Final Report
February 2022
USE OF FORCE ASSESSMENT OF
THE SAN JOSÉ POLICE
DEPARTMENT
Daniel S. Lawrence, Tom Christoff, Zoë Thorkildsen
This document contains the best opinion of CNA at the time of issue.
Suggested citation:
Lawrence, D.S., Christoff, T., Thorkildsen, Z. (202 2). Final Report: Use of Force Assessment of the San José Police Department. Arlington, VA: CNA Corporation.
CNA Document Number: IIM-2021-U-030446
Assessment Team:
Thomas Christoff, PhD
Jessica Dockstader, MA
Monique Jenkins, MA
Daniel S. Lawrence, PhD
Cody Stephens, PhD
Zoë Thorkildsen, MA
Distribution:
Unlimited.
February 2022
ii
Acknowledgements
CNA acknowledges the assistance we received from the San José Independent Police Auditor, the San José Police Department, and San José community members and organizations in completing this report. The documents, data, and input we received were crucial to the development of our findings and recommendations. In particular, we would like to thank Shivaun Nurre, Erin O’Ne ill, Telina Barrientos, and Chief Anthony Mata f or their coordination throughout the audit; community members who participated in listening sessions and focus groups ; and San José Police Department personnel who participated in interviews. Finally, we would like to thank Lt. Steve Donohue, who provide invaluable insights and worked quickly to fulfill every data request we sent his way.
iii
This page intentionally left blank.
iv
Executive Summary
Recognizing the urgent need for transparency, accountability, and legitimacy, the Mayor and City Council of San José, California directed staff to obtain an assessment of the San José Police Department’s (SJPD) use of force. CNA’s Center for Justice Research and Innovation was chosen through a competitive bid process coordinated by the Independent Police Auditor (IPA). This work coincided with an assessment of the SJPD’s efforts to bring the department in line with the recommended best practices promoted in the President’s Task Force on 21st Century Policing report and was completed by members of the same team. The use of force assessment focused on four key areas:
- A review of the SJPD’s use of force policies, procedures, training, and events. - An examination into the characteristics of use of force events, including disparity across racial and ethnic groupings.
- The impact COVID-19 and social justice movements for policing reform had on calls for service and use of force.
- Disparity in use of force behaviors and sustained injuries across racial and ethnic groupings.
In this executive summary, we present a summary of the findings of our assessment and a summary of the key recommendations offered to SJPD and the city. We encourage interested individuals to read the details in the body of this report, where they will find detailed the supporting evidence associated with our 39 findings and 51 recommendations. See Appendix B for the full list of findings and recommendations.
Through interviews, document reviews, community listening session, and data analyses, the team discovered the following key themes:
Summary of key findings
• Many of the SJPD use of force policies reflect best practices in the field. • The SJPD Duty Manual does not define levels of resistance and does not consistently indicate which level of resistance would justify various force options. • The SJPD does not have a use of force review board.
• The SJPD does not provide sufficient clarity in the definition of force. • The SJPD does not provide sufficient clarity on some elements related to electronic control weapons.
• The SJPD does not provide sufficient post -incident guidelines for officers, particularly f or incidents involving lethal force.
v
• The SJPD Duty Manual does not contain sufficient instruction related to officers’ duty to provide medical attention.
• The SJPD’s current use of force reporting system is outdated. • Black and Hispanic community members are arrested more frequently than would be predicted based on their proportion of San José ’s population compared with white community members ; however, among those arrest ed, use of force levels for Black and Hispanic community members are similar to white community members. • The amount of u se of force events was not significantly affected by COVID -19 and social justice movements in early 2020 . B oth the number of calls for service and arrests were significantly lower following these events.
• SJPD officers treated Black and Asian community members similar ly in use of force events compared with white community member s in regard to the amount of use of force, the severity of force, weapon discharges, community member injuries, the amount of inju ries, and the severity of injuries. On the other hand, SJPD officers were found to use more types of use of force among Hispanic community members that resulted in more severe injuries compared with white community members.
Summary of Key Recommendations
• The SJPD should better define levels of resistance and state the minimum resistance level needed for each use of force option.
• The SJPD should create a force review board or unit to identify policy, training, equipment, and personnel implications and include community representatives as part of its efforts. • The SJPD should adopt a “physical coercion against resistance” definition of force. • The SJPD should provide concrete prohibitions on the use of electronic control weapons. • The SJPD should revise the Duty Manual to provide comprehensive guidance on post-incident requirements, particularly for incidents involving lethal force. • The SJPD should use Section L 2610 (Providing First Aid) of the Duty Manual as a template for detailing the medical steps officers are required to take after using force. • The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation. • The SJPD should look further into racial disparities found in the quantitative analyses, identify potential reasons for the differences, and —where reasons are identified —take remedial steps.
vi
Contents
Introduction ............................................................................................................................................................ 1 Goals and objectives ....................................................................................................................................................... 4 Overview of the report .................................................................................................................................................. 4 Section 1: Methodology and Approach .......................................................................................................... 6 Document review............................................................................................................................................................. 6 Interviews and officer focus groups ........................................................................................................................ 6 Community listening sessions/community interviews ................................................................................... 6 Quantitative analyses ..................................................................................................................................................... 7 Section 2: Review of Use of Force Policy and Procedures ..................................................................... 10 Findings and recommendations .............................................................................................................................. 10 General findings about SJPD use of force policies .................................................................................... 10 Use of force categories ......................................................................................................................................... 18 Electronic control weapons ............................................................................................................................... 20 Use of deadly force ................................................................................................................................................ 21 Mass demonstrations ........................................................................................................................................... 26 Providing first aid .................................................................................................................................................. 29 Use of force report ................................................................................................................................................. 30 Section 3: Use of Force Events ......................................................................................................................... 34 Racial disparity in use of force events .................................................................................................................. 41 Findings and recommendations .............................................................................................................................. 45 Section 4: The Impact of COVID-19 and Social Justice Movements on Police Reform ................ 48 Findings and recommendations .............................................................................................................................. 51 Section 5: Use of Force and Injuries by Race and Ethnicity .................................................................. 52 Findings and recommendations .............................................................................................................................. 61 Conclusion .............................................................................................................................................................. 62 Appendix A: Data Reviewed by CNA Team ................................................................................................. 64 Appendix B: Table of Findings and Recommendations ......................................................................... 65 Appendix C: Interrupted Time Series Regression Tables ..................................................................... 71 Appendix D: Coding for Use of Force Severity ........................................................................................... 75 Appendix E: Group Balance and Regression Tables................................................................................ 76 Black compared with white tables ......................................................................................................................... 76 Asian compared with white tables ......................................................................................................................... 84 Hispanic compared with white tables ................................................................................................................... 92 vii
This page intentionally left blank.
1 Introduction
The city of San José, California , with a population of approximately 1.05 million, is the tenth-largest city in the United States and the third -largest city in California . The city covers approximately 181 square miles at the southern end of the San Francisco Bay and has a relatively low population density.
The city’s population has been growing over the past decade , with an 11 percent increase (roughly 100,000 people).1 The San José Police Department (SJPD) has also changed with the city’s demographic and economic growth as have the numbe r of index crimes committed annually .2 The department’s operating expenditures ha ve grown from roughly $290 million in 2010 –2011 to an operating budget of $47 2 million for 2020 –2021 (a 63 percent increase).3 During 2020–2021, the SJPD was budgeted for 1,157 sworn positions and 558 civilian positions; however, high vacancy rates among street -ready officers have consistently presented a challenge , with only 83 percent of authorized sworn positions filled with full-duty, street-ready officers as of June 2020.4 Community members’ attitudes toward the SJPD have remained persistently low. In the city’s 2020 yearly survey, only 36 percent of city residents rated police services as “excellent” or “good ,” even though 48 percent and 36 percent of respondents reported t hat they feel “very” or “somewhat” safe from violent and property crimes, respectively. The challenges associated with community perspectives of the SJPD and legitimacy in its work likely originate from high -profile use of force instances. The SJPD has had 25 deaths by police officers since January 2015, 6 of which involved unarmed community members.5 Community members we spoke with often expressed frustration at how SJPD officers treat community members during police-community interactions.
The COVID -19 pandemic has only exacerbated the challenges facing San José police -community relations. Beginning in March 2020, COVID-19 caused the longest disruption to city services and the longest active emergency response in the city’s recent history. 6 The pandemic caused economic activity to slow as the city’s unemployment rate increased to 13.8 percent in April 2020 from a pre - 1 Office of the City Auditor. (2020). Annual Report on City Services 2019-20. San José, CA: Office of the City Auditor. https://www.sanjoseca.gov/home/showpublisheddocument/67957/637467496715000000 2 Index crimes include homicide, rape, robbery, aggravated assault, burglary, larceny, motor vehicle theft, and arson. 3 Office of the City Auditor. (2020). Annual Report on City Services 2019-20. San José, CA: Office of the City Auditor. https://www.sanjoseca.gov/home/showpublisheddocument/67957/637467496715000000 Office of the City Auditor. (2021). Police staffing, expenditures, and workload: Staffing reductions have impacted response times and led to high overtime costs. San José, CA: Office of the City Auditor. https://www.sanjoseca.gov/home/showpublisheddocument?id=70064 4 Office of the City Auditor. (2020). Annual Report on City Services 2019-20. San José, CA: Office of the City Auditor. https://www.sanjoseca.gov/home/showpublisheddocument/67957/637467496715000000 5 As determined from https://fatalencounters.org/ data downloaded on 12/29/2021. 6 Office of the City Auditor. (2020). Annual Report on City Services 2019-20. San José, CA: Office of the City Auditor. https://www.sanjoseca.gov/home/showpublisheddocument/67957/637467496715000000 2 pandemic monthly average of 2.6 percent. 7 The pandemic caused a hunger crisis ; Bay Area food banks reported a surge in requests as people lost their jobs. In addition, the negative effects of the pandemic were disproportionate across the community, with Hispanic people accounting for nearly 60 percent of COVID-19 cases in the county but only 26 percent of county residents.
Coinciding with th e COVID-19 pandemic, communities and police departments across the country , including in San José, began to acknowledge and respond to the differential treatment experienced by different racial and ethnic groups . Specific to policing, national and local research has identified multiple indications of disparate treatment and outcomes by racial characteristics .8 For example, predominantly Black neighborhoods in New York City have been found to have higher rates of police- initiated contacts regardless of actual local crime rates.9 Furthermore, studies in North Carolina and Kansas City have found that Black adults are more likely than white adults to report having been unfairly stopped by police because of their race or ethnicity and that officers search Black and Hispanic drivers more often than they do white drivers, even though there are similar levels of illegal drug possession among these groups .10 Disparate outcomes of police interactions by race are also well documented. Although use s of force account for roughly two percent of all police -community interactions nationwide, numerous studies have found that police officers are more likely to use force and excessive force against people of color . For example, Black people are almost three times more likely to die than white people are when police result to using force during interactions.11 This differential treatment came to a head following the murder of George Floyd by Derek Chauvin on May 25, 2020, causing nationwide movements with activists and community members calling for police reform, defunding, and abolition , including in San José. As a result, the mayor and city councilmembers proposed a variety of police reform directives, including a use of force review consistent with the Reimagining Police Pledge , which aimed to review force policies, engage the 7 Office of the City Auditor. (2020). Annual Report on City Services 2019-20. San José, CA: Office of the City Auditor. https://www.sanjoseca.gov/home/showpublisheddocument/67957/637467496715000000 8 Davis, E., Whyde, A., & Langton, L. (2018). Contacts between Police and the Public, 2015. Washington, DC: US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.; Haldipur, J. (2020). No place on the corner: The costs of aggressive policing. New York: NYU Press.; Hetey, R. C., Monin, B., Maitreyi, A., & Eberhardt. J. L. (2016). Data for change: A statistical analysis of police stops, searches, handcuffings, and arrests in Oakland, Calif., 2013–2014. Stanford, CA: Stanford University, Social Psychological Answers to Real-World Questions. 9 Fagan, J., Geller, A., Davies, G., & West, V. (2010). Street Stops and Broken Windows Revisited. In Race, Ethnicity, and Policing: New and Essential Readings, (eds) Rice, S.K. & White, M.D. 309–48. New York: NYU Press. 10 Baumgartner, F. R., Epp, D. A., & Shoub, K. (2018). Suspect citizens: What 20 million traffic stops tell us about policing and race. Cambridge, GBR: Cambridge University Press.; Epp, C. R., Maynard-Moody, S., & Haider-Markel, D. (2014). Pulled over: How police stops define race and citizenship. Chicago: University of Chicago Press. 11 Davis et al. (2018); Fryer R. G. (2016). An empirical analysis of racial differences in police use of force. Cambridge, MA:
National Bureau of Economic Research.; Goff, P. A., Lloyd, T., Geller, A,, Raphael, S. & Glaser, J. (2016). The science of justice:
Race, arrests, and police use of force. Los Angeles: Center for Policing Equity.; Harrell, E., & David, E. (2020). Contact between police and the public, 2018. Washington, DC: US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.; DeGue, S., Fowler, K. A., & Calkins, C. (2016). Deaths due to use of lethal force by law enforcement:
Findings from the National Violent Death Reporting System, 17 U.S. States, 2009–2012. American Journal of Preventive Medicine, 51(5), S173–S187.
3 community in that review, report revie w findings to the community and seek feedback, and reform use of force policies. 12 A detailed Police Reforms Work Plan was created in San José and its implementation is underway.13
On December 2, 2020, the city released its consulting services request for proposals for a Review and Report on SJPD's Use of Force and 21st Century Policing Policies . Six proposals were received and competitively reviewed and evaluated on general requirements, experience and qualifica tions, and technical approach, with additional points awarded for local and small businesses.14 After review, the city recommended that the CNA be awarded the contract, which it was on February 12, 2021. The process to select CNA was completed before the commencement of the reimagining process.
CNA has more than 77 years of experience working with government clients and roughly 17 years of experience with public clients in preparing research findings, developing policy guidance, and identifying and incorpo rating best practices. CNA has supported 450 law enforcement agencies nationwide in implementing 21st Century Policing best practices and has worked with 39 cities and counties in California through many law enforcement and emergency management projects an d programs. CNA uses a multifaceted approach that includes data-driven methods and community engagement in its police department reviews. As part of that work, CNA was originally expected to work collaboratively with a separately selected community engagement and facilitation consultant to obtain community feedback. However, the consultant selected on February 25, 2021, dropped out of the process soon after and the initial community advisory group was disbanded in June 2021. A new Reimagining Public Safety Community Advisory Committee launched in August 2021. Despite the shift in the reimagining process, CNA continued work under its contract and conducted multiple community listening sessions to incorporate community feedback into its final deliverables, including this report on SJPD use of force.
The renewed national conversations about systemic bias and racism in justice systems have led the city of San José to initiate an examination of SJPD’s policies, procedures, and practices associated with use of force. This report details the methodology and findings of CNA’s examination into SJPD uses of force that occurred f rom February 17, 2017, to February 27, 2021 (roughly four years). We found that 936 unique SJPD officers completed a total of 4,817 unique use of force reports for 2,352 use of force events. In roughly the same time period, there were approximately 327,000 face-to-face police- community interactions resulting from a call for service, leading to a total of 0.72 percent of police interactions resulting in a use of force report being officially submitted (or, stated another way, one use of force event for every 119.9 interactions) . The SJPD requires all officers involved in a use of force event to fill out a report for each community member involved. A community member in this report is defined as an individual who had an interaction with a SJPD officer or who resides or works
12 Obama Foundation. (2021). Reimagine Policing. https://www.obama.org/policing-pledge/ 13 Maguire, J. A. (2020). Memorandum: Police reforms work plan. San José, CA: City of San José. https://www.sanjoseca.gov/home/showpublisheddocument/61498/637302498292170000 14 City of San José. (2020). Review and Report on SJPD’s Use of Force and 21st Century Policing Policies. San José, CA: City of San José. https://portal.biddingo.com/landingpage/sanjose/bid/1/41213543/34587128/verification
4 in San José. A use of force event can and should have multiple reports if there are two or more officers or community member s involved. Our analyses aggregated the report data to the community - member-per-event unit, which totaled 2,743 community members across the 2,352 use of force events. The following report details our data analyses of these use of force events, alongside o ther assessments that reviewed use of force policies and procedures, conducted stakeholder interviews, and documented community concerns during listening sessions.
Goals and objectives
The scope of work under the full contract encompassed two topics , the f irst of which is an examination of the SJPD’s policies and protocols on which to base recommendations on how to align San José with the best practices outlined in the Final Report of the President’s Task Force on 21st Century Policing (published in 2015).15 The second was an assessment of SJPD use of force that would review and address policies, procedures, training, tools, reporting, accountability , and investigative processes. The first area of focus was completed and is detailed in our 21st Century Policing Report.16 The work associated with th e second topic i s the focus of this report and was prepared in coordination with the 21st Century Policing Report by the same assessment team. This assessment was designed to review and provide recommendations to improve the following:
• SJPD’s policies and procedures that govern the use of force. • SJPD’s directives on how use of force incidents are reported and documented, including the collection of such incidents for data and trend analysis. • SJPD’s directives describing the process, roles , and responsibilities for the review of use of force incidents.
• SJPD’s training on use of force.
• SJPD’s equipment, tools, and tactics.
• SJPD use of force report data and, to the extent necessary, records of specific encounters involving use of force.
Overview of the report
Following the introduction, this report contains five sections. Section 1, the study’s methodology and approach, details how we assessed SJPD use of force through thorough document review, interviews, listening sessions, and data analyses. Section 2 discusses our assessment of the policies, procedures,
15 President’s Task Force on 21st Century Policing. (2015). Final Report of the President’s Task Force on 21st Century Policing. Washington, DC: Office of Community Oriented Policing Services. https://cops.usdoj.gov/pdf/taskforce/taskforce_finalreport.pdf 16 Christoff, T., Dockstader, J., Jenkins, M., Stephens, C. (2021). Draft Report: 21st Century Policing Assessment of the San José Police Department. Arlington, VA: CNA Corporation.
5 training, and events for best practices related to both deadly and non -deadly use of force. Section 3 provides descriptive analyses of use of force by SJPD officers in the city, with particular emphasis on identifying evidence of racial or ethnic disparities in the application of use of force. Section 4 discusses analyses of how the COVID-19 pandemic and coinciding social justice movements for police reform affected calls for service, arrests, and use of force. Finally, Section 5 covers analyses that examined how outcomes associated with use of force events differ across racial groupings. At the end of each section, we provide findings and recommendations . The report closes with a conclusion section.
The report also includes four appendices. Appendix A provides a complete list of d ata the assessment team reviewed. Appendix B includes a table of all findings and recommendations.
Appendix C presents the regression tables associated with analyses on the impact of COVID-19 and social justice movements discussed in Section 4. Appendix D details how the assessment team coded use of force severity levels for the analyses. Finally, Appendix E provides the descriptive statistics, group balance, and regression tables on how outcomes associated with use of force events differ across racial groupings, as discussed in Section 5.
6 Section 1: Methodology and Approach
The CNA assessment team based our force assessment on the following guiding principles:
1. Evidence-based technical assistance with emphasis on academic research , documented lessons learned, and best practices in the field.
