEPARTMENT 106 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 107 of 124 DISPARITY REVIEW – PART II Notes from Community Roundtables – Disparity Review – Night Three – Asian / Pacific Islander 12 community members 5. Incident – Trespassing / Burglary On-view As this was a frisk incident during a Terry stop, the group first inquired about why/how the officer so quickly proceeded to frisk the individual. SPD explained that the nature of the stop – dark, solo officer, bulky clothing on the subject – they all contribute to the articulable reasons for frisking the individual for safety. The community group was curious why this was charged as a Burglary. SPD explained that attempting to enter a secured premise for the intent of committing a crime is the definition of Burglary. There was a large discussion around officer discretion in how to and whether to charge. One community member had questions, later reflected by others, on the method in which the officer approached the individual. They came from behind the individual, not announcing they were there until they were right behind them. There was concern that someone could instinctively react, in a very protective way, to being surprised like that. There were concerns about the effects of startling someone, especially in the dark. SPD explained that the officer may have been trying to limit the opportunity for the subject to access a weapon – by not letting them know they were there. The official report referenced the likelihood that someone attempting to break into a location is likely to have burglary tools, which could be used as a weapon. SPD further explained that one approach might have been to “spotlight” the individual from the car and wait for backup. Another community member, later echoed by others, focused on the tone of the officer on the approach. The officer was very loud and stern – viewed as aggressive. The group had some concerns about how different cultures would handle/react to such a vocal approach. This progressed into a discussion around how in this incident, the officer quickly came in and acted as if the individual was known to be an issue. There was little opportunity for discussion. One community member noted that after the situation was in-hand, the officer seemed to be deeming or making fun of the individuals intelligence/language skills. There were questions, viewed as inappropriate, about what was in his wallet. There was more discussion about how SPD handles incidents where the subject does not understand English well. The group stressed how it is the officer’s responsibility to know how to “bring it down a level.” They should not be ramping up the intensity of the situation. A later discussion revolved around how the department handles getting Miranda rights understood if the person cannot understand English. SEATTLE POLICE DEPARTMENT 107 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 108 of 124 DISPARITY REVIEW – PART II SPD offered that it is exploring ways to use technology to help close this gap…the group noted that having a more diverse department also would help close this gap, in many ways. There also was a discussion about how the effects of not knowing English play out through the whole process, including knowing he could log a complaint if he felt he has been mocked, or more so, following through with the complaint process. Another community member asked if there was not an assumption about this person and what they were doing because of their obvious homelessness status. Was it just assumed they were doing something wrong. Is that why the officer wanted to know if he had a history or was DOC? Would an officer ask everyone that? One community member inquired about whether there is training around how to speak, including tone of voice and language. Does this include cultural/sensitivity training. Another community member noted that they agreed with the initial stop, though they definitely agreed many elements of the stop could have been done better. One community member inquired about the Body Worn Camera policy and the fact that the officer did not apparently alert the person why they were being recorded. There was some suggestion that the individual’s language issues may have been involved in that decision. SPD explained that they are required to do that, but an officer can forget and it is the responsibility of the supervisor to address the issue. SEATTLE POLICE DEPARTMENT 108 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 109 of 124 DISPARITY REVIEW – PART II 6. Incident –– Threats with a Gun A number of the community members immediately focused on and had questions about whether it was normal for officers to curse so much in an incident like this. They felt it was degrading to the person and likely escalating the situation. There were a lot of question about how that could be seen at all as deescalation. SPD explained that while not every officer/supervisor would be ok with that approach, it is the experience of some that in certain situations people only respond to that kind of forceful language. They are attempting to control the situation as best they can so they do not have to use force. The community members expressed that this wasn’t de-escalation, it was overwhelming force. The officers had a rifle, a handgun, and loud, cursing words to ensure the subject felt that they were nothing and had no rights. They were over-powering the situation, not de-escalating it. The