2. Multimethod assessment design, including interviews, listening sessions, policy and document review, and data analysis.
3. A commitment to conducting comprehensive reviews and ap plying best practices in police settings.
Our assessment of SJPD use of force included a review of document review, interviews, community listening sessions, and quantitative analysis.
Document review
The SJPD and the city of San José shared documents with the CNA assessment team to provide a better understanding of operational procedures and practices regarding use of force . These documents included the SJPD Duty Manual, bulletins, local ordinances, training lesson plans , training materials, and use of force case files. The CNA team reviewed each of these sources as part of our assessment.
Interviews and officer focus groups
CNA conducted over 20 semi-structured interviews with a diverse set of SJPD personnel (including representatives from th e chief’s office, training division , and research and development, among others) to develop an understanding of formal policies and procedures, as well as to gain insights into the agency culture and community relations. We also conducted 12 virtual focus groups with SJPD patrol officers, sergeants, and lieutenants. Officers discussed the SJPD’s community engagement activities, officer wellness and safety, problem -oriented policing approach to solving crime, and culture within the d epartment. During the pat rol officer focus group, officers discussed organizational culture and officer morale. The interviews provided qualitative data for our assessment of policing practices, culture, leadership, and use of force approaches, expanding our understanding of the agency’s culture and unique dynamics.
Community l istening session s/community interviews
The assessment team reached out to over 75 community stakeholders and organizations to help promote and invite communit y members to listening sessions. The organizations were identified either through CNA’s own research, by direction of community stakeholders who participated in interviews, or from SJPD liaisons who mentioned specific community organizations they worked
7 with during interviews and/or focus groups . The SJPD did not directly recommend or introduce the assessment team to any specific community organization. CNA hosted two virtual listening sessions with residents in San José on July 27, 2021, and August 25, 2021, with over 50 attendees between the two listening sessions. T he information gleaned at these sessions helped the assessment team understand community perspectives regarding the SJPD and its applications of use of force. Based on their direct experiences interacting with officers and the agency, community members were able to provide insight into how the application of SJPD policies may inadvertently result in disparate outcomes. The listening sessions also provided insight on community policing topics , engagement, and on interactions with youth.
Additionally, the assessment team interviewed nearly 40 community stakeholders representing the diverse range of community members within San José , including representatives from Black, Hispanic, Asian, mental hea lth, youth, and religious organizations. As stakeholders are the most direct source of information about impressions of SJPD community outreach and engagement efforts, we found conversations with these individuals to be vital to our assessment.
Quantitati ve analyses
Our data analysis primarily relied on the SJPD’s use of force report data; although we also used SJPD’s open-source calls for service data, data on arrests, and other supporting information (such as Census data) in support of our analyses. The data we examined covered all use of force reports from February 17, 2017, to February 27, 2021. During that roughly four-year period, 936 unique SJPD officers were involved in 2,352 use of force events, completing a total of 4,817 unique use of force r eports. In roughly the same period, there were approximately 327,000 face -to-face police -community interactions resulting from a call for service, leading to a total of 0.72 percent of police interactions resulting in a use of force report being officially submitted (or, stated another way, one use of force event for every 119.9 interactions) . The SJPD requires all officers involved in a use of force event to fill out a report for each community member involved. As such, an event can and should have multiple reports if there are two or more officers or community members involved. Our analyses aggregated the report data to the community-member-per-event unit, which totaled 2,743 community members across the 2,352 use of force events. Some community members had multiple instances of use of force in the data, meaning that the total number of unique community members in the database was less than 2,743.
It is worth noting and emphasizing here that the data are from the use of force reports that officers fill out after a use of force event. Furthermore, the data used in the following analyses were not verified using body-worn camera ( BWC) footage. Nonethel ess, in most cases, a use of force event included corroborating information from multiple use of force reports from different officers (e.g., consistent field entries between officers) . We also note that BWCs—which SJPD deployed in
8 November 2016—have also been found to improve the documentation and reporting of events by emergency personnel in other agencies.17
Identifying community members in the data and ensuring that the appropriate amount were counted was an incredibly complex task. The assessment team needed to reformat the data to the community- member-per-event level to best measure how many community members were involved in use of force events. In addition, early in the data cleaning phase, the team quickly identified multiple problems that would make this task difficult.
An initial problem was that the General Offense Report Numbers that identify specific use of force events were found to be the same across community members involved in a single event. For example, a single use of force event that i nvolved three community members would have the same report number, which made it difficult to identify the individual community members involved in the event.
A second problem was that use of force report filled out by SJPD officers include s five fields for the involved community member’s name. The first field uses a drop -down menu of a known list of entities to select from, the second field is meant for the community member’s last name, and the remaining three fields are meant for any other “given names” that could be entered in any order. The problem was that all officers involved in a use of force event are expected to fill out an associated use of force report; meaning that multiple officers may each create a report and enter community members’ name s in different ways. Furthermore, there was no identification number attached to individuals to determine if they were involved in multiple different use of force events within the examined time period.
A third problem was that there were 46 instances in which the General Offense Report Number was entered incorrectly across officer reports for the same use of force event, inflati ng the apparent number of unique use of force events. There were also 20 instances where officers created multiple use of force reports for the same community member within an event . In these instances, the assessment team relied on the first report.
Most of these errors were corrected through a combination of manual review and a statistical character matching procedure (i.e., MATCHIT command in Stata) to create a database with a unique identifier for each community member within an event . The assessment team used the MATCHIT command in Stata to create a similarity score across all pairs of cases using the report numbers a nd separate name fields available in the data. Once similarity scores were calculated, the assessment team manually reviewed 419 mismatched cases within report numbers that had similarity scores of
0.80 or lower. This process allowed us to identify instances where different community members were involved in the same use of force event and create unique identifiers for each individual. When
17 Dawes, D., Heegard, W., Brave, M., Paetow, G., Weston, B., & Ho, J. (2015). Body-worn cameras improve law enforcement officer report writing accuracy. Journal of Law Enforcement, 4(6), 1-21.
Ho, J.D., Dawes, D. M., McKay, E. M., Taliercio, J.J., White, S. D., Woodbury, B.J., Sandefur, M.A., Miner, J.R. (2016). Effect of body-worn cameras on EMS documentation accuracy: A pilot study. Prehospital Emergency Care, 21(2), 263-271.
9 names were not available for matching, the assessment team grouped community members within a use of force event by demographic characteristics, when possible.
Although the above process identified unique individuals within a use of force event, we also wanted to ensure that we were not over counting the same individual within an event. For example, the procedure would have created separate unique identifying numbers in instances where two officers were involved in a use of force event against one community member , but each officer spelled the community member’s name slightly differently. Thus, we conducted a second round of the character matching procedure but instead reviewed cases with different unique identification numbers within an event but with similar names. We identified a total of 47 cases that were being treated as separate community members because of minor spelling differences across reports. We corrected these cases were to use the same identification number for each individual.
Once we created unique identification numbers for each community -member-per-event, we aggregated use o f force reports at the community member level (rather than the event or officer level). This process essentially reduced the 4,817 unique use of force reports to the 2,743 community member cases. We noted any recorded presence of a data field across the sometimes differing officer reports within an event. For example, if there were two officer reports for an event and one officer reported that the threat of a firearm was used (by an officer) but the other officer did not, we aggregated the cases to include the presence of the threat of a firearm by an officer. In fields that were categorical (e.g., the community member’s race), we relied on the response option with the most recordings across the officers’ reports . For example, if one officer said a community member was white, and two officers said the community member was Hispanic, we recorded the community member as Hispanic.
With a fully cleaned and aggregated dataset, we were able to report descriptive information about use of force events in San José and conduct analyses on the disparity of use of force across racial characteristics (Section 3), assess the impact of COVID -19 and social justice movements for police reform (Section 4), and conduct statistical comparison analyses on the amount and severity of force and injuries across racial characteristics (Section 5), all of which are detailed below.
10 Section 2: Review of Use of Force Policy and
Procedure s
The CNA assessment team worked with the SJPD and identified the documents relevant to use of force. These documents included policies, general orders and directives, standard operating procedures, specialized section manuals, bulletins, local ordinances, and training lesson plans. We separate the following findings and recommendations from those reviews by policy topic area, which include general policies on the use of force, policies on the different categories of use of force, the use of electronic control weapons, the use of deadly force, policies on behaviors during mass demonstrations, policies on providing first aid, and findings and recommendations associated with the use of force reporting form that officers are required to fill out.
Findings and recommendations
General findings about SJPD use of force policies
Finding 1: Segments of the San José community have diverse perspectives and experiences with use of force.
We discussed use of force by the SJPD d uring our interviews and focus groups with community members. A primary takeaway fr om these sessions was that the community has very diverse perspectives and experiences with the SJPD, including with use of force, which tracks with the assessment team’s quantitative analysis of use of force incidents. Black and Hispanic community members often shared their own use of force experiences and concerns (or those of their immediate social circles). Asian and Muslim community members frequently cited awareness of high -profile incidents but had not personally experienced use of force by SJPD offi cers. Acknowledging these experiences and concerns will be important as the SJPD works to build and rebuild relationships with specific communities within San José that have experienced harm by and mistrust of the department.
As the SJPD modifies its operations in accordance with this assessment, its own internal evaluations, and future independent reviews, we recommend such modifications reflect such diverse perspectives and experiences.
Recommendation 1: Future modifications to departmental operations should reflect the diversity of perspectives and experiences of the San José community.
11 Finding 2: The SJPD’s use of force tools and options that are listed in the Duty Manual are consistent with other agencies.
The SJPD Duty Manual identifies the primary force tools and options that SJPD members are authorized to use during encounters with community members. In reviewing the Duty Manual and the force tools identified, we find that they are consist ent with force tools and options found within other agencies. The SJPD Duty Manual includes discussion of the following force tools and options:
• Carotid restraints
• Chemical agents
• Chokeholds
• Electronic control weapons (ECWs)18
• Electronic restraint transportation belts
• Firearms
• Impact weapons
o Expandable batons
o Kendo sticks
o Side handle batons
o Straight batons
o Yawara sticks
• Physical contact/body weapons
o Control holds
o Takedowns
• Police service dogs
• Police vehicles
• Projectile impact weapons
o Stun bag shotguns
o 40mm less-lethal munition
• Violent patient transports
• Voice commands
• Wrap restraints
Although we note concerns with SJPD policy for some of these tools and options below, the overall force options and tools used by the SJPD are consistent with conventional police standards.
Recommendation 2: The SJPD should maintain its current practice for this finding.
Finding 3: The SJPD Duty Manual and training contain consistent elements across all use of force tools and options.
For nearly all force tools and options, the SJPD Duty Manual includes guidelines for officers regarding the criteria for using the tools and options, prohibited uses, providing first aid after using the tool or option, supervisor responsibilities , categorization of review, and review procedures. In some instances, additional guidance is given about using particular force tools for crowd control. For each of these elements, the Duty Manual gives clear guidance. Furthermore, many of these elements a re
18 SJPD's choice of ECW is Axon’s Taser. The company Axon was formerly known as TASER prior to April 2017.
12 reinforced through comprehensive training for force tools and options. We reviewed the 2021 Defensive Tactics training , which goes though the tools and options, identifies their abilities and limitations, sets guidelines for use, and describes first aid steps officers should take after using the force. Overall, we find the SJPD Duty Manual and related training maintain consistent elements across force tools and options.
Recommendation 3: The SJPD should maintain its current practice for this finding.
Finding 4: The SJPD Duty Manual does not define levels of resistance and does not consistently indicate which level of resistance would justify various force options.
In reviewing the Duty Manual, we found no discussion of the types of resistance that officers may encounter and no definition of the types of resistance. Furthermore, while the Duty Manual is explicit about the levels of resistance needed for some force options, this is not consistent. For instance, the Duty Manual states in Section L 2609 (Use of Chemical Agents) that “the use of chemical agents must be justified by a suspect’s active resistance” (emphasis added). However, for t ools such as impact weapons, the Duty Manual states that they will be used only “to incapacitate suspects armed with a weapon likely to cause serious bodily injury or death” or “in situations wh ere its use is likely to prevent any person from being seriously injured.” Both of these situations appear to describe active aggression as the level of resistance.19 It may benefit the SJPD to explicitly state this (with, of course, a sufficient definition of the term).
The Duty Manual more often just includes the broad term “resistance” and states that force must be objectively reasonable to overcome that resistance. 20 This is a Graham v. Connor 21 standard, and therefore the SJPD meets the constitutional minimum. However, the Graham standard is a legal threshold in the context of determining Fourth Amendment violations, and nothing prohibits SJPD policy from being more explicit about when force options are acceptable and when they are n ot. In the same way that SJPD policy limits chemical agents to situations involving active resistance, the department can set the minimum thresholds for other force options. We note this differs from a force continuum in the sense that departments with a c ontinuum indicate officers should use a specific force option when faced with a certain resistance level . Instead, our position is that SJPD should clearly identify the force options officer should not use unless faced with certain resistance levels.22 A lack of definition for “resistance” creates internal inconsistency within the policy. For instance, both the Carotid Restraint and Chokehold sections note that the tactics are “prohibited as an authorized control technique to overcome resistance ,” though later say, “After resistance is overcome with the
19 Levels of resistance typically fall into three levels: passive resistance (e.g., going limp), active resistance (e.g., attempting to escape the officer’s control), and active aggression (e.g., attempting to fight the officer). 20 For lethal force events, the SJPD also requires the force to be “necessary,” as required by Assembly Bill 392. 21 Graham v. Connor is considered the guiding Supreme Court decision on the constitutionality of use of force. The decision holds that the reasonableness of use of force should be determined under the totality of circumstances as judged by a “reasonable officer” without the benefit of 20/20 hindsight. 22 See, for instance, Portland Police Bureau Directive 1010.00.
13 [carotid restraint/chokehold]…” The Duty Manual should make explicit that both tactics require the subject to pose a threat of death or great bodily harm. We recommend that SJPD state the minimum resistance level needed for each force option while also emphasizing that those force options may still be unreasonable based on the totality of the circumstances.
Recommendation 4.1: The SJPD should better define levels of resistance.
Recommendation 4.2: The SJPD should state the minimum resistance level needed for each use of force option.
Finding 5: The SJPD does not have a force review board.
Currently, the SJPD has an Officer -Involved Incident (OII) Training Review Panel that is designed to review critical incidents for policy, training, equipment, and personnel implications. Although we discuss the operation of the OII Training Review Panel in our 21 st Century Policing Report (see Recommendation 2.3 in that report), we note that as of the writing of this report, there is no broader force review board to identify department-wide trends in use of force, encompassing lower-level but more common uses of force.23 Alternatively, there appears to be no one responsible for auditing uses of force and identifying trends in a way that would allow for policy, training, equipment, and personnel enhancements.
The SJPD currently contracts with a third -party research organization to produce yearly reports on quantitative assessment s that discuss longitudinal and comparative frequencies of force. 24 In addition, supervisors are required to review every use of force, evaluating an ind ividual officer’s actions leading up to and including the application of force. However, no mechanism exists for evaluating more global trends, including officer positioning, officer communication skills, or actions after using force (e.g., providing medic al aid). Furthermore, as noted above, there is no audit mechanism to ensure that supervisory reviews are comprehensively addressing these elements.
That is not to say that such assessments are not conducted informally. For instance, the SJPD verbally informed the assessment team of a training developed in house that addresses officers firing their weapons at moving vehicles. This responsive training was born after the SJPD identified a concern with officers engaging in this action (for further information, see our 21st Century Policing Report).
Although we would need to further review the training to comment on the content, the process used by the SJPD in this situation is wholly in line with what we would expect from a more formal force review board. Howeve r, without a formalized process, it is possible that other similar trends have gone unnoticed and unaddressed.
Several agencies incorporate a force auditor or a force review board. For instance, Portland, Oregon, has a force inspector, a sworn position that is supported by non-sworn analysts. The force inspector’s team evaluates force reports to ensure that the reports are comprehensive and accurate. In addition,
23 The SJPD advised the assessment team that a Force Review Board is in the process of being developed; however, at the time of this report, the board had not been implemented. 24 See yearly Police Strategies, LLC reports at https://www.sjpd.org/records/crime-stats-maps/force-analysis-data.
14 the force inspector reviews force events to identify issues that do not rise to the level of an administrative investigation but may require a debrief with the officer to offer instruction on how a different approach might have avoided the use of force. Furthermore, the agency has a Police Review Board, a primarily disciplinary board though one which also evaluates lethal force events to identify policy, training, equipment, and personnel concerns. The board contains sworn members as well as community representatives.
The SJPD may also choose to use a broader board to review all uses of force. For in stance, New Orleans uses a force review board made up of NOPD Deputy Superintendents to look at lethal force events as well as other higher level uses of force. The Board assesses “the actions of each participant in the use of force event, often starting with the actions of the 9-1-1 center call taker and ending with the first NOPD supervisor who arrives on the scene and establishes incident control.”25 As stated above, the SJPD already includes elements of a broader, formalized review in a quantitative sense. Furthermore, we have seen the SJPD demonstrate an ability to identify trends indicating a departmental need and implement responsive training. Therefore, we believe the SJPD should take a more systematic and purposeful approach as this is what appears to be primarily missing. We also recommend the SJPD include community representatives as part of these efforts where possible. For instance, community representatives could be used as part of non -disciplinary reviews of force, providing an opportunity for civilian input into policy, training, tactics, and equipment considerations. Since the SJPD is unable to incorporate community oversight into making disciplinary decisions, they could bifurcate the process to allow community input into the non -disciplinary elements. Alternatively, this process could be facilitated through the Independent Police Auditor (IPA), allowing the SJPD to provide a fact -based briefing for the community. IPA could then issue recommendations based on the feedback received.
Recommendation 5.1: The SJPD should create a force review board or unit to identify policy, training, equipment, and personnel implications.
Recommendation 5.2: The SJPD should include community representatives as part of its efforts.
Finding 6: The SJPD does not have a policy requiring officers to attempt de-escalation.
Overall, the SJPD Duty Manu al needs a stronger emphasis on de -escalation. For instance, Section L 2602.5 (Tactical Conduct) states that the proper use of cover, concealment, distance, and time “may help a Department member avoid and/or minimize the use of physical force.” Although we agree with this statement, it appears to be the sole mention of an affirmative duty to attempt de -escalation methods prior to using force in the Duty Manual.
25 Office of the Consent Decree Monitor (2017). Special report of the consent decree monitor for the New Orleans Police Department consent decree on NOPD uses of force. Washington, DC: Sheppard Mullin Richter & Hampton, LLP. https://www.nola.gov/getattachment/NOPD/NOPD-Consent-Decree/12-1924-517-Monitor-s-Special-Report-on-the- NOPD-s-Uses-of-Force-(002).pdf/
15 Additionally, the Duty Manual does not provide officers with guidance on the range of de -escalation tactics that they are expected to know and use. For instance, the Duty Manual does not discuss de - escalation considerations such as maintaining one-on-one communication (as opposed to crosstalk), potential alternative reasons that a subject is not com plying (e.g., mental health crisis or language barrier), or disengaging from the subject. We point to the Baltimore Police Department’s use of force policy in this respect.26
The Duty Manual also misidentifies the actual application of force as a de -escalation tool. For instance, section L 2609 (Use of Chemical Agents ) states that “when properly used, chemical agents are a valuable de -escalation tool,” and that the use of chemical agents “can reduce or eliminate the necessity to use other force options to gain compliance.” While lower levels of force certainly may be used to avoid a higher use of force, these actions do not conform to the spirit of de -escalation, since de-escalation is designed to be a force avoidance tool.27 By including a use o f force as a force avoidance tool, the SJPD policy contradicts itself and may cause confusion among officers. In addition, the language of this section may hold the potential for any improper use of chemical agents to be mitigated as necessary to avoid higher levels of force (though we did not see any evidence that this happening in the SJPD).