community members had questions about the policies/training on using different firearms. In the incident one officer has an assault rifle pointed at the subject, who is visibly not holding a weapon, about 6 feet away. SPD explained that only certain officers are trained on rifles/shotguns. They would likely discuss/decide on their way to the call if they were taking the specialized gun with them. Once they take the gun out of the car with them, they are committed to that gun. They aren’t going to fling it over their back or drop it, so they can use their sidearm handgun. Some of the community members noted that from the tone of the officer’s voice, the words used, and their actions, it was apparent that the officer(s) were dealing with stress and trauma. The community members were concerned that a prior call or some other incident had the officer(s) hyped up and was leading to the elements of this interaction…and putting the subject and others at risk. The officer was pissed and ready to shoot/kill somebody was the reaction of the room, in general. The community asked that officers have trauma follow-up sessions and get training in stewardship and how to manage and identify that stress. SEATTLE POLICE DEPARTMENT 109 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 110 of 124 DISPARITY REVIEW – PART II 7. Incident –– Assault with a Knife Several community members initially focused on the size of the response by the police. There were at least 4 police cars visible in the video, for an apparent older man alleged to have stabbed/slashed a person with a knife. SPD explained that when there is a call of someone with a weapon, particularly who has hurt another person, a lot of officers are going to show up to ensure no one else gets hurt. It is then the supervisor’s responsibility to release officers as soon as the scene is secure. One community member was curious as to the protocol/training for using a firearm when the issue is the person might have a knife. In this incident the officer comes upon the subject walking up a sidewalk, with no knife visible, and the officer is yelling, pointing the gun, and ordering the person to lay down on the ground. The community member was wondering why they got close enough that a knife would be dangerous and necessitate pointing the firearm at the subject. SPD explained that when someone has allegedly used a weapon to threaten or injure another person, the officer is going to attempt to ensure that no one else gets hurt by taking the person into custody as quickly as possible, without putting themselves at undue risk. There was a general conversation about the need for officers to have awareness of the culture of the people they are encountering. There was a conversation around how the very thing many Pacific Islanders are “recognized” for (i.e., being “a big boy”) are the very things that make police officers scared of them on the street. The community stressed that the best way to close that gap was to have more diverse officers who know the community, the cultures, and the values. One community member expressed an interest in having sessions around how an individual stopped on the street/encountering the police could best comply with the officers as well as let them know, if they so, that they are carrying a weapon (gun or knife). Are there things they could know not to do to help keep the situation from escalating. SEATTLE POLICE DEPARTMENT 110 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 111 of 124 DISPARITY REVIEW – PART II APPENDIX F Additional Seattle Police Department Data on Stops and Force 2 Stops, Frisk Rate, and Weapon Hit Rate During the study period, January 1st, 2016 to June 30th, 2018, a total of 19,511 Terry Stops were conducted by the Seattle Police Department. It was reported that during 4,209 (21.6% of all stops) of these stops a frisk for weapons was conducted which resulted in the discovery of a weapon in 840 (20.0%) of those frisks. Figure 1: Stop, Frisk, and Weapon Hit Count Time Series More than 80% of Terry Stops occurred during Priority 018, 1, or 2 events, 37% (931 stops). Most frisks for weapons occurred during Priority 1 stops (38.5%, 371 frisks) and 22.1% of those frisks resulted in the discovery of a weapon. Though only accounting for 6.2% (219) of all Terry Stops, Priority 7 events resulted in a 22.4% weapon hit rate (see Table 1). 18 Priority 0 indicates a missing relationship to the underlying CAD event. SEATTLE POLICE DEPARTMENT 111 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 112 of 124 DISPARITY REVIEW – PART II Table 1: Stops, Frisks, and Weapon Hit by Call Priority: Most Terry Stops occurred as the result of a dispatched call for service (47.7%, 9,250 stops). 28.5% of these stops involved a frisk which resulted in a weapon hit rate of 20.9%. Of the 15.6% of stops that were initiated with an officer Onview, 20.9% involved a frisk with 17.2% of those frisks resulting in the discovery of a weapon. During the study period this information is unavailable for 37.0% of Terry Stops (see Table 2). Table 2: Stops, Frisks, and Weapon Hit by Call Type The North precinct, SPD’s largest, accounts for the most Terry Stops, with 24.6%. The fewest stops are in the Southwest precinct, which includes accounts for just 7.1% of all stops. SEATTLE POLICE DEPARTMENT 112 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 113 of 124 DISPARITY REVIEW – PART II Figure 2: Terry Stops by Precinct The distribution of Terry Stops occurred fairly evenly across all SPD Sectors, though the data for 27% of stops during the study period do not have a Sector recorded. The frisk rate for stops with recorded Sectors varied between 17.5% and 43.3% while hit rate varied between 13% and 31.1%. Sam Sector reported the highest rate of frisks (43.3%, 324 stops) but had the lowest weapon hit rate (13.0%, 42 stops). Conversely, Mary Sector had the highest weapon hit rate (31.1%, 69 stops) though the sector had the second lowest frisk rate (17.5%, 222 stops), (see Table 3). SEATTLE POLICE DEPARTMENT 113 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 114 of 124 DISPARITY REVIEW – PART II Table 3: Stops, Frisks, and Weapon Hit by Sector Just over half of all Terry Stops recorded during the study period were of White subjects (50.5%). Black subjects accounted for 30.6% of all stops, Hispanics accounted for 4.8%, Native American/Alaskan Natives accounted for 3.3%, and Asians accounted for 2.8%. Hispanic subjects had the highest frisk rate (27.0%) with a weapon hit rate of 18.7%. Though White subjects were frisked at the lowest rate (18.2%), they accounted for the highest weapon hit rate (25.2%), (see Table 4). Table 4:Stops, Frisks, and Weapon Hit by Subject Race Male subjects accounted for more than three quarters of all Terry Stops (77.8%). They were frisked at a rate of 24.5% with a weapon hit rate of 20.9%. Female subjects were frisked at a rate of just 10.9% with a hit rate of 12.3% (see Table 5). SEATTLE POLICE DEPARTMENT 114 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 115 of 124 DISPARITY REVIEW – PART II Table 5: Stops, Frisks, and Weapon Hit by Gender Figure 3: Terry Stops by Gender and Race Male Female Figure 4: Frisks by Gender and Race Male Female SEATTLE POLICE DEPARTMENT 115 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 116 of 124 DISPARITY REVIEW – PART II Figure 5: Arrest is Indicated Disposition by Gender and Race Male Female People between the ages of 26 and 35 were most frequently the subjects of Terry Stops (33.6%) and had a frisk rate of 21.1%. Juvenile subjects accounted for the highest frisk rate (26.1%), though they accounted for only 5.1% of all Terry Stops (989 stops) and had the lowest weapon hit rate (8.5%). Subjects between 18 and 25 accounted for 21.8% of all Terry Stops and were frisked at a rate of 23.0%. The weapon hit rate for this age range was 17.7%. Subjects over the age of 56 accounted for only 4.4% of the total Terry Stops and were frisked a rate of 17.9% of the time (see Table 6). Figure 6 shows the weapon hit rate by age range and race of the stop subject. Table 6:Stops, Frisks, and Weapon Hit by Age SEATTLE POLICE DEPARTMENT 116 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 117 of 124 DISPARITY REVIEW – PART II Figure 6: Rate of Weapon Hit by Age Range and Race During the study period, January 1, 2016 to June 30, 2018, 78.5% (15,322 stops) of all Terry Stops were conducted by White officers with a frisk rate of 21.9%, just slightly higher than the overall frisk rate. In the same time period, 70.1% of all sworn SPD employees were White. Hispanic officers conducted 5.0% (170 stops) of Terry stops while accounting for a similar percentage (5.5%) of sworn SPD employees. Black officers made up 7.8% of sworn SPD officers during the study period and conducted 4.2% of all Terry Stops at a frisk rate of 23.8% (see Table 7). Table 7:Stops, Frisks, and Weapon Hit by Officer Race The gender of officers conducting Terry Stops also appeared to be fairly consistent with the sworn SPD employee gender rates. During the study period 85.2% of sworn employees were make and 14.6% were female. 88.9% (17,347 stops) of stops were conducted by make officers with a frisk rate of 22.0% and weapon hit rate of 20.0%. Female officers conducted 11.1% (397 stops) of stops with a frisk rate of 18.4% and weapon hit rate of 19.6% (see Table 8). SEATTLE POLICE DEPARTMENT 117 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 118 of 124 DISPARITY REVIEW – PART II Table 8:Stops, Frisks, and Weapon Hit by Officer Gender The majority of Terry Stops (75.1%) were conducted by officers assigned to 911 Response units and a corresponding percentage of frisks (78.6%) were also conducted by officers functionally assigned to 911 Response (see Figures 7 and 8). Figure 7:Percentage of Terry Stops by Officer Assignment Figure 8: Percentage of Frisks by Officer Assignment 3 Use of Force and Firearm Pointing Rate During the study period there were a total of 4,392 Use of Force (UoF) incidents that involved 917 incidents of firearms pointing (20.9% rate). Three quarters (76.7%) of all UoF reported during the study period were Level 1 – Use of Force incidents. These incidents had a rate of firearm pointing of SEATTLE POLICE DEPARTMENT 118 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 119 of 124 DISPARITY REVIEW – PART II 26.2%. Level 2 – Use of Force incidents accounted for 21.6% of all UoF and were associated with just a 3.0% rate of firearm pointing (see Table 9). Table 9: Use of Force Count and Firearm Point Rate by UoF Incident Type The majority (70.7%) of UoF incidents occurred during a dispatched call for service while 29.2% occurred when an officer initiated the contact (Onview). The firearm point rate for dispatched calls was 19.7% while the rate for Onview incidents was 23.9% (see Table 10). Table 10:Use of Force Count and Firearm Point Rate by Call Type Priority 1 calls accounted for 40.4% of all Uses of Force and involved a firearm point rate of 25.5% (1,776 incidents). 