Conversely, SJPD training does comprehensively identify and discuss a wide range of de -escalation skills and tactics. In 2020, the SJPD gave officers a six -hour train ing developed based on a Police Executive Research Forum ( PERF) training that included lecture s as well as live scenarios emphasizing communication skills. Additionally, the training covered non -communication-based tools, including gathering information prior to engaging with the subject when possible. We reviewed the course outline and instructor’s notes for this class and find it to be consistent with de -escalation trainings seen in other agencies. Where possible, we recommend the SJPD incorporate critica l elements of the training into departmental expectations that are memorialized into the Duty Manual.
Finally, we note that community members consistently discussed SJPD officers’ reported lack of de - escalation tactics during our interviews with community members. Many shared examples of specific use of force incidents in which unarmed or noncombative community members experienced escalating use of force without appar ent justification. Community members fe lt that officers are focused primarily on their own safety and “getting home safe” and are therefore inclined to escalate force more rapidly as a self -protective measure. Others also discussed other communication -based issues that could lead to a use of force event. For instance, some noted that if there is a language barrier, incidents often escalate because the involved community member does not understand the officers’ commands. We therefore recommend the SJPD engag e the community in defining de - escalation and providing concrete tools for officers in the Duty Manual.
26 Baltimore Police Department Policy 1115 Use of Force 27 See also: International Association of Chiefs of Police (IACP) (2020). De-escalation: Guidelines for how to begin evaluating your agency’s de-escalation practices. Alexandria, VA: IACP. https://www.theiacp.org/sites/default/files/Research%20Center/Combined%20v3.pdf
16 Recommendation 6.1: The SJPD should elevate and emphasize affirmative duty to attempt de-escalation during encounters when time and
circumstances permit.
Recommendation 6.2: The SJPD should better define the concept of de-escalation.
Recommendation 6.3: The SJPD should revise the Duty Manual to remove the reference to use of force being a de-escalation tool.
Recommendation 6.4: The SJPD should engage the community when defining de- escalation and providing concrete tools in the Duty Manual.
Finding 7: The SJPD has levels of force but does not elevate force levels if the event involves vulnerable populations.
As discussed in the President’s Task Force on 21 st Century Policing report, the use of force against persons in vulnerable populations “can undermine public trust and should be used as a last resort” (p.15). The Task Force defined "vulnerable populations" as encompassing “children, elderly persons, pregnant women, people with physical and mental disabilities, limited English proficiency, and others” (p.15). However, the only reference in the Duty Manual in the context of force to vulnerable populations is found in Section L 2614 ( Use of Electronic Control Weapons and Reporting Requirements), though the language in this section is vague (see Finding 13 ), and there are no other concrete directives that speak to the issue as it relates to any other for ce tool or option. The SJPD should revise the Duty Manual to include the language of the Task Force’s report and apply the language to all force tools and options.28
Additionally, the categorization of force in Section L 2605 ( Supervisor & Command Officer Responsibility) does elevate force based on other qualifiers. Some qualifiers pertain to how the force was used. For instance, although the use of an impact weapon typically is categorized as a Category II Use of Force, an impact weapon to the head is elevated to a Category III Use of Force .29 Other qualifiers pertain to who uses the force. For instance, two or more officers deploying less-than-lethal force on one subject is similarly elevated from a Category II to a Category III. Requiring enhanced review of use of force incidents against vulnerable populations would be consistent with the SJPD’s approach with other qualifiers.
Recommendation 7.1: The SJPD should revise its Duty Manual to include the language of the President’s Task Force on 21st Century Policing when describing vulnerable populations in all uses of force.
Recommendation 7.2: The SJPD should require enhanced review of use of force incidents against vulnerable populations.
28 For a broader discussion of vulnerable populations, see also: IACP. (2018). Practices in modern policing: Policing in vulnerable populations. Washington, DC: Office of Community Oriented Policing Services. https://www.theiacp.org/sites/default/files/2018-11/IACP_PMP_VulnerablePops.pdf 29 See Finding 15 as related to the categorization of impact weapons to the head.
17 Finding 8: The SJPD Duty Manual does not require each application of force to be evaluated independently.
After a use of force event, the SJPD Duty Manual describes the process for supervisors to respond to the scene, collect evidence, approve reports, and forward the in formation to Internal Affairs. Thus, the investigative steps necessary to evaluate the use of force event are laid out comprehensively in the Duty Manual. However, constitutional force is not assessed at the event level. Each individual application of force needs to be evaluated independently under the totality of the circumstances.
Despite this need, the Duty Manual contains no explicit discussion that each application of force needs to be independently evaluated for all force tools and options. 30 While SJPD personnel inform us that this occurs in practice, w e recommend SJPD practice in this regard be memorialized in the Duty Manual.
Recommendation 8: The SJPD should revise the Duty Manual to require that each application of force be evaluated under the totality of the circumstances independent of other force applications.
Finding 9: The SJPD Duty Manual unnecessarily includes a section on excited delirium.
Section L 2614 (Use of Electronic Control Weapons and Reporting Requirements) of the Duty Manual includes a paragraph citing concern about excited delirium, describing it as “highly agitated individuals whose condition may put those individuals at heightened risk of serious injury or death.” The paragraph goes on to say that because of the individual’s state, there are times when the “use of [tasers] may, in some instances, be the preferred method of quickly subduing an agitated and aggressive individual and minimize the subject’s physical exertion.” We recommend the SJPD remove this paragraph from the Duty Manual as excited delirium is a disputed diagnosis 31 and the inclusion of this paragraph may cause officers to feel they are justified in using a taser on any “highly agitated individual.”
Recommendation 9: The SJPD should remove the paragraph on excited delirium from the Duty Manual.
30 Duty Manual section L 2614 (Use of Electronic Control Weapons and Reporting Requirements) discusses independent evaluation in the context of ECWs though the concept should be more broadly discussed in the context of all force tools and options.
31 See, for instance, American Medical Association’s June 2021 policy opposing “excited delirium” diagnosis.
18 Use of force categories
Finding 10: The SJPD Duty Manual does not provide sufficient clarity in the definition of force and bases force on injury or complaint of pain, rather than on the physical act that caused it.
Section L 2644 (Definition of Reportable Force ) of the Duty Manual defines a reportable use of force as follows:
Any incident in which officers, either on or off duty, exercise their police powers and use deadly force or any force option including physical force in conformance with L 2603, Force Options Policy.
However, this statement does not actually define “force” and instead states that reportable use of force is any incident in which officers use the force they are authorized to use . The Duty Manual should provide an explicit definition of force based on physical coercion of the subject against resistance.32
Presently, the SJPD appears to define force as a function of outcome rather than officer actions. Within Section L 2644 (Definition of Reportable Force) is text outlining exceptions to reportable use of force, including: (1) “firm grip control ,” which does not result in an injury or compl aint of pain, and (2) “force reasonable to overcome resistance due to physical disability or intoxication ,” which does not result in an injury or complaint of pain. Furthermore, Section L 2605.5 (Command Officer’s Responsibility by Use of Force Category) defines Category I Use of Force as any use of force “that causes minor injury or a complaint of pain.”
By focusing on the outcome to define force rather than the act, the Duty Manu al incorporates an unnecessary (and potentially subjective) qualifier, namely whether the subject was injured as a result of the officer’s actions. Furthermore, this puts the officer into the position of determining causation, which can be problematic. While this may represent the department’s position that force containing injury or complaint of pain warrants a heightened review (i.e., supervisor review of force), the more critical element is the proclivity of officers to engage in physical coercion. For instance, the intent of officers who deploy a n ECW or a 40mm less -lethal round but miss is the same intent as those who make contact. Whereas the act of using lethal force is investigated in the same manner regardless of whether the shot hit its intended target, so too should lower levels of force be investigated to better measure officer proclivities (as well as measure the effectiveness of the training they’ve received). As with the above recommendations, doing so would require a definition of use of force that is action - oriented rather than outcome-oriented.
Our conversations with community members reinforce this position, as some of those we spoke with were worried that SJPD members were too eager to engage physically, regardless of whether a
32 See, for example, IACP. (2020). National consensus policy and discussion paper on use of force. Alexandria, VA: IACP. https://www.theiacp.org/sites/default/files/2020-
07/National_Consensus_Policy_On_Use_Of_Force%2007102020%20v3.pdf
19 community member felt pain. In addition, recipients of officers’ force may not al ways be aware of injury at the time of the force, particularly those who are under the influences of alcohol or narcotics.
If the recognition of pain comes at a later time, SJPD will have missed the opportunity to conduct an immediate and comprehensive for ce review (including supervisors responding to the scene and interviewing witnesses).
Finally, we recognize that nearly all use of force tactics identified by SJPD are intended to (and likely will) lead to some type of injury or pain that would characteriz e it as a “reportable use of force.” Therefore, we do not believe that defining use of force as action-based rather than outcome-based is likely to significantly increase the number of force reports written by officers. However, while use of force data may not ultimately vary much by defining force as action -based, internal processes ( e.g., an early intervention system) may benefit if there are potential officer trends in low -level physical coercion.
Recommendation 10: The SJPD should adopt a “physical coercion against resistance” definition of force and apply it throughout the Duty Manual.
Finding 11: The SJPD’s categorization of force is not comprehensive.
As outlined in Section L 2605.5 ( Command Officer’s Responsibility by Use of Force Category ) in the Duty Manual, the SJPD uses a four -tiered categorization system for use of force, with the lowest category being Category I Use of Force, defined as “ any use of force not listed in Categories II, III , or IV, that causes a minor injury or a complaint of pain.” The highest category, Category IV Use of Force, is deadly force that “the user knows would pose a substantial risk of death or serious bodily injury.” Figure 1: SJPD use of force
20 However, the categories (as provided in the Duty Manual) are incomplete, as they do not include uses of force such as takedowns, strikes/kicks that are not to the head, and the categorization of improvised tools (e.g.., flashlights, vehicles). As these force options have the potential to cause more than “minor injury or a complaint of pain” (Category I) but may not require “hospital admission” (Category III), it is unclear where these force tools fit into the review process based on policy.
Recommendation 11: The SJPD should categorize all force tools and options in the categories identified in the SJPD Duty Manual.
Electronic control weapons
Finding 12: The SJPD Duty Manual covers many conventional sections related to ECWs.
In reviewing policy related to ECWs, we find that the Duty Manual covers many of the typical sections we would expect to find in a comprehensive ECW policy, including the following:
o Who can or cannot carry an ECW.
o Avoiding multiple deployments against one person.
o Limitations on bodily targets.
o Limitations on subjects for whom ECWs can be used against (e.g., pregnant women). o Limitations on time and place (e.g., subject is a fall risk or presence of flammable material). o Requirements for switching to alternative force options if ECW is not effective. o Post-deployment requirements.
Each of these elements are important and SJPD should be commended for comprehensive policies on use of ECWs.
Recommendation 12: The SJPD should maintain its current practice for this finding.
Finding 13: The SJPD Duty Manual doe s not provide sufficient clarity on some elements related to ECWs.
While many standard sections related to ECWs are found in the Duty Manual, the Duty Manual is incomplete in some sections and vague in others, thereby lacking in sufficient guidance for off icers.
For instance, the Duty Manual does not require an elevated standard of review for when officers use more than three ECW cycles against a subject. As written in the Duty Manual, officers shall refrain from using more than three five-second cycles unless a reasonable officer would conclude that each subsequent cycle is warranted based on the threat of harm. However, this is generally the same standard for any use of ECW cycle (or any other use of force) in the sense that the application must be reasonable under the totality of the circumstances.
Other sections related to ECW also appear to be vague, including requirements that officers “avoid” certain actions but not expressly prohibiting them or requiring a higher standard of review for such
21 actions. For instance, officers are instructed to “avoid activating more than one [ECW] against a single subject at the same time.” Similarly, officers are informed that the use of ECW should be generally avoided against “women who are known to be pregnant, very old or very young persons, physically frail persons , and persons with known heart conditions.” Furthermore, officers are informed that they should avoid targeting a subject’s head, neck , genitalia, or chest ; however, such avoidance is required only “if circumstances permit.”
The Duty Manual also states that officers shall not use an ECW against a person if that person is in “physical control of a vehicle in motion unless…the risk of serious bodily injury or death resulting from the subsequent movement of the vehicle is outweighed by the need to capture the subject.” While prohibiting officers using an ECW against an individual controlling a vehicle is a generally positive practice, basing the decision on whether the vehicle could cause injury is unnece ssarily limiting. Rather, the danger posed by the subject is more important in this context. For instance, a fleeing, violent felon in an empty parking lot would not pose a danger for the vehicle to cause serious injury; however, the danger posed by the subject in this situation would justify the use of an ECW.
Recommendation 13: The SJPD should revise the Duty Manual to provide concrete prohibitions where appropriate or substitute a higher standard of review (i.e., when deadly force would be authorized) for such uses of ECWs.
Use of deadly force
Finding 14: The SJPD Duty Manual allows for warning shots but does not provide sufficient framework around them.
Warning shots are inherently dangerous and strict guidelines should govern them if policy allows them. For instance, officers should have a significant degree of certainty that the warning shots will be effective in reducing the need for additional shots , and a number of factors (including backdrop , trajectory, and the subject’s mindset) should be explicitly laid out in policy. 33 Currently, the Duty Manual identifies only situations in which warning shots are allowed (i.e., events where lethal force would be justified) but does not include other considerations.
The SJPD may also rethink the benefit of allowing warning shots given the danger associated with them. Many gun owners (police and community members alike) are trained to fire only when left with no other option. If the subject is not considered dangerous enough to engage with immediately, one may wonder why deadly force would be needed in the first place. There has been little to no research on the effectiveness of officer warning shots, and community expectations for police use of firearms has changed over the years , leading agencies to move away from the practice . If SJPD does not decide to ban outright the use of warning shots, it may consider a general ban on the practice
33 IACP (2020). National consensus policy and discussion paper on use of force. Alexandria, VA: IACP. https://www.theiacp.org/sites/default/files/2020-
07/National_Consensus_Policy_On_Use_Of_Force%2007102020%20v3.pdf
22 except in extreme situations. As written, the Duty Manual appears to allow the practice in any lethal force situation.
Recommendation 14: The SJPD should reconsider the benefits on the use of warning shots; however, the SJPD should explicitly detail the situations and factors in which warning shots are allowed if it continues to allow their use.
Finding 15: The SJPD Duty Manual is inconsistent regarding which actions constitute deadly force and/or require an investigation consistent with a deadly force event.
The Duty Manual identifies actions by police officers requir ing deadly force justification. These actions include strikes to the head with a baton, use of ECWs, where the risk of falling is likely to cause serious bodily harm or death, and carotid neck holds. However, although the onus appears to be on the officer to justify deadly force, the investigation of such incidents do es not appear to be consistent with deadly force events (e.g., use of firearms). For instance, an intentional strike to the head with an impact weapon is listed as a Category III Use of Force despite officers needi ng deadly force justification. To reflect the potential for death or serious bodily injury resulting from such actions, uses of force that have the potential to cause serious bodily harm should be categorized and evaluated in a manner consistent with deadly force. While such events may not fit the definition of an OII in terms of requiring an investigation consistent with the Santa Clara County Police Chiefs’ Association OII guidelines, there are steps that officers and supervisors can take to secure the scene and safeguard the integrity of the investigation similar to th ose taken after a use of force that does meet the guidelines’ definition.
Recommendation 15: Where officers require deadly force justification, uses of force should be investigated in a manner consistent with deadly force events.
Finding 16: The SJPD Duty Manual section related to officers’ duty to provide a lethal force warning is vague.
As written, Section L 2602.1 (Deadly Force ) of the Duty Manual is vague and does not provide adequate guidance on when warnings are required by policy and when they are not. For instance, warnings are necessary “when feasible.” Officers are required to “make reasonable efforts” to identify themselves and provide a force warning unless it is “objectively reasonab le” to believe the subject already knows those facts. The degree of subjectivity in interpreting this section of the Duty Manual leaves officers without an explicit understanding of what is expected of them prior to a lethal force event. We recommend that the SJPD revise the language to create an affirmative duty for officers to identify themselves and provide a force warning unless it would present a clear and immediate danger to them or others to do so.
More broadly, the Duty Manual is not clear regarding whether a force warning is needed for any use of force (let alone deadly force). For instance, the introductory paragraph of Section L 2600 ( Use of
23 Force) states “attempts are made to achieve control through advice, warnings, and persuasion.” We agree such attempts should be made; however, there appears to be no affirmative duty on behalf of the officers to give force warnings where possible. Section L 2623 (Use of Police Service Dogs ) has a warning requirement for conducting searches using police dogs, t hough this may be negated when there are safety concerns. Aside from this situation, there is no broader requirement for a force warning. A review of use of force cases found real-world examples where force was not preceded by warnings, including one event where stun bag rounds were deployed without any warning. The SJPD should revise the Duty Manual to require warnings for all uses of force unless doing so would represent a danger to the officer or other community members or would significantly h inder the officer’s ability to make an arrest.
Recommendation 16: The SJPD should revise the Duty Manual to require force warnings for all uses of force unless time and circumstances do not allow for a warning.
Finding 17: The SJPD Duty Manual does not provide sufficient guidance for officer actions after a lethal force event.
Although responding supervisors and criminal/administrative investigators will complete many of the responsibilities for post -incident a ctions, first responding officers , as well as witness ing and involved officers, can take steps to ensure the integrity of the investigation. For instance, while the Duty Manual requires supervisors (via the Santa Clara County Police Chiefs’ Association OII guidelines) to ensure that witnessing and involved officers are separated, the Duty Manual does not require those officers to separate themselves. Furthermore, the Duty Manual does not list the responsibilities for personnel arriving after the application of lethal force, which may include beginning a scene log, marking evidence, or canvassing for witnesses. While the supervisor is ultimately responsible for these tasks, setting expectations for officers in the Duty Manual reinforces an affirmative duty to act before supervisors arrive on scene.
Recommendation 17: The SJPD should revise the Duty Manual to provide comprehensive guidance on post-incident actions that should be taken after a lethal force event.
Finding 18: The SJPD Duty Manual places the responsibility for post-lethal force notifications on the involved officer.
The Duty Manual states:
An officer who discharges a firearm either accidentally or in the performance of a police duty, except at an approved range, will protect the scene consistent with the safety of officers and members of the public and evidence and will ensure that an on-duty officer of the officer’s subdivision with the rank of lieutenant or highe r is notified as soon as time and circumstances permit. If an officer of the required rank is not on duty in the officer’s subdivision at the time of discharge, the officer will ensure that the watch
24 commander on duty, or any on -duty officer with the rank of lieutenant or higher, is notified.