32.3% (1,417 incidents) UoF incidents were initiated during a Priority 2 event with a firearm pointing rate of 13.6%. Though Priority 9 events accounted for only 1.5% (68 incidents), they had a firearm pointing rate of 25.0% (see Table 11). SEATTLE POLICE DEPARTMENT 119 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 120 of 124 DISPARITY REVIEW – PART II Table 11:Use of Force Count and Firearm Point Rate by Priority The largest percentage (39.4%, 1,732) of all Uses of Force involved a White subject, but incidents with a White subject had the lowest rate of firearm pointing (16.6%). Hispanic subjects had the highest rate of firearm pointing UoF incidents (37.8%) but accounted for just 3.4% (148). Black subjects were involved in 32.1% (1,408) of all UoF with a firearm pointing rate of 26.6% (see Figure 12). Table 12:Use of Force Count and Firearm Point Rate by Subject Race 77.5% (3,405) of all UoF incidents involved a male subject, these UoF had a firearm pointing rate of 25.1%. Female subjects accounted for 20.4% (898) of UoF incidents with a firearm pointing rate of 10.6% (see Table 13). SEATTLE POLICE DEPARTMENT 120 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 121 of 124 DISPARITY REVIEW – PART II Table 13:Use of Force Count and Firearm Point Rate by Subject Gender Subjects between the ages of 26 and 35 were involved in the highest percentage (30.3%) of UoF incidents. The highest firearms pointing rate was 28.3% and involved subjects age 18 to 25. Juveniles (age 1-17) were involved as subjects of 253 UoF incidents, 5.8% of the total, and had a firearm pointing rate of 26.9% (see Table 14). Table 14: Use of Force Count and Firearm Point Rate by Subject Age White officers accounted for 78.1% of UoF incidents (3,432) during the study period and had a firearm pointing rate of 21.1%. Native American/Alaskan Native officers accounted for just 1.5% (51) UoF incidents but had the highest rate of firearm pointing at 37.3%. Black officers were involved in 5.4% of all UoF incidents (189) and had a firearm pointing rate of 21.2% (see Table 15). SEATTLE POLICE DEPARTMENT 121 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 122 of 124 DISPARITY REVIEW – PART II Table 15:Use of Force Count and Firearm Point Rate by Officer Race Though male officers were involved in a much larger percentage of all UoF incidents (89.6%, 3,935) than female officers, the male officers had a slightly lower firearm pointing rate (20.8%). Female officers were involved in 10.4% (457) UoF incidents and had a firearm pointing rate of 21.9% (see Table 16). Table 16:Use of Force Count and Firearm Point Rate by Officer Gender Though 21.5% (943) UoF incidents did not have a Sector assigned in the report. Of those incidents that did, King Sector had the largest share of incidents with 331 during the study period (7.5%) but King Sector also reported the second lowest rate of firearm pointing at 10.0%. Frank Sector reported the highest firearm pointing rate at 35.0% (55 incidents) but had only 3.6% (157) of the total UoF incidents. William Sector had only 1.6% (71) of the total UoF incidents but had the second highest firearm pointing rate with 29.6% (see Table 17)19. The process of “geocoding” force is manual. Some lag exists in preparing geocoded data. All force will be geocoded and all sectors available for reporting in the Use of Force Outcomes report due January 2020. 19 SEATTLE POLICE DEPARTMENT 122 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 123 of 124 DISPARITY REVIEW – PART II Table 17:Use of Force Count and Firearm Point Rate by Sector SEATTLE POLICE DEPARTMENT 123 Case 2:12-cv-01282-JLR Document 600-1 Filed 12/31/19 Page 124 of 124 DISPARITY REVIEW – PART II APPENDIX G Internal SPD Review Survey Instrument 1. What was the case number? 2. Imagine yourself as the officer, how safe did you feel during this interaction? Were there behaviors you observed or facts you were given that gave you concern? That made you feel less concern? Cite facts, impressions, observations of the subject, the scene and the case, generally, that support your response. 3. Imagine yourself as the subject, how safe did you feel during this interaction? Were there things the officer did that made you feel more or less safe? Cite facts, impressions, observations of the officer or officers, the scene and the case, generally that support your response. 4. Imagine yourself as a member of the community, observing this event. How safe did you feel during this interaction? Did you fear for your safety? The safety of the officer? The safety of the subject? Were there things the officer(s) or subject(s) did that made you feel more or less safe? Cite facts, impressions, observations of the subject(s) or officer(s), the scene and the case, generally that support your response. 5. Imagine yourself as the subject, do you feel as though you were treated fairly, with dignity? Cite facts, impressions, observations and examples of the officer’s behavior that support your response. 6. Imagine yourself as the officer, if this event were observed by a member of the community. Would your actions or behaviors be taken out of context or misunderstood or do you think they speak for themselves? Were they justified? Cite facts, impressions, observations and examples of the officers’ behavior that support your response. SEATTLE POLICE DEPARTMENT 124