Lethal force events are extremely traumatic for everybody involved. After using lethal force, officers are likely to be in some state of shock , and expectations for immediate actions by involved officers should be tempered. For instance, while involved members should be expected to protect the scene until others arrive, they should not be expected to navigate a phone tree if their immediate lieutenant is not a vailable. This situation can be resolved easily by requiring the officer to notify the communications unit, and then having the communications unit bear primary responsibility for making notifications to the officers’ chain of command. The SJPD should also give direction as to what information is given during this notification so as to not violate the Officers’ Bill of Rights.34 Recommendation 18: After using lethal force, the involved officer should communicate the force to the communications unit, which would assume responsibility for making notifications.
Finding 19: Family and friends of officer -involved shooting victims expressed concern about the quality of communication after the shootings and during the investigations.
As part of our interviews with community members, we spoke with family and friends of individuals who were subjects in officer-involved shootings. During our interviews, these community members expressed concern that officers did not treat them with respect and dignity after these shootings and during the investigation. In the same way that everyone responds to the death of a loved one, the community members we spoke with noted that such families are experiencing intense loss and grief, regardless of the circumstances of the officer-involved shooting. These community members also felt that the investigation of such events can further subjugate friends and family to additional emotional harm, a concept they did not believe was understood by the SJPD. We recommend the SJPD further explore the way such communication is handled, including interviewing SJPD members responsible for communicating with friends and family members (at the initial point of contact as well as during follow-up investigat ion). Where appropriate and where welcome d, we also recommend the SJPD initiate contact with friends and family members of prior officer -involved shooting victims to hear their firsthand experiences.
Recommendation 19: The SJPD should explore and gather feedback about communication with friends and family of officer-involved shooting victims. This should include interviewing SJPD members and, where appropriate and welcomed, the friends and family of subjects in the officer-involved shootings.
34 California Statewide Law Enforcement Association (2021). Government Code Sections 3300-3313, Peace Officer Bill of Rights. https://cslea.com/legal/peace-officer-bill-of-rights/
25 Finding 20: The SJPD Duty Manual does not provide sufficient guidance for supervisor actions after a lethal force event.
The Duty Manual references supervisors’ responsibility to ensure that post -lethal force events are managed in accordance with the Santa Clara County Police Chiefs’ Association OII g uidelines.
However, the guidelines act largely as a reference for what is to be done rather than who is expected to do it. For instance, while the guidelines discuss separating all involved and witness ing officers to ensure independent recollection, nothing in the Duty Manual identifies who is responsible for doing so (e.g., the first responding supervisor) nor do the guidelines discuss the need for monitoring the separation (e.g., assigning a supervisor to each involved officer). In addition, the Guidelines state that after a lethal force event, certain notifications shall be made, the first of which are “Intra-department, as required by that agency’s procedures.” Yet, notification requirements within the Duty Manual are few: the involved officer is required to notify a supervisor, and a command officer is required to notify the Homicide Unit Commander. The Guidelines also indicate that various individuals within the District Attorney’s Office should be notified, though the Duty Manual does not list who is responsible for notifying these individuals. Furthermore, lethal force events often require other notifications, including the chief, Internal Affairs, and city executives. The Duty Manual should identify how and by whom these individuals will be notified.
The Guidelines identify many of the steps we would expect to see though some important pieces are missing that can be further spelled out in the Duty Manual. For instance, the Guidelines do not discuss officers making on-scene public safety statements to ensure that the threat to the public has been neutralized. Since all lethal force events are criminally investigated and to safeguard against potential Garrity concerns, SJPD should ensure that public safety statements from the involved officer are requested only when the information cannot be gathered from another source (e.g., witness ing officer).35 The statement should also be limited to matters of public safety (i.e., number and direction of shots fired, injured persons, and whether the suspect is still at large). Such considerations should be included in the Duty Manual to lay out clear expectations for public safety statements as well as identify who is respo nsible for such statements. The Guidelines also do not discuss imposing restrictions on officer communication. Aside from a select set of individuals ( e.g., spouse, attorney), the involved officer and witness ing officers should be restricted from discussin g details of the incident with others until the completion of administrative and criminal investigation s. Having this information in the Duty Manual will help safeguard the integrity of the investigation.
Recommendation 20.1: The SJPD should revise the Duty Manual to identify the required post-incident tasks to be performed.
Recommendation 20.2: The SJPD should revise the Duty Manual to identify who is responsible for performing post-incident tasks.
Recommendation 20.3: The SJPD should revise the Duty Manual to include requirements and constraints on gathering public safety statements.
35 Garrity v. New Jersey holds that an officer's compelled statements during an administrative investigation cannot be used as part of a criminal investigation.
26 Mass demonstrations
Finding 21: The SJPD conducted a comprehensive after-action report of the 2020 social justice movements related to the murder of George Floyd , though many of the resulting recommendations have yet to be implemented.
During the summer of 2020, like other agencies across the country, San José experienced significant social justice mass demonstration movements related to the murder of George Floyd. Also similar to agencies across the country, issues related to policy, training, personnel, and operations hindered the SJPD’s response to the movements, some of which we describe in more detail below. However, SJPD should be credited with swiftly and thoroughly evaluating its response through a comprehensive after-action report (AAR).36 Within three months of the movements ending, the SJPD had completed the AAR, reviewing 25 different topical areas, including, among others, the background of the events, the information known prior to the events, actions taken during each day of the events, equipment, use of force, and responses to each day of the events (including shifting tactics across the entire set of events). In total, the SJPD identified 41 separate recommendations across five thematic categories.
In all, the AAR appears to demonstrate a sincere effort on the part of the SJPD to assess its response, identify areas for improvement, and create responsive recommendations.
However, despite undertaking the effort to conduct the AAR, many of the recommendations have still not yet been implemented. For instance, the AAR gave significant attention to the fact that ma ny officers had received minimal substantive training on crowd control response and were unprepared for responding to the social justice movements . On page 4 of the AAR, the SJPD readily admits this, saying, “Most of the Department’s officers have never ex perienced civil unrest of this type. Crowd control training has been minimal and infrequent as mass training requires times away from already depleted patrol staffing.” Yet the only training that officers have received since the AAR have been roll-call trainings, which are limited in their ability to allow extended discussion. No in-service crowd training was provided in 2021 nor does it appear there is planned crowd control training for 2022 in-service trainings. This is just one example (though a critical one), given the importance of this issue to the SJPD. We recommend the SJPD provide a public update on each of these recommendations about the status of implementation, including a timeline for implementation (if applicable) , to incorporate an element of transparency and accountability.
Recommendation 21: The SJPD should provide a public update on the implementation of the AAR, including a timeline for implementation, if applicable.
36 San José Police Department (2020). Police department preliminary after action report for the public protests, civil unrest, and law enforcement response from May 29 – June 7, 2020. San José, CA: San José Police Department. https://sanjose.legistar.com/View.ashx?M=F&ID=8769493&GUID=3ED4A6F5-F069-4E7F-BADE-99421D9991B3
27 Finding 22: The SJPD Duty Manual does not provide sufficient guidance for officers when responding to a crowd control event.
Duty Manual Section L 2300 ( Demonstrations and Civil Disturbances ) does not offer sufficient guidance for officers responding to a social justice movement . Of primary concern is that the Duty Manual does not identify a command structure for events that are escalating, nor does it identify who is responsible for Mobile Field Force (MFF) or specialized unit call-outs. For instance, if the crowd is relatively small and can be monitored by a first responding sergeant, then no additional notifications need to be made. However, as the crowd grows (either in number or in destructive demeanor), additional resources will be needed. The Duty Manual should clearly identify the criteria needed to make such call -outs. Related to this issue, the Duty Manual does not provide definitions for characterizing crowd behavior, including definitions for protest, unlawful assembly, crowd management, crowd control, and riot. Such definitions would be helpful in determining whether additional call-outs would be appropriate. Overall, there is an absence of clear authority and decisive factors in the Duty Manual, which may lead to a delay in decisions being made.
Many social justice events are also subject to at least some degree of planning, meaning police departments also have some ability to plan for a response. Department members should reach out to movement leaders in an attempt to coordinate efforts for a safe exercise of constitutional rights. Per the SJPD’s AAR, this happened to some extent during the George Floyd social justice movements ; however, this practice is not memorialized in the Duty Manual. Similarly, nothing exists in the Duty Manual about the need to maintain ongoing communication, particularly when surgical arrests need to be made that may require the cooperation of movement leaders. While these things may occur in practice, memorializing them as duty -bound requirements will emphasize their importance throughout the department.
Finally, post -event debriefings between the incident manager and command staff should be a requirement in the Duty Manual . The Duty Manual should also include guidance about what such debriefings should entail (e.g., assessment of the response, uses of force, and potential policy/training implications). As with the above, the SJPD appears to have done an excellent job of this as it relates to shifting tactics to constitute a more hidden presence rather than an authoritative force.
Memorializing these practices in the Duty Manual reinforces the importance of these debriefings.
Certainly, these are not the only policy revisions that will be necessary based on the extensive review conducted by SJPD as well as based on an upcoming independent AAR. In fact, a primary finding from the SJPD AAR explicitly states that SJPD should “engage the community in a comprehensive review of the Department’s policies and procedures applicable to crowd control events and use of force” (p. 99). We bring these issues up solely as examples of significant holes in the current policy and recommend the SJPD use its AAR, the upcoming independent evaluation, and community feedback to identify additional areas where the policy might be enhanced.
Recommendation 22.1: The SJPD should revise the Duty Manual to identify a command structure for escalating events, who is responsible for call- outs, and the criteria for such call-outs.
28 Recommendation 22.2: The SJPD should revise the Duty Manual to memorialize the requirement for initial and ongoing communication with social justice movement organizers.
Recommendation 22.3: The SJPD should memorialize the requirements for post- event debriefings between the incident commander and command staff and what those debriefings should entail.
Finding 23: Community members expressed concern regarding SJPD’s use of force during the 2020 social justice movements.
While o ur assessment looked at use of force more broadly, one theme that emerged from our interviews with community members was concern about the application of force in response to social justice movements, particularly in the immediate aftermath of the death of George Floyd and other social justice movements on a variety of topics since May 2020. Community members expressed significant concerns that officers were inclined to use force during the social justice movements with little or no provocation or justification. Several community members shared firsthand experience of being at the events and witnessing peaceful participants being targeted with rubber bullets, including shots to the face and head. Community members also noted that officers seemed to respond differently depending on the topic or political tone associated with the event, with some social justice movements, resulting in a far more combative and aggressive response from the SJPD while during other social justice movements, officers were congenial and supportive of participants. Although the SJPD’s AAR sufficiently discusses the legal and operational elements of force used during the 2020 social justice movements , the perspectives of community members at the events should also be addressed. Whereas the SJPD may find such actions to be in policy, community members may still find them to have been ultimately unnecessary and the SJPD will need to keep this in mind wh en engaging the community in a comprehensive review in the department’s crowd control and use of force polices.
Recommendation 23: The SJPD should recognize that community members’ perspectives of use of force during social justice movements may differ from departmental findings and incorporate that recognition when engaging the community in a review of policies.
Finding 24: The SJPD has not provided regular training on crowd control to all officers.
We begin by acknowledging that this was a self-reflective finding by the SJPD in the AAR, noting that “while special operations train s in MFF tactics at least quarterly, patrol officers have not been provided any MFF training since December 2018, unless they happen ed to graduate from the Academy since that time” (p.30). However, as noted above, no formal classroom - or scenario-based training has occurred since the findings were made, leaving the potential for officers to believe that the response to the social justi ce movements had little room for improvement. While we do not believe that crowd control training specifically needs to occur annually, there are ways that SJPD could incorporate elements into other types of training. For instance, if the SJPD puts on a sc enario
29 involving force, they may consider having “community members” (e.g., officer actors or hired actors) see the force event and begin shouting at the officers. The officers would then have to consider whether to notify a supervisor and whether addition al resources would be necessary to control the crowd. We recommend the SJPD implement its own findings and consider how crowd control dynamics can be introduced in other types of training to reinforce key concepts.
Recommendation 24.1: The SJPD should provide training relevant to the findings of its 2020 AAR.
Recommendation 24.2: The SJPD should incorporate elements of crowd control into other types of training.
Providing first aid
Finding 25: The SJPD Duty Manual does not contain sufficient instruction related to officers’ duty to provide medical attention.
For each use of force type listed in the Duty Manual, there is information regarding the requirements for providing medical attention. However, the instruction in the Duty Manual is at times incomplete or requires officers to make determinations they are not qualified to make. For instance, Section L 2608.6 (Providing First Aid ) and Section L 2 622 (Providing First Aid ) state that when a use of force “causes injury which would reasonably require medical attention, the officer using force shall ensure the injured individual received proper medical attention as soon as practicable.” As written, the terms “causes” and “reasonably require medical attention” are subjective and, given that officers are not medical professionals, they should not be making this determination.37 Rather, officers should take a risk-averse approach by summoning medical units any time there is a complaint of injury or pain and providing interim first aid in accordance with the training they have received. Additionally, while Section L 2608.6 (Providing First Aid) requires the officer who used force to ensure that the subject receives proper medical attention, this responsibility should apply to all officers on the scene. In addition to reinforcing the primacy of community member safety, there is also the potential that the recipient of force may desire for another officer to provide the first aid and the policy should account for that.
Other sections of the Duty Manual are more expansive in the instruction given to officers. For example, Section L 2610 (Providing First Aid) discusses steps that officers are expected to take aft er the use of chemical agents, including placing the subject in an upright position, flushing the subject’s eyes with clean water, requesting an on -scene medical response, and observing the subject until medical personnel arrive. The SJPD should consider this section a template for detailing the steps officers are required to take after using other force tools and options.
37 Related to this, as part of providing first aid after the use of a chokehold or carotid restraint, the Duty Manual states that officers will inform medical professionals that the arrestee “should be placed under observation.” While officers should certainly inform medical professionals of the use of chokeholds or carotid restraints, medical decisions should not be dictated by police policy.
30 Recommendation 25.1: SJPD officers should take a risk-averse approach by summoning medical units any time there is a complaint of injury or pain and providing interim first aid in accordance with the training they have received.
Recommendation 25.2: The SJPD should use Section L 2610 (Providing First Aid) as a template for detailing the medical steps officers are required to take after using other force tools and options.
Finding 26: SJPD officers are not required to prepare to provide first aid.
Several sections of the Duty Manual require officers to render first aid when necessary. However, they are not required by policy to carry first aid kits or individual first aid kits (IFAKs). Officers should be required to carry IFAKs—which are designed to treat potentially deadly injuries with items such as chest seals —in their vehicles to be prepared when encountering a severely injured person. In addition, Section L 2610 (Providing First Aid) discusses the requirement for officers to flush subjects’ eyes after using chemical agents; h owever, policy does not require officers to carry water to accomplish this task, leaving the officer with a requirement but not the means to act. This is particularly important in the context of crowd control, where officers must be prepared to fulfill that duty (assuming a reasonable opportunity to do so).
Recommendation 26: The SJPD should revise the Duty Manual to require officers to carry individual first aid kits and water for use after a force event.
Use of force report
Finding 27: The SJPD use of force report uses an outdated reporting system.
The SJPD provided the use of force reporting template to the assessment team as a pdf file. This template detailed how officers and supervisors are expected to go through the automated use of force report and respond to it s items and questions. A r eview of this template and the data that are developed from its use reveal an inadequate use of force reporting system for an agency the size of the SJPD. The reporting system seemed to not be user friendly, ensure checks for data quality, allow for quick review or analysis, nor provide reports on use of fo rce information. We recommend the SJPD update its use of force reporting procedure by implementing a use of force reporting system that can better allow for information entry, case tracking, review, analys es, and summary report creation. For example, the IAPro software provides a set of application s that better collect, assess, and manage professional standards for supervisory officers compared with the current SJPD system.
Recommendation 27: The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation.
31 Finding 28: The SJPD uses a use of force template to report use of force and has recently released a Use of Force Documentation Guide to aid in completing the template.
The SJPD uses a use of force template for officers to report their use s of force and in June of 2021, SJPD released a Use of Force Documentation Guide as a companion to the template to “identify issues to be considered for any use of force incident.” The Use of Force Documentation Guide includes more than 30 points that should be considered as part of force documentation. The guide includes directives on documenting information about actions preceding the use of force, the use of force itself, and post -force actions. Overall, both the guide and the template are comprehensive and when combined, appear to demonstrate a departmental desire for officers to thoroughly describe the totality of circumstances. However, there is room for improvement , specifically in how these efforts are emphasized across trainings, guidelines, and reporting requirements. For instance, the format of the Guide and its areas of consideration are at the event level, whereas the use of force reporting template is at the officer -community member level and officers are expected to provide a nar rative summary that includes a description of each application of force. As such, the guide does not emphasize that each application of force is to be coded independently, as the reporting template and narrative require. We encourage the SJPD to stress that uses of force are documented at a more micro level across all trainings, guidelines, and reporting requirements to ensure that each application of force is reasonable. For instance, we refer to the downloadable data provided by the Portland Police Bureau, which separates each application and requires corresponding information for each application.38 This will be especially pertinent if the SPJD implements a new use of force reporting system.
In addition, although a more minor point, the SJPD guide often includes yes/no questions that could be phrased differently to help officers refine their thinking when describing the force used. For instance, one question asks, “Was the level of force reasonably proportionate to the level of crime, level of threat, and level of resistance?” For any officer using constitutional force, the answer to this question should always be “Yes,” and this prompt does little to address the question of “how” the force was proportionate given the standards laid out in Graham v. Connor.
Recommendation 28: The SJPD should ensure that all trainings, guides, and reporting requirements emphasize that each application of force is to be coded independently.
Finding 29: The SJPD’s use of force data collection form allows for each specific combination of event, involved officer, type of force, sustained injuries, and involved community member to be assessed.
Cases in the SJPD’s use of force database developed from its use of force report template account for each combination of event, officer, force used, injuries sustained (by both community member and officer), and involved community member. This allows for analysis of eac h individual event of force
38 Portland Bureau of Police Use of Force Report. https://www.portlandoregon.gov/police/76875
32 within an overall event and is a best practice for use of force data management. For events with multiple officers and multiple community members, it is straightforward to understand which officers used force, what type of force they used, and against which community members. This allows for a better understanding of disparity in use of force.
The information on sustain ed injuries by community members and officers as a result of the use of force event is an additional component not often documented by police departments. This is a unique and important data field that the SJPD requires officers to report.
Recommendation 29: The SJPD should maintain its current practice of documenting use of force incidents at this level of detail.
Finding 30: The manual entry of information into the SJPD use of force report creates data inconsistencies.
As detailed in the Quantitative Analyses section in Section 2 of this report , we observed that the manual entry of names and event identification numbers increased the likelihood of errors in the use of force data. For example, the use of force report fill ed out by SJPD officers include s five fields for the involved community member’s name. The first field uses a drop -down menu of a known list of entities to select from, the second field is meant for the community member’s last name, and the remaining three fields are meant for any other “given names” that could be entered in any order. This created an opportunity for inconsistencies to exist in the data, as multiple officers creating a use of force report may enter community members’ names in different ways . Furthermore, there was no identification number attached to individuals to determine if they were involved across different use of force events. Furthermore, there were instances where the General Offense Report Number was entered incorrectly across officer reports for the same use of force event, resulting in an over inflation of unique use of force events. And there were instances where officers created multiple use of force reports for the same community member within an event. These data entry issues created multiple inconsistencies in the dataset, requiring significant data cleaning prior to any analyses. A new use of force reporting system that automatically ensures data quality may alleviate these problems.
Recommendation 30: The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation.
Finding 31: Duplicate items in the SJPD use of force report create data inconsistencies.
In our review of the use of force report template, we observed multiple items that asked the same question in slightly different formats. A new use of force reporting system that does not include duplicate items may al leviate these problems. Below, we detail the specific duplicate items in the current template ; although the wording may be slightly different than what is used in the actual report:
• Duplicate items on firearm discharge:
o Did officer discharge a firearm?
33 o Type of force on suspect: Discharge of firearm (miss) o Type of force on suspect: Discharge of firearm (hit)
• Duplicate items on use of force on community member:
o Did officer use force on a suspect?
o Did officer use force on this suspect?
o Did officer use force against a suspect (or multiple suspects)? • Duplicate items on community member injury:
o Did this cause serious injury or death to a suspect?
o Was suspect injured (even if minor)?
• Duplicate items on community member assault of officer:
o Did the suspect(s) assault officer?
o Did the suspect assault officer?
o Were officers assaulted by suspect(s)?
o Level of suspect resistance
• Duplicate items on officer injury:
o Did this cause serious injury to officer?
o Was officer injured (even if minor)?
• Duplicate items on officer death:
o Did this incident cause death to an officer?
o Officer's injury level
• Duplicate items on community member arrest:
o Applies to incident: An arrest was made
o Encounter Outcome - In Custody
o Did the use of force result in taking the suspect into custody? Recommendation 31: The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation.
Finding 32: The lack of data quality check allows for data entry errors in the SJPD use of force report.
The SJPD use of force reporting system does not conduct within-system data quality checks, resulting in entry errors across items. This oc curs both across the duplicate items, where an officer may respond in the affirmative to one item and the negative to another, as well as in items that cover similar topics. For example, an officer could respond that the community member did not exhibit any irregular behaviors, but then proceed to affirm specific irregular behaviors that were present. Of the 938 cases that were marked as having no irregular behavior, 90 were also marked as having alcohol impairment, 90 were also marked as having mental disability, and 185 were also marked as having drug impairment. A new use of force reporting system that automatically ensures data quality may alleviate these problems.
Recommendation 32: The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation.
34 Section 3: Use of Force Events
From February 17, 2017, to February 27, 2021 (roughly 4 years), 936 unique SJPD officers completed 4,817 unique use of force reports related to 2,352 use of force events. The SJPD requires all officers involved in a use of force event to fill out a report for each involved community member and therefore one event can have multiple reports if there are two or more officers or two or more community members involved. Our analyses aggregated the report data to the community-member-per-event unit of analysis (i.e., each case in the dataset represented a community member who experienced a unique use of force event ), which totaled 2,743 community members across the 2,352 use of force events. It is worth noting that there were some community members with multiple instances of use of force in the data, meaning that the total amou nt of unique community members in the database was less than 2,743.
Figure 2 details the counts of use of force by unique events, community members, officer reports, and officers from March 2017 to February 2021 . Figure 3 displays the same as monthly avera ges.
Use of force levels were relatively stable prior to the COVID -19 pandemic, with an average of 50.2,
53.3, and 51.8 use of force events per month for 2017, 2018, and 2019, respectively. In 2020 , the average amount of use of force events per month dropp ed to 41.8, coinciding with the large -scale onset of COVID-19 as well as social justice movements that took place in San José for policing reform.
Only two months were examined in 2021, but they averaged 37.0 use of force events, keep ing pace with 2020 counts.
Figure 2: Count of use of force, by unit and year
0 200
400
600
800
1000
1200
1400
1600
03/2017-12/2017 2018 2019 2020 01/2021-02/2021
Events Community Members Reports Officers
35 Figure 3: Use of force averages per month, by unit and year
The average number of use of force events at the individual officer level was 5.14, with a median of 3, across the 931 officers involved in a use of force event from March 2017 to February 2021. At a yearly level, the average use of force events per officer per year was 2.44 from 2018 to 2020. Figure 4 details the distribution of individual use of force events across the officers. Of these 931 officers, 352 officers (38 percent) had only 1 or 2 use of force events , 215 officers (23 percent) had 3 or 4 events, and 117 (13 percent) h ad 5 to 6 events. The remaining 247 officers (27 percent) had more than 7 use of force events from March 2017 to February 2021. A total of 56 officer (6 percent) had 15 or more use of force events. It is worth noting that the 2019–2020 annual city report advised that, as of Ju ne 2020, the number of street-ready sworn SJPD officers was 954, indicating that the vast majority of SJPD recorded at least one use of force event during the examined time period.
Figure 4: Distribution of the amount of use of force events across 931 officers
0 20
40 60
80 100
120
03/2017-12/2017 2018 2019 2020 01/2021-02/2021
Events Community Members Reports Officers
206
146
117
98 62 55 45
31 33 26 17 11 11 17 9 12 5 9 1 5 1 2 0 5 2 0 0 1 1 1 0 0 0 0 0 2 0 50
100
150
200
250
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Number of Officers
Number of Use of Force Reports from March 2017 to February 2021
36 The SJPD offers a field for the reporting officer to detail the initial contact reason, which the assessment team recoded into five groupings, detailed in Figure 5. The majority of use of force events (78 percent) began as officers responding to calls for service and/or a crime report. The second leading contact reason that resulted in a use of force was officer-initiated traffic and pedestrian stops, which accounted for 15 percent of cases. A use of force was recorded in 94 apprehensions, which include arrests, and make up only 3 percent of all use of force cases across community members.
Another 4 percent of cases were the result of other initial contact reasons, such as assisting an agency, officer request for assistance, and ambushes.
Figure 5: Initial reason for community contact that resulted in use of force
The SJPD use of force report offers three categories of reasons that officers used force during community encounters, detailed in Figure 6, which include use of force to effect an arrest, prevent an escape, or to overcome resistance. The officer can choose to not select any of these reasons or can select all three. Figure 6 details the mutually exclusive reasonings officers provided for the use of force. The majority of cases (42 percent) indicated that the officer used force to prevent escape and overcome resistance from the community member , while 3 8 percent of cases reported all three behaviors (effect arrest, prevent escape, and overcome resistance) as the reason for the use of force.
A total of 92 percent of cases included overcoming resistance (independently or in combination with another reason) as a reason for the use of force. It’s worth noting that 75 cases (3 percent) did not mark any reason for the use of force.
Responding to a Crime
Report or Call for
Service, 2147, 78%
Apprehension, 94, 3%
Traffic or Pedestrian
Stop, 404, 15%
Other Reason, 98, 4%
37 Figure 6: Reason for use of force
The SJPD requires the officer to report whether or not they perceived the community member to be armed with a weapon. If officers reported that they perceived the community member to be armed, they are then required to report what type of weapon they perceived, either a firearm; knife, blade, or stabbing instrument; other dangerous weapon; and/or an unknown weapon. Of the 2,743 use of force events, officers in 1,467 events (53.5 percent) advised that they perceived the suspect to be armed; however, there were some discrepancies in the data. First, 11 events reported the perception of an armed community member but no specific weapon was reported in follow-up questions. These cases were subsequently treated as unknown. Second, an additional 126 events did not report an officer’s perception of an armed community member but fields for the specific perceived weapons were reported in the follow-up questions; although 120 were for the “unknown” category. With the inclusion of 126 events, the total number of events where a community member was perceived to be armed was 1,593 events (58 percent). Another way to state this is that officers did not perceive the community member to be armed in 1,150 use of force events (42 percent) from March 2017 to February 2021.
Figure 7 details the unique combinations of the types of weapons in the 1 ,593 use of force events where the officer perceived an armed community member. Of note, officers in 640 events (40 percent) reported an “unknown” weapon, by far the most frequent res ponse option noted . The second most frequently reported perceived single weapon type was a knife, blade, or stabbing instrument with officers in 253 events (16 percent) reporting a knife as the only weapon. In total, knifes, blades, or stabbing instrument were reported in 550 events (35 percent). The “other dangerous weapon” category was the second most frequency single weapon type being reported in 182 events (11 percent); however, other dangerous weapons were reported being perceived in 483 1,040
1,160
134
48 201
60 25
75 0 250 500 750 1,000 1,250
To effect arrest, prevent escape, and overcome
resistance
To prevent escape and overcome resistance
To effect arrest and overcome resistance
To effect arrest and prevent escape
To overcome resistance
To prevent escape
To effect arrest
No reasons selected
Number of Use of Force Events
38 events in total. A firearm being the only weapon was reported in 143 events (9 percent) but across all combinations of weapon types, firearms were present in 359 events (23 percent) where a community member was perceived to be armed . It’s also worth noting that 67 events had all three weapon categories marked in the report (firearm, knife, and other) and an additional 91 events had all four weapon categories marked . These 158 events, in addition to the 640 “unknown” weapon events, perhaps allude to officers not filling the report in as much detail as they should be.
Figure 7: Unique combinations of the types of perceived weapons on community members
The SJPD records the officers’ description of the community member ’s resistance during the use of force event. Figure 8 details that 65 percent of community members was described as actively resisting officers’ commands. Another 22 percent of community members’ resistance was described as assaultive toward the officers. A total of 9 percent of community members were described as passively resisting the officer(s), and the remaining 4 percent were described as either fleeing or fled the officer, or their resistance level was described as unknown.
640
46 182
31 34
48 253
7 14
11 91
9 67
17 143
0 100 200 300 400 500 600 700
Unknown
Other & Unknown
Other dangerous weapon
Knife, Other & Unknown
Knife & Unknown
Knife & Other
Knife
Firearm, Other, & Unknown
Firearm & Unknown
Firearm & Other
Firearm, Knife, Other, & Unknown
Firearm, Knife, & Unknown
Firearm, Knife, & Other
Firearm & Knife
Firearm
Number of Use of Force Reports with Listed Weapon Type
39 Figure 8: Officer descriptions of community member resistance
The SJPD also assesses the presence of irregular behaviors from the involved community member , which is defined in the use of force report as “erratic” behavior. Figure 9 shows the amount of community members who officers described as presenting such behaviors, which were separated by different disabilities and impair ments. In 793 cases (29 percent), the officer reported that the community member showed no signs of a disability or impairment. The remaining cases had one or more of the irregular behaviors listed as present during the encounter. Figure 8 details the numb er of cases that that listed the specific behavior as being present (i.e., these categories are not mutually exclusive). A physical disability was perceived in 1 percent of cases, a developmental disability in 3 percent of cases, and mental disability in 3 0 percent of cases. In total, 31 percent of community members had one or more disabilities as assessed by the officer.
Figure 9: Irregular behaviors of community member involved in use of force event
Drug an d alcohol impairment were also recorded , detailed in Figure 9 . Officers reported that community members had an alcohol impairment in 32 percent of cases and a drug impairment in 47 percent of cases. In total, 62 percent of community member were observed to have an alcohol or drug Assaultive, 591, 22%
Active, 1784, 65%
Passive, 244, 9%
Fleeing/Unknown, 124, 4%
1,282
869
823
82 29
793
0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 Signs of drug impairment
Signs of alcohol impairment
Signs of mental disability
Signs of developmental disability
Signs of physical disability
No signs of disabilities or impairments
Number of Use of Force Events
40 impairment during the police encounters that resulted in a use of force. Combining these behaviors, 71 percent of community members presented some level of irregular behavior as assessed by the officer.
The SJPD includes 17 categories to describe the type of use of force used during the event, presented in Figure 10, loosely ordered from least to most severe. 39 The lowest levels of force that include physical contact and control holds/ takedowns accounted for the majority of use of force applied during these events. Across all cases, these lower-level use of force activities were used in 65 percent of the cases and were effective without any additional use of force activities, and applied in another 17 percent cases with other, more severe, levels of force. Use of ECW, including taser deployments, was present in roughly 15 percent of use of force cases, whereas the threat of an officer using their firearm (i.e., pointing their firearm) occurred in 2 percent of cases and officers firing in less than one percent (0.33 percent) of use of force cases.
Figure 10: Incidences of use of force by type of force used
39 The analyses in Section 5 used a more complex severity hierarchy 7 3
59 10
0 113
184
8 8
282
294
115
76 376
90 2,044
869
0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 Discharge of firearm (hit)
Discharge of firearm (miss)
Threat of firearm
Officer vehicle contact
Knife, blade, or stabbing instrument
K-9 contact
Impact projectile
Carotid restraint control hold
Other dangerous weapon
Blunt/impact weapon
TASER probes deployed
TASER drive stun
TASER probes deployed (miss)
Electronic control device
Chemical Spray (e.g., OC/CS)
Other control hold/takedown
Other physical contact (fists, feet, etc.)
Number of Use of Force Events
41 Racial disparity in use of force events
A critical aspect of our examination of SJPD use of force was to assess its force applications across different racial categories. Figure 11 presents the trend lines of the count of community members each month involved in a use of force event in total and by race. San José community members experienced a monthly average of 56.8 use of force events from March 2017 to February 2021 ; whereas the monthly average amount of face-to-face police-community interactions resulting from a call for service was 6,812 (or, stated another way, one use of force event for every 119.9 interactions).
The overall trend line for the data (depicted as the black -dotted line in Fig ure 11) shows a slight decline in the amount of community members involved in a use of force event across the review period; however, much of that decline is likely a result of fewer community contacts due to the COVID-19 pandemic (see Section 4 for a more detailed analysis). The trend lines also show relatively flat distributions of use of force across the racial categories, with Hispanics experiencing an average of roughly 32 use of force events per month, white community members experiencing about 12 events per month on average, Black community members experiencing about 8 per month, Asians experiencing 4 per month, and individuals of other or unknown races experiencing roughly 1 use of force event per month.
Figure 11: Use of force events per month by community member race
0 10
20 30
40 50
60 70
80 90
white Black Hispanic Asian Other Race Total
42 In addition to the raw numbers, we also assessed differences in use of force by the race of the community member. However, u nderstanding racial disparities in law enforcement activity is difficult given the complexity of developing a baseline against which to compare law enforcement activity.40 Using Census Bureau population numbers will typically result in overestimates of disparity, since these numbers do not account for diff erential rates of contact with police among different demographic groups. That is to say, policing activities are not uniform across a city. On the other hand, using law enforcement data sources as a baseline will typically underestimate disparity, since these data may mask systemic racism and bias inherent in the criminal justice system as a whole (e.g., over-policing or selective arrests).
To visually highlight such issues, Figure 12 presents demographics from three different datasets: the racial distribution of all 1.02 million San José community members as determined from the 2019, 5 - year American Community Survey estimates 41; the racial distribution of 87,880 SJPD arrests from March 2017 to February 2021; and the racial distribution of 2,743 use of forc e events from March 2017 to February 2021.
From a population perspective, San José is a diverse community, with a majority population that is Asian (36 percent) or Hispanic (31 percent). White community members make up approximately 26 percent of the population, and Black community members are the smallest racial group, just 3 percent and less than those who are of another or multi-race (4 percent). When evaluating force solely based on population, the data indicate a disparity between certain racial group ings and the use of force experienced by those groups. In particular, Black and Hispanic community members are arrested and experience use of force at higher levels compared with their populations in the community. The proportion of Hispanics who experienc e use of force was 25 percentage points greater than their proportion in the community (56 percent compared with 31 percent) and it was 11 percentage points greater for Black community members (14 percent with use of force compared with 3 percent in the population). These differences are made up in reductions of the proportion of white and Asian community members compared with their use of force events. The proportion of white individuals who experienced use of force was 5 percentage points lower than their proportion in the community, and for Asians, the proportion of use of force was 29 percentage points lower than their proportion in the community.
There is much greater parity between race and use of force when using arrests as the baseline.
Hispanic comm unity members were the most frequently involved in these events, making up approximately 55 percent of community members arrested or involved in use of force, 21 percent being white, 13 to 14 percent being Black, and 7 to 9 percent being Asian. In roughly 2 percent of the cases, the community member was of another race or ethnicity or their race was unknown.
40 Tregle, B., Nix, J., & Alpert, G. P. (2019). Disparity does not mean bias: Making sense of observed racial disparities in fatal officer-involved shootings with multiple benchmarks. Journal of Crime and Justice, 42(1), 18-31. 41 American Community Survey, Table B03002, 2019, 5-year estimates:
https://data.census.gov/cedsci/table?q=B03002&g=1600000US0668000&tid=ACSDT5Y2019.B03002
43 Figure 12: Percent of race groups in San José population, arrests, and use of force events
To further assess these issues, we calculated three disproportionality indexes that examine disparity levels in different ways. This index is expressed as a compound ratio—the ratio of the percentage of police interactions with individuals of color involved in use of force to the corresponding percentage for white individuals. We first calculated this index by examining the amount of use of force events by community members compared with the populations in San José:
𝑈𝑠𝑒 𝑜𝑓 𝑓𝑜𝑟𝑐𝑒𝐵 𝑜𝑟 𝐻 𝑜𝑟 𝐴
𝑆𝑎𝑛 𝐽𝑜𝑠é 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝐵 𝑜𝑟 𝐻 𝑜𝑟 𝐴
⁄
𝑈𝑠𝑒 𝑜𝑓 𝑓𝑜𝑟𝑐𝑒𝑊
𝑆𝑎𝑛 𝐽𝑜𝑠é 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑊
⁄
The above method offers the least restrictive model , leading to observations of greater disparity levels. We also calculated the disparity index by examining the amount of use of force events by community members compared with arrests:
𝑈𝑠𝑒 𝑜𝑓 𝑓𝑜𝑟𝑐𝑒𝐵 𝑜𝑟 𝐻 𝑜𝑟 𝐴
𝐴𝑟𝑟𝑒𝑠𝑡𝑠𝐵 𝑜𝑟 𝐻 𝑜𝑟 𝐴
⁄
𝑈𝑠𝑒 𝑜𝑓 𝑓𝑜𝑟𝑐𝑒𝑊
𝐴𝑟𝑟𝑒𝑠𝑡𝑠𝑊
⁄
This method restricts the ratios to law enforcement data, which typically underestimates disparity levels since such data include known disparities by race due to systemic racism and bias in the criminal justice system as a whole. To illustrate this point further, we offer a third disparity index that examines the amount of arrests by community members compared with the populations in San José:
𝐴𝑟𝑟𝑒𝑠𝑡𝑠𝐵 𝑜𝑟 𝐻 𝑜𝑟 𝐴
𝑆𝑎𝑛 𝐽𝑜𝑠é 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝐵 𝑜𝑟 𝐻 𝑜𝑟 𝐴
⁄
𝐴𝑟𝑟𝑒𝑠𝑡𝑠𝑊
𝑆𝑎𝑛 𝐽𝑜𝑠é 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑊
⁄
25.7%
31.6%
35.7%
2.8% 4.1%
20.8%
55.1%
9.3%
13.0%
1.8%
20.6%
55.6%
7.3%
14.0%
2.4%
0%
10%
20%
30%
40%
50%
60%
White Hispanic Asian Black Other or Multi-
RacePopulation Arrests Use of Force
44 Note that none of the se three methods provide a single correct estimation in the level of disparity and all come with a degree of error .42 Instead, we offer all three to show that disparity levels likely fall somewhere between our provided minimum and maximum values. In an ideal analysis, we would calculate the disparity index using use of force events to the number of face-to-face, police- community interactions. Unfortunately, demographic information cannot be reliably collected by the SJPD for each community interaction, a common issue in policing data.
Figure 13 presents the population base disparity indexes. Using this approach, the disparity index for Black community members is 6.12 across all use of force events , meaning that Black community members are involved in use of force incidents 6.12 times more often than white community members when using population as the baseline . For Hispanic community members, the disparity index was 2.14 across all use of force events, and it was 0.24 for Asian community members.
Figure 13: Disparity indexes of use of force to San José population by race 03/2017 –
12/2017
01/2018 –
12/2018
01/2019 –
12/2019
01/2020 –
12/2020
01/2021 –
02/2021 Total
Compared with White
Black 7.15 5.24 5.47 7.07 7.00 6.12
Hispanic 2.02 1.86 2.10 2.66 2.84 2.14
Asian 0.21 0.27 0.31 0.16 0.22 0.24
By comparison, the disparity indexes using arrests as the baseline are detailed in Figure 14. Using this benchmark, Black community members were involved in use of force incidents only 1.09 times more often than white community members. For Hispanic community members, the disparity index using arrests was 1.02, and it was 0.80 for Asian community members compared with white community members. Therefore, the disparity indexes in Figure 14 show that, consistent with each racial group’s proportion of arrests, use of force is mostly equal among Black and Hispanic community members compared with white community members, though force was used less against Asian community members compared with white community members.
Figure 14: Disparity indexes of use of force to arrests by race 03/2017 –
12/2017
01/2018 –
12/2018
01/2019 –
12/2019
01/2020 –
12/2020
01/2021 –
02/2021 Total
Compared with White
Black 1.15 1.00 0.94 1.33 1.30 1.09
Hispanic 0.96 0.96 0.97 1.20 1.43 1.02
Asian 0.67 0.95 1.03 0.49 0.76 0.80
42 For example, we note that not everyone arrested by the SJPD actually resides in San José. As such, using the US Census population as a baseline statistic will underestimate the number of individuals who potentially could come into contact with an SJPD officer (i.e., a “community member” as defined in this report). The assessment team was not capable of removing non-San José residents from the analyses; however, the results were unlikely to be significantly affected by this limitation. Furthermore, the US Census population values are estimates and come with a degree of error.
45 Finally, we looked to assess disparity in arrest levels compared with population as a contributor to use of force compared with population. A s detailed in Figure 15, arrests among Black community members occur 5.61 times more often than arrests among white community members. For Hispanic community members, the disparity index using arrests and population was 2.09, and it was 0.30 for Asian community members compared with white community members.
Figure 15: Disparity indexes of arrests to San José population by race 03/2017 –
12/2017
01/2018 –
12/2018
01/2019 –
12/2019
01/2020 –
12/2020
01/2021 –
02/2021 Total
Compared with White
Black 6.20 5.26 5.79 5.32 5.38 5.61
Hispanic 2.10 1.94 2.17 2.22 1.99 2.09
Asian 0.31 0.28 0.30 0.33 0.29 0.30
In examining these three disparity indexes, we can say that Black and Hispanic community members are arrested at higher rates than white community members, but within those arrests, use of force levels are similar. Asian community members – on the other hand – experience fewer arrests than white c ommunity members, and even fewer experience of use of force within those arrests. It is worth nothing that the disparity index measurements do not take into account any information about the specific incidents that involved use of force, or the outcomes associated with force. The analyses in Section 5 applied more rigorous statistical analyses to better assess those differences.
Findings and recommendations
Finding 33: Eighteen percent of SJPD officers were involved in nine or more use of force events during the period analyzed.
Eighteen percent of SJPD officers, or 171 officers in total, have been involved in use of force events considerably more often than their peers over the period from March 2017 to February 2021. While certain job duties, assignments, or involvement in particular responses can provide reasonable explanation for officers being more frequently involved in use of force incidents, these outliers should be reviewed and evaluated regularly. The SJPD currently provides a list to the Chiefs on an annual basis, and high use of force officers are reviewed to determine the reasons behind the frequency of force. If no clear explanation emerges for the pattern of more frequent involvement in these incidents (e.g., assignment to riot response or tactical teams), appropriate action should be taken, to include supervisory interventions and referral for refresher training on relevant topics, including de - escalation. Processes and protocols should be in place to ensure this analysis is conducted regularly, potentially via an early intervention system, and appropriate action is taken to remediate any individual patterns of excessive escalation or use of force.
Recommendation 33: SJPD should continue to review the records and patterns of behavior of officers with high levels of use of force to understand why they are involved in use of force more frequently than typical, and if
46 necessary, refer officers for appropriate intervention, including refresher training.
Finding 34: SJPD’s pattern of types of force used generally fits the pattern of an agency that uses appropriate force escalation procedures.
As seen in Figure 9, SJPD uses lower levels of force relatively more frequently (i.e., there have been 2,913 physical contact and holds/takedowns uses of force) and higher levels of force relatively less frequently (e.g., there have been 10 firearm uses of force). This pat tern is consistent with practices wherein officers escalate only when necessary and attempt to resolve an incident using the lowest level of force necessary.
Recommendation 34: The SJPD should maintain its current practice and continue to review patterns in type and category of force used to assess relative rates.
Finding 35: The SJPD’s “other” and “unknown” perceived weapon categories represent a substantial number of events.
In the review of the reported perceived weapons on armed community members, we noted that the categories “other,” “unknown,” and “other and unknown” accounted for 868 (55 percent) of the events where the community member was perceived to be armed. The use of these categories for such a high volume of events diminishes transparency and can result in community mistrust of police use of force responses. It a lso hinders data analyses since it is unknown whether these perceived weapon types result in different officer behaviors.
Recommendation 35: The SJPD should consider additional, more detailed, categories for the use of force report item regarding perceived weapon types.
The SJPD should examine these use of force events to determine whether these weapon categories should have been included in existing categories and whether additional categories are needed to capture information from these events.
Finding 36: Black and Hispanic community members are arrested more frequently than would be predicted based on their proportion of San José’s population compared with white community members . Among those arrest ed, use of force levels are similar for Black and Hispanic community members compared with white community members.
As seen in Figure 1 2 and with the disparity index es presented in Figures 13 and 15, Black and Hispanic community members are arrested more and have force used against them more frequently than would be predicted based on their proportion of San José’s population compared with white community members. However, population is a poor baseline comparison for use of force as it does not account for potential disparities in overall contacts or arrests. Figure 1 4 shows the disparity
47 indexes using arrests as the baseline, which indicates that among those arrested, use of force levels are similar for Black and Hispanic community members compared with white community members.
The SJPD should therefore evaluate its enforcement practices to determine potential explanations for the disparities in arrests.43
Recommendation 36: The SJPD should further explore the reasons for differences in arrest and, where necessary, take remedial action.
43 Disparities in arrests should not be dismissed whole-cloth by noting disparities in crime reports. Other explanations, even within higher crime areas, should be considered. See, for example, Goff, P.A., Jackson, M.C., Di Leone, B.A.L., Culotta, M., & DiTomasso, N.A. (2014). The essence of innocence: Consequences of dehumanizing black children. Journal of Personality and Social Psychology, 106, 526–545. https://doi.org/10.1037/a0035663
48 Section 4: The Impact of COVID -19 and Social
Justice Movements on Police Reform
The assessment team also assessed what effect the COVID-19 pandemic and 2020 social justice movements for police reform had on San José policing in the context of calls for service, arrests, and use of force. For instance, as a result of the COVID-19 pandemic, the SJPD altered its operations and services to comport with the guidelines set forth b y the County of Santa Clara while maintaining a commitment to ensuring the safety of San José residents. Furthermore, the SJPD identified opportunities for continued improvements as a result of the social justice movements during the months of May and June 2020.44
To examine the impact of these events, the assessment team collected calls for service data from the SJPD open data repository that covered March 2017 to February 2021 .45 Calls for service without a police encounter were removed from the data. The se included calls with dispositions that were marked as canceled, no disposition, no response, unfounded event, no report required, and gone on arrival/unable to locate. Using these data and the collected use of force data, we conducted interrupted time series analyses on the number of calls for services, arrests made during calls for service, and the amount of use of force events. The interrupted time series provides information on the immediate change in the first month following an intervention, such as policy changes as a result of the COVID-19 pandemic, as well as the change in the post-period trend line compared with the pre- period trend. 46 Because the count of the events is aggregated to the monthly level for the analysis, individual-level characteristics are not included as covariates. However, the models (detailed in Appendix C) included a month of the year covariate to control for seasonal effects on the outcomes , as well as the amount of arrests made for the models examining the amount of use of force.
We used March 2020 as our intervention date because Santa Clara County issued its COVID-19 order to shelter in place on March 16, 2020, and the murder of George Floyd and subsequent social justice movements occurred shortly afte r. Our statistical models account for autocorrelation within each model and adjust lag accordingly when necessary.
Figure 16 presents the interrupted time series trend associated with the total amount of calls for service responded to by SJPD officers. Prior to the shelter -in-place order, the amount of calls for service per month were relatively stable, at an average of 7,409 per month in 2019. As can be visually seen in the trend, there was a substantial decrease in calls for service as a result of COVID -19 of approximately 2,300 calls for service (b(SE)= -2298.89(296.83), p<.001). This decrease stayed at a
44 Office of the City Auditor. (2020). Annual Report on City Services 2019-20. San José, CA: Office of the City Auditor. https://www.sanjoseca.gov/home/showpublisheddocument/67957/637467496715000000 45 San José, CA Police Calls for Service Open Data. https://data.sanjoseca.gov/dataset/police-calls-for-service 46 Linden, A. (2015). Conducting interrupted time-series analysis for single- and multiple-group comparisons. The Stata Journal, 15(2), 480-500.
49 similar level across 2020 and into 2021 , with an average of 5,511 calls for service each month. The slopes of the pre-COVID-19 and post-COVID-19 trend lines were equal, and both were relatively flat (b(SE)=6.12(34.97), p>.05).
Figure 16: Interrupted time series of calls for service with March 2020 break
Not surprisingly, a similar pattern was observed in the amount of arrests connected to calls for service. Prior to the shelter -in-place order, the amount of call for service arrests per month was increasing. The average arrests per month in 201 7 was 746, while the average per month in 2019 increased to 871. As can be seen in Figure 17, there was a substantial decrease in arrests connected to calls for service as a result of COVID -19 of approximately 382 arrests (b(SE)=-382.51(49.66), p<.001). Arrest levels after this drop decreased slightly over time, with an average of 544 arrests per month from April 2020 to February 2021; however, the trend was not significantly different from the arrest trend prior to COVID-19 (b(SE)=-2.88(6.58), p>.05).
50 Figure 17: Interrupted time series of arrests from calls for service with March 2020 break
While significant decreases were observed in the month following the shelter-in-place order for both the amount of ca lls for service and associated arrests, the same was not observed for use of force.
The amount of community members involved in a use of force event, as presented in Figure 18, averaged about 59 per month prior to the pandemic, and while this did drop slightly to an average of 50 per month after March 2020, the change was not statistically significant (b(se)= -4.43 (11.54), p>.05). The final model controlled for the amount of arrests each month, but we also examined the model without including the arrest control and found similar, non-significant results. The difference in the slope of trend lines pre -COVID-19 and post-COVID-19 were also found statistically similar in both models.
Appendix C also provides regression tables on interrupted time series of the amount of community members who experienced use of force by racial groupings. We examined whether the COVID -19 pandemic and 2020 social justice movements affected use of force levels for Black, white, Hispanic, and Asian individuals. Each of the models produce similar results as the model with the total amount of use of force , namely that there was no significant change in the levels after the shelter -in-place orders were implemented, nor did the trend of use of force change following March 2020.
51 Figure 18: Interrupted time series of use of force events with March 2020 break
Overall, the results from the interrupted time series indicate that calls for service and arrests declined significantly as a result of shelter -in-place orders associated with the COVID -19 pandemic that also coincided with the 2020 social justice movements for police reform after the murder of George Floyd.
However, these events did not have an impact on the amount of community members in total or by separate racial groups against which the SJPD used force.
Findings and recommendations
Finding 37: The amount of use of force events was not affected by COVID-19 and social justice movements in early 2020; whereas both the amount of calls for service and arrests were significantly lower following these events.
As noted above, both calls for service and arrests declined significantly in connection with the shelter-in-place orders associated with the COVID -19 pandemic and with the 2020 social justice movements for police reform. However, while uses of force decreased during the same time period, the decreases were not statistically significant and were not consistent with the degree of changed experienced by calls for service and arrests.
Recommendation 37: The SJPD should look further into this and identify potential reasons for the difference. Where reasons are identified, the SJPD should take remedial steps.
52 Section 5: Use of Force and Injuries by Race and
Ethnicity
To analyze the outcomes associated with use of force incidents, the assessment team used a quasi- experimental approach called propensity score matching ,47 which compares incidents that are otherwise extremely similar but differ in terms of the race of the involved community member. In simplified terms, in reviewing use of force incidents, the propensity score matching method would attempt to match two similar incidents: one involving a white community member and one involving a community member of color. We conducted separate propensity score matches for Black, Asian, and Hispanic community members using the following information to create equal groups:
• The community member’s age (0-25, 26-35, and 36+)
• The community member’s sex (male and female)
• The shift the use of force event occurred within (1st, 2nd, 3rd, and other/unknown) • The initial reason for the contact (crime report or a call for service, a pprehension, traffic or pedestrian stop, other reason)
• The amount of community members presence at the use of force event (one, two or more) • The number of officers present at the use of force event (one, two, three, four, five or more) • The number of female officers at the use of force event • The number of white officers at the use of force event • The number of Black officers at the use of force event • The number of Hispanic officers at the use of force event • The number of Asian officers at the use of force event • The number of officers of another race at the use of force event • The number of officers at the use of force event with a tenure less than 2 years • The number of officers at the use of force event with a tenure of 2 to 5 years • Whether any of the officers at the use of force event perceived the community member to be armed (yes or no)
• The degree of community member resistance (passive no compliance, active resistance, assaultive, life threatening)
47 Rosenbaum, P. B., & Rubin, D. B. (1983). The Central Role of the Propensity Score in Observational Studies for Causal Effects. Biometrika, 70(1): 41–55.
53 • The d egree of irregular behavior from community member (count of follow ing characteristics: signs of mental disability, signs of physical disability, signs of alcohol impairment, signs of drug impairment, signs of developmental disability) This procedure produced groups of white community members who matched the individuals of color based on the above list of variables. No cases were dropped from the samples as a result of not having a match , and propensity score matching with replacement was necessary only for the matching conducted with Hispanic community member s (i.e., the white comparison group was weighted appropriately to match the Hispanic group). Figures in Appendix E present the descriptive statistics of these variables and full regression models associated with each outcome . Balance between the groups was assessed using the Cohen’s d effect size and t statistics, also presented in Appendix E.
Imbalance would be exhibited by Cohen’s d in excess of +/ - 0.20 and a t in excess of +/ - 1.96. The propensity score match models produced highly equivalent white comparison groups for the Black and Asian samples. While there are some variables with minor degrees of imbalance for the Hispanic sample, we determined th at, overall, the comparison grouping with whites was well balanced in its composition.
Because the two racial groupings for each analysis are otherwise similar, absent disparate treatment or bias, we would expect to see both incidents result in the same outcome (e.g., level of force used, count of force used). Although propensity score matching cannot establish that racial bias exists with certainty, it provides stronger evidence than past techniques, such as correlational analysis or disparity indexes, alone.
These analyses did not assess the difference in whether use of force was used across racial groups (as detailed in the disparity indexes above) but instead focus on the racial differences in behaviors and outcomes present in use of force events. Specifically, we examined the following outcomes:
• The number of use of force activities used during the event • The most severe use of force activity used during the event • Whether a weapon was discharged during the event
• Whether the community member was injured to any extent during the event • The amount of different injuries sustained by the community member during the event • The most severe injury sustained by the community member during the event Figures 19 and 20 detail the descriptive statistics associated with these outcomes by different community member demographic characteristics. The amount of use of force activities was a measurement of the count of whether any of the following different use of force behaviors were present during the event:
- Discharged service weapon (firearm)
- Discharged taser probes
- Carotid restraint control hold
- Threat of a firearm (i.e, pointing their
weapon)
- Other control hold/takedown
54 - ECW
- Other physical contact (e.g., fists, feet)
- Blunt/impact weapon
- Chemical spray (e.g., OC/CS)
- Officer vehicle contact
- Impact projectile
- K-9 contact
- Other dangerous weapon
- Taser drive stun
Another use of force activity included in the force report was officer use of a knife, blade, or stabbing instrument; however, no officers reported this activity.
Across the 2,743 community members who experienced a use of force, the average count of use of force activities was 1.66 (SD=1.02), as detailed in Figure 19. Black community members had the highest average amount of use of force activities at 1.70 (SD=1.02) compared with the lowest among white community members (M=1.63, SD=1.00), but Black males (M=1.83, SD=1.09) accounted for this level more than did Black females (M=1.28, SD=0.56).
The outcome measuring the most severe level of use of force used the SJPD’s four -category use of force categories from its policy manual (SJPD Policy# L 2605.5). These categories use a combination of the above use of force activit ies, injury information, and hospital admission information to create the four categories, as detailed in Appendix D. The higher the category the more sever e the level of use of force. The average most severe use of force across all community members was 1.92 (SD=0.94), as detailed in Figure 19. The average among Asian community members was the highest at 2.02 (SD=0.98) and lowest among Hispanics (M=1.91, SD=0.94).
55 Figure 19: Use of force by demographic characteristics
Use
of
Force
Cases
Number of Use
of Force
Activities Used
Mean (SD)
Most Severe Use
of Force Used
Mean (SD)
Taser
Discharged
n (%)
Firearm
Discharged
n (%)
Weapon
Discharged
Total n (%)
Race
White 565 1.63 (1.00) 1.95 (0.92) 80 (14.16%) 3 (0.53%) 82 (14.51%) Black 385 1.70 (1.02) 1.87 (0.92) 52 (13.51%) 3 (0.78%) 54 (14.03%) Hispanic 1,526 1.66 (1.03) 1.91 (0.94) 170 (11.14%) 8 (0.52%) 176 (11.53%) Asian 201 1.66 (1.01) 2.02 (0.98) 26 (12.94%) 6 (2.99%) 30 (14.93%) Other/Unknown Race 66 1.56 (1.07) 2.15 (1.00) 7 (10.61%) 1 (1.52%) 7 (10.61%) Age
1 to 25 years old 783 1.51 (0.88) 1.81 (0.93) 65 (8.30%) 4 (0.51%) 69 (8.81%) 26 to 35 years old 957 1.72 (1.04) 2.03 (0.96) 121 (12.64%) 12 (1.25%) 128 (13.38%) 36 years or older 999 1.71 (1.10) 1.92 (0.92) 149 (14.91%) 5 (0.50%) 152 (15.22%) Age unknown 4 1.00 (0.00) 2.00 (1.41) 0 (0.00%) 0 (0.00%) 0 (0.00%) Sex
Female 498 1.30 (0.64) 1.66 (0.91) 14 (2.81%) 5 (1.00%) 18 (3.61%) Male 2,236 1.74 (1.07) 1.98 (0.94) 321 (14.36%) 16 (0.72%) 331 (14.80%) Trans 6 1.17 (0.41) 1.50 (0.84) 0 (0.00%) 0 (0.00%) 0 (0.00%) Sex unknown 3 1.00 (0.00) 2.33 (1.53) 0 (0.00%) 0 (0.00%) 0 (0.00%) Race x Sex
White male 454 1.72 (1.06) 2.02 (0.91) 77 (16.96%) 2 (0.44%) 79 (17.40%) White female 111 1.25 (0.56) 1.67 (0.91) 3 (2.70%) 1 (0.90%) 3 (2.70%) Black male 295 1.83 (1.09) 2.00 (0.92) 50 (16.95%) 3 (1.02%) 52 (17.63%) Black female 87 1.28 (0.56) 1.45 (0.77) 2 (2.30%) 0 (0.00%) 2 (2.30%) Hispanic male 1,272 1.72 (1.07) 1.95 (0.94) 163 (12.81%) 5 (0.39%) 166 (13.05%) Hispanic female 251 1.33 (0.71) 1.72 (0.95) 7 (2.79%) 3 (1.20%) 10 (3.98%) Asian male 164 1.74 (1.06) 2.09 (0.97) 24 (14.63%) 5 (3.05%) 27 (16.46%) Asian female 37 1.32 (0.63) 1.73 (0.99) 2 (5.41%) 1 (2.70%) 3 (8.11%) Other/Unknown Race male 51 1.71 (1.17) 2.25 (0.98) 7 (13.73%) 1 (1.96%) 7 (13.73%) Other/Unknown Race female 12 1.08 (0.29) 1.67 (0.89) 0 (0.00%) 0 (0.00%) 0 (0.00%) Total 2,743 1.66 (1.02) 1.92 (0.94) 335 (12.21%) 21 (0.77%) 349 (12.72%)
56 The measurement of whether a weapon was discharged during the use of force event considered if the officer discharged his or her firearm or taser, independent of whether the community member was hit. This variable was measured as Yes = 1 and No = 0. As detailed in Figure 20, the majority of these instance s of a weapon discharge were the result of a taser deployment (95.99 percent) as opposed to a firearm shooting. Across all 2,743 cases, SJPD officers discharged their firearms 21 times (0.77 percent) and deployed their tasers 335 times (12.21 percent). As such, the average weapon discharge was 12.72 percent across all cases, with relatively equal discharge rates among white, Black, and Asian community members at 14.51 percent, 14.03 percent, and 14.93 percent, respectively. Hispanic community members had the lowest rate of weapon discharges at 11.53 percent of community members involved in a use of force. Interestingly, older community members had higher rates of weapon discharges than those who were younger. By race and sex, Black males had the highest rate of weapon discharges at 17.63 percent, with white males being a close second at
17.40 percent, while all females in general had much lower rates (3.61 percent).
The outcome measuring whether the community me mber was injured corresponded to the officer’s assessment of whether any injury occurred, even if it was minor. Recorded injuries included complaint of pain , c ontusion, c oncussion, u nconsciousness, a brasion/laceration, i nternal injury , obvious disfigurement, bone fracture, stabbing wound, and gunshot wound. In total, 78.13 percent of the community members involved in a use of force event had one or more of the recorded injuries.
This amount was relatively even across the racial groupings, with 76.10 percent of Black community members having an injury and 79.36 percent of Hispanic community members having an injury. The highest amount of injuries was among Asian males (81.10 percent), closely followed by white males (81.06 percent). Additionally, community members who were older had a greater amount of injuries.
The amount of sustained injuries by the community member was a measurement of the count of whether any of the above recorded injuries occurred during or as a result of the use of force event.
The average across all community members was 1.05 (SD=0.78) types of injuries , and this number did not vary greatly across the demographic characteristics. The highest average amount of injuries was observed among white males (M=1.13, SD=0.84), while the lowest average was among Black females (M=0.72, SD= 0.62).
The ranking of these injuries to determine the most severe sustained injuries categorized the above injury types into one of four categories. The lowest ranking (coded as 1) was individuals who did not sustain an injury. Community members who complained of pain, had a contusion, or concussion were the second ranking (coded as 2). Unconsciousness, abrasions and lacerations, and internal injuries were the third ranking (coded as 3). Finally, bone fractures, stabbing wounds, and gunshot wounds were the highest ranking (coded as 4). As mentioned earlier, there were no instance of officers reporting that they used force with a knife, blade, or stabbing instrument; however, officers reported that 12 community members had a stabbing wound injury. It is possible that these people sustained these injuries from self-inflicted wounds during a shuffle with the officer, or the wound was present prior to the officer’s arrival. Overall, the average most severe injur y was 2.23 (SD=0.81) across all community members who experienced a use of force. Again, this number did not vary greatly across
57 the demographic groupings, with the highest average of most severe injuring being observed among Asian males (M=2.34, SD=0.83), and the lowest among Black females (M=1.85, SD=0.74).
Figure 20: Injuries by demographic characteristics
Use of Force
Cases
Community
Member
Injured
n (%)
Amount of
Sustained
Injuries
M (SD)
Highest
Sustained
Injury
Mean (SD)
Race
White 565 444 (78.58%) 1.07 (0.81) 2.24 (0.81)
Black 385 293 (76.10%) 1.01 (0.77) 2.16 (0.80)
Hispanic 1,526 1,211 (79.36%) 1.06 (0.78) 2.25 (0.80)
Asian 201 159 (79.10%) 1.05 (0.76) 2.28 (0.83)
Other/Unknown Race 66 36 (54.55%) 0.67 (0.71) 1.76 (0.79)
Age
1 to 25 years old 783 582 (74.33%) 0.95 (0.73) 2.14 (0.81) 26 to 35 years old 957 749 (78.27%) 1.07 (0.80) 2.24 (0.81) 36 years or older 999 810 (81.08%) 1.11 (0.80) 2.28 (0.80) Age unknown 4 2 (50.0%) 0.50 (0.58) 1.75 (0.96)
Sex
Female 498 353 (70.88%) 0.83 (0.65) 1.97 (0.74)
Male 2,236 1,784 (79.79%) 1.10 (0.80) 2.29 (0.81)
Transgender 6 4 (66.67%) 0.67 (0.52) 1.67 (0.52)
Sex unknown 3 2 (66.67%) 0.67 (0.58) 2.00 (1.00)
Race x Sex
White male 454 368 (81.06%) 1.13 (0.84) 2.32 (0.81)
White female 111 76 (68.47%) 0.81 (0.67) 1.90 (0.74)
Black male 295 235 (79.66%) 1.09 (0.79) 2.25 (0.80)
Black female 87 56 (64.37%) 0.72 (0.62) 1.85 (0.74)
Hispanic male 1,272 1,024 (80.50%) 1.10 (0.80) 2.30 (0.81) Hispanic female 251 185 (73.71%) 0.88 (0.66) 2.04 (0.75) Asian male 164 133 (81.10%) 1.12 (0.79) 2.34 (0.83)
Asian female 37 26 (70.27%) 0.76 (0.55) 2.00 (0.78)
Other/Unknown Race male 51 24 (47.06%) 0.63 (0.77) 1.71 (0.83) Other/Unknown Race female 12 10 (83.33%) 0.83 (0.39) 1.92 (0.51)
Total 2,743 2,143 (78.13%) 1.05 (0.78) 2.23 (0.81)
The differences by race groups on these outcomes were assessed using the matched groups identified from the propensity score matching and Poisson r egressions for the counts of use of force activities and different injuries the community member sustained, linear regressions for the most severe use of force activity and most severe injury sustained by the community member, and logistic regressions for whether a weapon was discharged and whether the community member was injured to any extent. All models were run using four steps. The first model, or Model A, was the base model that
58 included only the variable on the matched racial groupings. Model B added characteristics of the community member, specifically their age and sex. Model C added characteristics associated with the use of force event , specifically how the event initiated, the shift of the event, the number of community members and officers present, the number of officers by sex and race, and count of those present based on their tenure. Finally, Model D added critical characteristic s of the community member, including whether the officer perceived the community member to be armed, the resistance level, and the level of irregular behavior as determined by the officer. We included the covariates in the models in this way to better assess how specific fields of information may affect the measured outcomes. The following charts display the predicted margins of the outcomes determined fr om the full models, while the regression tables for each model are presented in Appendix E.
Figure 21 displays the predicted margins and significance levels associated with differences observed between Black community members and the matched white comparison group for each outcome. Across all outcomes and in all models detailed in Appendix E, the two groups were found statistically equal, indicating that officers treat Black and white community members similarly in use of force events. However, there are some caveats of these results that are discussed in detail further below.
Figure 21: Predicted values on main outcomes, Black vs. white community members (n = 382 per group)
1.56 1.88
0.12 0.77
1.06 2.21
1.70 1.87
0.14 0.76
1.01 2.16
0 0.5
1 1.5
2 2.5
Amount of
different use of
force activities
used
(n.s.)
Most severe use of
force used
(n.s.)
Weapon
discharged
(n.s.)
Community
member injured
to any extent
(n.s.)
Amount of
different injuries
CM sustained
(n.s.)
Most severe injury
sustained by CM
(n.s.)
Matched White Group Black
59 Similar findings were observed in the models that examined racial differences for Asian community members and the matched white community members, as detailed in Figure 22. Across all outcomes and in all models detailed in Appendix E, the two groups were found statistically equal, indicating that officers treat Asian and white community members similarly in use of force events.
Figure 22: Predicted values on main outcomes, Asian vs. white community members (n = 201 per group)
Results associated with the models that examined racial differences for community members who were Hispanic and the matched white community members are detailed in Figure 23. While no significant differences were ob served for the severity of the force, discharged weapons, injury, and amount of sustained injuries, there were significant differences between the two racial groups in the amount of use of force activities that were used and the severity of sustained injuries. The regression results indicate that Hispanic community members experience roughly 6 percent more use of force activities than their white counterparts do (IRR=1.06, p < .05). This is also shown in the predicted margins, which estimate the count of use of force activities to be 1.65 for Hispanic community members and 1.56 for white community members. The differences in the severity of injuries was also higher for Hispanic community members —with a predicted margin of 2.25 —compared with white community members, with a predicted margin of 2.19 (b(SE)= 0.06(0.03), p < .05), indicating that the injuries sustained by Hispanic community members were more severe than those sustained by white community members.
1.60 2.04
0.15 0.81
1.10 2.29
1.63 2.00
0.16 0.79
1.03 2.27
0 0.5
1 1.5
2 2.5
Amount of
different use of
force activities
used
(n.s.)
Most severe use of
force used
(n.s.)
Weapon
discharged
(n.s.)
Community
member injured
to any extent
(n.s.)
Amount of
different injuries
CM sustained
(n.s.)
Most severe
injury sustained
by CM
(n.s.)
Matched White Group Asian
60 Figure 23: Predicted values on main outcomes, Hispanic vs. white community members (n = 1,522 per group)
A reader might notice that the difference in predicted margins on the amount of use of force between Black community members and its matched white group (0.14) was larger than the difference for the same among Hispanic community members and its matched white group (0.09), but the difference for Blacks was non-significant while it was significant for Hispanics. There might be multiple reasons why this is the case. First, the white comparison groups are not the same group across these analyses.
The white comparison group for the Black group analyses used a portion of all wh ite community members who experienced use of force that most closely matched the Black individuals, whereas the white comparison group for the Hispanic analyses used all white individuals , but weighted them appropriately to match the Hispanic group of community members. As such, there may be underlying relationships that affect the statistical models.
Second, the sample size for the Black analyses was 764, split evenly across Black and white community members, while the sample size for the Hispanic analyse s was much larger. There was a total of 1, 522 Hispanic individuals who experience d use of force but only 565 white individuals.
These white individuals were weighted in the analyses to create an equal 1, 522 sample of white individuals that best matched the Hispanic group. As such, the total sample size for the Hispanic analyses was 3,044. In statistical analyses, larger samples increase the statistical power of the models and reliability of their results. 48 In fact, there can be many instances w here very small differences between groups are found significant when the sample size is large enough. Essentially – and well
48 Blalock, H. M. (1979). Social Statistics. New York City, NY: Mcgraw-Hill
1.56 1.95
0.13 0.77
1.03 2.19
1.65 1.90
0.11 0.79
1.06 2.25
0 0.5
1 1.5
2 2.5
Amount of
different use of
force activities
used
(p < .05)
Most severe use of
force used
(n.s.)
Weapon
discharged
(n.s.)
Community
member injured
to any extent
(n.s.)
Amount of
different injuries
CM sustained
(n.s.)
Most severe injury
sustained by CM
(p < .05)
Matched White Group Hispanic
61 known to statisticians – it is the size of the difference (i.e., the effect size) that is more important than the degree of signifi cance of the difference. As such, it is possible (and even likely given the upper and lower bounds of the predicted margins ; see Figure A.9 in Appendix E) that the differences observed among Black community members and the white comparison group would be f ound significant had there been more cases in the statistical models.
Findings and recommendations
Finding 38: SJPD officers treat Black and Asian community members similarly in use of force events compared with white community members.
As noted above, the data do not indicate any statistical significant difference in the treatment of either Black community members or Asian community members when compared with white community members. This was true for all main outcomes included in the model.
Recommendation 38: The SJPD should maintain its current practice for this finding.
Finding 39: SJPD officers treat Hispanic community members differently in use of force events compared with white community members.
While there were no statistical differences between Black or Asian community members when compared with white community members, the data do indicate some differences in the treatment of Hispanic com munity members when compared with whites. Hispanic community members experienced a greater amount of independent use of force activities per event than their white counterparts did. Additionally, the injuries sustained by Hispanic community members were mo re severe than those sustained by white community members. For all other assessed outcomes, there were no statistical differences found between the two groups.
Recommendation 39: The SJPD should further explore the reasons for the differences found and, where necessary, take remedial action.
62 Conclusion
Given the authority bestowed on the police by society to take persons into custody and, where necessary, use physical means to do so, police departments must be vigilant in how force is proscribed in policy, taught in training, and used during street interactions with community members. As a result of nationwide discourse on the topic, the city of San José through the Independent Police Auditor’s office, requested a review of the SJPD’s use of force practices with a focus on the overall application of force as well as potential disparities based on the demographics of force recipients. The present report includes 39 findings and 51 recommendations based on such a review, with each findin g/recommendation either validating the SJPD’s current practices or providing clear guidance to come in line with current policing standards. Although it is beyond the scope of this assessment to conduct rigid comparative analysis between San José and other agencies regarding use of force, the overall finding of this assessment is that the SJPD has several elements of use of force that are in line with best practices as well as several elements that require additional reform to be considered consistent with industry standards. Where we have found room for improvement, we have identified other agencies that exemplify best practices for the SJPD’s referral.
Several of our findings were based on SJPD’s current practices as detailed in the SJPD Duty Manual.
Of pa rticular note is our finding that SJPD’s definition of a use of force event is poor, both in its wording (defining a force event as an event in which an officer uses force) as well as in its construct (focusing on an injurious outcome rather than the actio ns of the officer). Related to this, the SJPD does not define levels of resistance and does not consistently indicate which level of resistance is necessary to justify different force options. Furthermore, while we identified areas of the Duty Manual that should be revised, we also found several critical elements that were absent from the Duty Manual and therefore absent from departmental operations. For instance, the SJPD does not have a use of force review board or any other third -party entity to audit us es of force and identify potential policy, training, or operational implications.
In addition to broader findings on use of force, this assessment focused on particular subtopics of force, including the categorization of force, use of deadly force, mass demonstrations, providing first aid, and the use of force report used by officers to document their force. Similar to our broader findings related to force, these focal areas indicate that in some areas the SJPD is in line with best practices whereas in other areas, there is room for improvement. For instance, the manner by which SJPD collects data on use of force allows for the data to be assessed at the event, subject, officer, force type, and injury levels. However, the overall reporting system is outdated , requiring manual entry and prohibiting easy review and analysis. Additionally, the SJPD uses the Santa Clara County Police Chief’s Association OII g uidelines to inform their practices after an OII event. However, the Duty Manual does not memorialize the guidelines and SJPD members are not provided guidance on who is responsible for carrying out specific tasks. In some areas, the SJPD has already performed the ground-work necessary for departmental improvements. For instance, after the 2020 social justice movements, the SJPD conducted a wide -ranging after-action review, which identified gaps in the
63 department’s crowd control policies, training, and operation. Although these gaps have been identified, many of SJPD’s own findings and recommendations have not yet been fully implemented.
We also conducted quantitative analysis on SJPD’s use of force data, examining events over a four - year span between 2017 and 2021. Overall , the data indicated that use of force levels had been relatively stable prior to the COVID-19 pandemic and the social justice movement during the summer of 2020. For instance, between 2017 and 2019, there were approximately 52 use of force events per month; however, in 2020, uses of force dropped to approximately 42 events per month. Across al l use of force events, trends emerged related to event, officer, and subject characteristics. For instance, a majority of force events resulted from the officer responding to a crime report or a call for service (as opposed to officer-initiated contacts). Additionally, for a majority of use of force events, officers perceived the suspect to be armed though the most frequent response option for type of weapon selected was an “unknown” weapon. Community members were most often described as being actively resistant (e.g., attempting to escape the control of the officer) and many were reported as showing signs of mental disability, alcohol impairment, or drug impairment.
We also assessed the data to identify racial disparities in SJPD use of force. However, unde rstanding racial disparities in law enforcement activity is difficult given the complexity of developing a baseline against which to compare groups. To highlight such issues, we provided three separate analyses, looking at racial differences across the ove rall San José population, differences within arrests, and differences in experiencing a use of force event. The results indicate that while differences in use of force are found for some racial categories relative to their proportion in the San José popula tion, these differences largely disappear when compared with the arrest statistics.
Finally, we used propensity score matching to evaluate whether differences existed between racial categories with regards to force outcomes, including the amount of force, severity of force, and injuries resulting from force. For most outcomes we reviewed, we did not find any significant differences between racial groups. However, for Hispanic community members, we found they experienced a greater amount of different use of force activities used as well as more severe injuries from the use of force when compared with the matched white group.
This report recognizes areas where SJPD operations are consistent with best practices and where the data indicate fair and equitable treatment. Alternatively, this report also identifies areas where SJPD can improve their operations and provide greater equity in treatment. As SJPD seeks to implement the recommendations from this report, we urge them to consult with officers, community memb ers, external stakeholders, and subject matter experts in order to make changes in accordance with principles of community policing. Where appropriate, we have providing guidance to the department for best and emerging practices though note that any change in departmental operations will still need to be reflective of the San José community.
64 Appendix A: Data Reviewed by CNA Team
Type Document
Data Arrests
Calls for Service
Census Demographic
Use of Force
Forms Incident Action Plans
Training Evaluation
Use of Force Reports
Policies SJPD Duty Manual
Training Active Shooter
De-escalation
Defensive Tactics
Firearms
Force Options Simulator
Officer Safety
Other Documents SJPD After Action Review
65 Appendix B: Table of Findings and
Recommendations
Finding
No. Finding Recommendation
1 Segments of the San José
community have diverse
perspectives and
experiences with use of
force.
1: Future modifications to departmental operations should reflect the diversity of perspectives and experiences of the San José community.
2 The SJPD’s use of force tools
and options that are listed
in the Duty Manual are
consistent with other
agencies.
2: The SJPD should maintain its current practice for this finding.
3 The SJPD Duty Manual and
training contain consistent
elements across all use of
force tools and options.
3: The SJPD should maintain its current practice for this finding.
4 The SJPD Duty Manual does
not define levels of
resistance and does not
consistently indicate which
level of resistance would
justify various force options.
4.1: The SJPD should better define levels of resistance.
4.2: The SJPD should state the minimum resistance level needed for each use of force option.
5 The SJPD does not have a
force review board.
5.1: The SJPD should create a force review board or unit to identify policy, training, equipment, and personnel
implications.
5.2: The SJPD should include community representatives as part of its efforts.
6 The SJPD does not have a
policy requiring officers to
attempt de-escalation.
6.1: The SJPD should elevate and emphasize affirmative duty to attempt de-escalation during encounters when time and circumstances permit.
6.2: The SJPD should better define the concept of de-
escalation.
6.3: The SJPD should revise the Duty Manual to remove the reference to use of force being a de-escalation tool.
6.4: The SJPD should engage the community when defining de-escalation and providing concrete tools in the Duty
Manual.
66 7 The SJPD has levels of force
but does not elevate force
levels if the event involves
vulnerable populations.
7.1: The SJPD should revise its Duty Manual to include the language of the President’s Task Force on 21st Century
Policing when describing vulnerable populations in all uses of force.
7.2: The SJPD should require enhanced review of use of force incidents against vulnerable populations.
8 The SJPD Duty Manual does
not require each application
of force to be evaluated
independently.
8: The SJPD should revise the Duty Manual to require that each application of force be evaluated under the totality of the circumstances independent of other force applications. 9 The SJPD Duty Manual
unnecessarily includes a
section on excited delirium.
9: The SJPD should remove the paragraph on excited
delirium from the Duty Manual.
10 The SJPD Duty Manual does
not provide sufficient clarity
in the definition of force and
bases force on injury or
complaint of pain, rather
than on the physical act that
caused it.
10: The SJPD should adopt a “physical coercion against resistance” definition of force and apply it throughout the Duty Manual.
11 The SJPD’s categorization of
force is not comprehensive.
11: The SJPD should categorize all force tools and options in the categories identified in the SJPD Duty Manual.
12 The SJPD Duty Manual
covers many conventional
sections related to ECWs.
12: The SJPD should maintain its current practice for this finding.
13 The SJPD Duty Manual does
not provide sufficient clarity
on some elements related to
ECWs.
13: The SJPD should revise the Duty Manual to provide
concrete prohibitions where appropriate or substitute a higher standard of review (i.e., when deadly force would be authorized) for such uses of ECWs.
14 The SJPD Duty Manual
allows for warning shots but
does not provide sufficient
framework around them.
14: The SJPD should reconsider the benefits on the use of warning shots; however, the SJPD should explicitly detail the situations and factors in which warning shots are allowed if it continues to allow their use.
15 The SJPD Duty Manual is
inconsistent regarding
which actions constitute
deadly force and/or require
an investigation consistent
with a deadly force event.
15: Where officers require deadly force justification, uses of force should be investigated in a manner consistent with deadly force events.
16 The SJPD Duty Manual
section related to officers’
duty to provide a lethal
force warning is vague.
16: The SJPD should revise the Duty Manual to require force warnings for all uses of force unless time and circumstances do not allow for a warning.
67 17 The SJPD Duty Manual does
not provide sufficient
guidance for officer actions
after a lethal force event.
17: The SJPD should revise the Duty Manual to provide
comprehensive guidance on post-incident actions that
should be taken after a lethal force event.
18 The SJPD Duty Manual
places the responsibility for
post-lethal force
notifications on the involved
officer.
18: After using lethal force, the involved officer should communicate the force to the communications unit, which would assume responsibility for making notifications.
19 Family and friends of
officer-involved shooting
victims expressed concern
about the quality of
communication after the
shootings and during the
investigations.
19: The SJPD should explore and gather feedback about
communication with friends and family of officer-involved shooting victims. This should include interviewing SJPD members and, where appropriate and welcomed, the friends and family of subjects in the officer-involved shootings. 20 The SJPD Duty Manual does
not provide sufficient
guidance for supervisor
actions after a lethal force
event.
20.1: The SJPD should revise the Duty Manual to identify the required post-incident tasks to be performed.
20.2: The SJPD should revise the Duty Manual to identify who is responsible for performing post-incident tasks.
20.3: The SJPD should revise the Duty Manual to include requirements and constraints on gathering public safety statements.
21 The SJPD conducted a
comprehensive after-action
report of the 2020 social
justice movements related
to the murder of George
Floyd, though many of the
resulting recommendations
have yet to be implemented.
21: The SJPD should provide a public update on the
implementation of the AAR, including a timeline for
implementation, if applicable.
22 The SJPD Duty Manual does
not provide sufficient
guidance for officers when
responding to a crowd
control event.
22.1: The SJPD should revise the Duty Manual to identify a command structure for escalating events, who is responsible for call-outs, and the criteria for such call-outs.
22.2: The SJPD should revise the Duty Manual to
memorialize the requirement for initial and ongoing
communication with social justice movement organizers.
22.3: The SJPD should memorialize the requirements for post-event debriefings between the incident commander
and command staff and what those debriefings should entail.
68 23 Community members
expressed concern
regarding SJPD’s use of
force during the 2020 social
justice movements.
23: The SJPD should recognize that community members’
perspectives of use of force during social justice movements may differ from departmental findings and incorporate that recognition when engaging the community in a review of
policies.
24 The SJPD has not provided
regular training on crowd
control to all officers.
24.1: The SJPD should provide training relevant to the findings of its 2020 AAR.
24.2: The SJPD should incorporate elements of crowd
control into other types of training.
25 The SJPD Duty Manual does
not contain sufficient
instruction related to
officers’ duty to provide
medical attention.
25.1: SJPD officers should take a risk-averse approach by summoning medical units any time there is a complaint of injury or pain and providing interim first aid in accordance with the training they have received.
25.2: The SJPD should use Section L 2610 (Providing First Aid) as a template for detailing the medical steps officers are required to take after using other force tools and options. 26 SJPD officers are not
required to prepare to
provide first aid.
26: The SJPD should revise the Duty Manual to require
officers to carry individual first aid kits and water for use after a force event.
27 The SJPD use of force report
uses an outdated reporting
system.
27: The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation.
28 The SJPD uses a use of force
template to report use of
force and has recently
released a Use of Force
Documentation Guide to aid
in completing the template.
28: The SJPD should ensure that all trainings, guides, and reporting requirements emphasize that each application of force is to be coded independently.
29 The SJPD’s use of force data
collection form allows for
each specific combination of
event, involved officer, type
of force, sustained injuries,
and involved community
member to be assessed.
29: The SJPD should maintain its current practice of
documenting use of force incidents at this level of detail. 30 The manual entry of
information into the SJPD
use of force report creates
data inconsistencies.
30: The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation.
31 Duplicate items in the SJPD
use of force report create
data inconsistencies.
31: The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation.
69 32 The lack of data quality
check allows for data entry
errors in the SJPD use of
force report.
32: The SJPD should pursue implementation of a new use of force reporting system that allows for better information entry, case tracking, review, analyses, and summary report creation.
33 Eighteen percent of SJPD
officers were involved in
nine or more use of force
events during the period
analyzed.
33: SJPD should continue to review the records and patterns of behavior of officers with high levels of use of force to understand why they are involved in use of force more
frequently than typical, and if necessary, refer officers for appropriate intervention, including refresher training. 34 SJPD’s pattern of types of
force used generally fits the
pattern of an agency that
uses appropriate force
escalation procedures.
34: The SJPD should maintain its current practice and
continue to review patterns in type and category of force used to assess relative rates.
35 The SJPD’s “other” and
“unknown” perceived
weapon categories
represent a substantial
number of events.
35: The SJPD should consider additional, more detailed, categories for the use of force report item regarding
perceived weapon types. The SJPD should examine these use of force events to determine whether these weapon
categories should have been included in existing categories and whether additional categories are needed to capture information from these events.
36 Black and Hispanic
community members are
arrested more frequently
than would be predicted
based on their proportion of
San José’s population
compared with white
community members.
Among those arrested, use
of force levels are similar for
Black and Hispanic
community members
compared with white
community members.
36: The SJPD should further explore the reasons for
differences in arrest and, where necessary, take remedial action.
37 The amount of use of force
events was not affected by
COVID-19 and social justice
movements in early 2020;
whereas both the amount of
calls for service and arrests
were significantly lower
following these events.
37: The SJPD should look further into this and identify potential reasons for the difference. Where reasons are identified, the SJPD should take remedial steps.
38 SJPD officers treat Black and
Asian community members
similarly in use of force
events compared with white
community members.
38: The SJPD should maintain its current practice for this finding.
70 39 SJPD officers treat Hispanic
community members
differently in use of force
events compared with white
community members.
39: The SJPD should further explore the reasons for the differences found and, where necessary, take remedial
action.
71 Appendix C: Interrupted Time Series Regression
Tables
Figure A.1: ITS on calls for service, March 2020 break
b (SE) 95% CI
Constant 6714.98 (251.83) *** 6202.62 | 7227.33
_t 25.13 (7.20) ** 10.49 | 39.78
_x37 -2298.89 (296.83) *** -2902.80 | -1694.98
_x_t37 6.12 (34.97) -65.02 | 77.27
February (vs. January) -370.41 (208.74) -795.11 | 54.28 March (vs. January) 336.87 (238.08) -147.51 | 821.25
April (vs. January) -65.54 (258.40) -591.27 | 460.19
May (vs. January) 183.05 (447.04) -726.46 | 1092.56
June (vs. January) -100.11 (286.94) -683.89 | 483.66
July (vs. January) 210.47 (223.18) -243.58 | 664.53
August (vs. January) 304.31 (198.48) -99.51 | 708.13
September (vs. January) 190.65 (200.41) -217.09 | 598.39 October (vs. January) 413.49 (187.88) * 31.24 | 795.73 November (vs. January) -221.93 (179.40) -586.92 | 143.07 December (vs. January) -53.59 (152.15) -363.14 | 255.97
Number of Observations 48
Chi-squared 35.37 ***
* p < .05, ** p < .01, *** p < .001
Notes: Lag = 1, Period 37 = March 2020, Months included
72 Figure A.2: ITS on arrests from calls for service, March 2020 break b (SE) 95% CI
Constant 663.45 (43.28) *** 575.39 | 751.5
_t 5.72 (0.59) *** 4.52 | 6.92
_x37 -382.51 (49.66) *** -483.55 | -281.47
_x_t37 -2.88 (6.58) -16.26 | 10.51
February (vs. January) -8.50 (28.91) -67.32 | 50.32
March (vs. January) 130.49 (67.50) -6.83 | 267.82
April (vs. January) 99.49 (48.15) * 1.54 | 197.45
May (vs. January) 113.74 (44.53) * 23.16 | 204.33
June (vs. January) 28.00 (56.51) -86.97 | 142.97
July (vs. January) 59.25 (39.39) -20.90 | 139.39
August (vs. January) 53.00 (38.80) -25.95 | 131.94
September (vs. January) 136.00 (57.31) * 19.41 | 252.59 October (vs. January) 106.50 (57.90) -11.30 | 224.30
November (vs. January) 4.00 (54.03) -105.93 | 113.93
December (vs. January) -29.25 (21.90) -73.81 | 15.31
Number of Observations 48
Chi-squared 34.36 ***
* p < .05, ** p < .01, *** p < .001
Notes: Lag = 5, Period 37 = March 2020, Months included
Figure A.3: ITS on use of force, March 2020 break
b (SE) 95% CI
Constant 45.99 (20.08) * 5.08 | 86.90
_t 0.07 (0.23) -0.39 | 0.53
_x37 -4.43 (11.54) -27.93 | 19.08
_x_t37 -0.85 (0.82) -2.52 | 0.81
Calls for Service Arrests 0.00 (0.03) -0.06 | 0.07
February (vs. January) -0.84 (4.88) -10.78 | 9.10
March (vs. January) 2.16 (9.84) -17.89 | 22.21
April (vs. January) 4.18 (4.96) -5.92 | 14.28
May (vs. January) 15.98 (8.49) -1.31 | 33.27
June (vs. January) 14.28 (5.53) * 3.02 | 25.53
July (vs. January) 14.49 (6.84) * 0.55 | 28.43
August (vs. January) 13.64 (6.44) * 0.53 | 26.75
September (vs. January) 9.85 (6.22) -2.81 | 22.51
October (vs. January) 9.36 (9.16) -9.30 | 28.03
November (vs. January) 6.99 (2.17) ** 2.57 | 11.41
December (vs. January) 2.78 (6.98) -11.45 | 17.00
Number of Observations 48
Chi-squared 5.62 ***
* p < .05, ** p < .01, *** p < .001
Notes: Lag = 3, Period 37 = March 2020, Months included
73 Figure A.4: ITS on use of force with white San José residents, March 2020 Break b (SE) 95% CI
Constant 10.56 (7.39) -4.49 | 25.62
_t -0.01 (0.11) -0.24 | 0.22
_x37 -2.04 (4.93) -12.08 | 8.00
_x_t37 -0.24 (0.30) -0.86 | 0.37
Calls for Service Arrests 0.00 (0.01) -0.02 | 0.02
February (vs. January) 0.82 (1.82) -2.88 | 4.53
March (vs. January) 1.01 (3.37) -5.85 | 7.88
April (vs. January) -1.41 (2.86) -7.24 | 4.42
May (vs. January) 3.66 (2.90) -2.26 | 9.58
June (vs. January) 3.50 (2.82) -2.24 | 9.23
July (vs. January) 6.81 (4.89) -3.14 | 16.77
August (vs. January) 4.13 (1.90) * 0.25 | 8.01
September (vs. January) 2.69 (2.62) -2.64 | 8.03
October (vs. January) 2.77 (2.16) -1.63 | 7.16
November (vs. January) 1.36 (2.99) -4.74 | 7.45
December (vs. January) -0.07 (2.02) -4.18 | 4.05
Number of Observations 48
Chi-squared 2.96 **
* p < .05, ** p < .01, *** p < .001
Notes: Lag = 1, Period 37 = March 2020, Months included
Figure A.5: ITS on use of force with Black San José residents, March 2020 Break b (SE) 95% CI
Constant 10.74 (2.78) ** 5.08 | 16.40
_t -0.04 (0.03) -0.11 | 0.03
_x37 -2.10 (1.64) -5.43 | 1.24
_x_t37 0.18 (0.12) -0.07 | 0.43
Calls for Service Arrests -0.00 (0.00) -0.01 | 0.01
February (vs. January) -2.51 (1.60) -5.76 | 0.74
March (vs. January) -1.83 (1.99) -5.89 | 2.22
April (vs. January) -0.88 (2.67) -6.32 | 4.57
May (vs. January) -0.60 (1.92) -4.52 | 3.31
June (vs. January) 0.27 (2.18) -4.17 | 4.72
July (vs. January) 0.07 (1.73) -3.44 | 3.59
August (vs. January) 1.57 (1.83) -2.16 | 5.29
September (vs. January) 1.69 (2.45) -3.30 | 6.67
October (vs. January) -1.85 (2.27) -6.47 | 2.76
November (vs. January) 0.50 (1.93) -3.44 | 4.44
December (vs. January) -2.29 (1.96) -6.28 | 1.69
Number of Observations 48
Chi-squared 7.37 ***
* p < .05, ** p < .01, *** p < .001
Notes: Lag = 5, Period 37 = March 2020, Months included
74 Figure A.6: ITS on use of force with Hispanic San José residents, March 2020 break b (SE) 95% CI
Constant 19.97 (12.61) -5.73 | 45.66
_t 0.03 (0.14) -0.27 | 0.32
_x37 3.58 (8.09) -12.91 | 20.06
_x_t37 -0.61 (0.46) -1.56 | 0.33
Calls for Service Arrests 0.01 (0.02) -0.03 | 0.05
February (vs. January) 0.41 (2.69) -5.06 | 5.89
March (vs. January) 1.42 (7.09) -13.02 | 15.87
April (vs. January) 2.81 (4.52) -6.39 | 12.01
May (vs. January) 7.75 (5.29) -3.03 | 18.52
June (vs. January) 5.43 (4.23) -3.18 | 14.04
July (vs. January) 4.44 (3.75) -3.20 | 12.09
August (vs. January) 4.08 (4.38) -4.85 | 13.02
September (vs. January) 3.58 (4.89) -6.38 | 13.55
October (vs. January) 5.71 (6.94) -8.42 | 19.83
November (vs. January) 2.06 (2.65) -3.34 | 7.45
December (vs. January) 4.21 (3.96) -3.86 | 12.29
Number of Observations 48
Chi-squared 3.16 **
* p < .05, ** p < .01, *** p < .001
Notes: Lag = 3, Period 37 = March 2020, Months included
Figure A.7: ITS on use of force with Asian San José residents, March 2020 break b (SE) 95% CI
Constant -1.39 (5.36) -12.30 | 9.52
_t 0.05 (0.07) -0.10 | 0.20
_x37 -3.23 (3.00) -9.34 | 2.87
_x_t37 0.16 (0.21) -0.26 | 0.59
Calls for Service Arrests 0.00 (0.01) -0.01 | 0.02
February (vs. January) -0.33 (1.27) -2.91 | 2.25
March (vs. January) 1.58 (1.48) -1.43 | 4.58
April (vs. January) 2.85 (1.87) -0.96 | 6.65
May (vs. January) 2.17 (1.83) -1.56 | 5.90
June (vs. January) 2.43 (1.39) -0.41 | 5.26
July (vs. January) 2.18 (2.09) -2.08 | 6.43
August (vs. January) 3.34 (2.36) -1.46 | 8.14
September (vs. January) 1.12 (1.04) -0.99 | 3.23
October (vs. January) 1.63 (1.88) -2.20 | 5.46
November (vs. January) 2.46 (1.63) -0.86 | 5.78
December (vs. January) 0.74 (1.27) -1.84 | 3.32
Number of Observations 48
Chi-squared 5.97 ***
* p < .05, ** p < .01, *** p < .001
Notes: Lag = 4, Period 37 = March 2020, Months included
75 Appendix D: Coding for Use of Force Severity
Location of Force
Head
Neck/
throat
Front
upper
torso/
chest
Rear
upper
torso/
back
Front
lower
torso/
abdomen
Rear
lower
torso/
back
Front
below
waist/
groin area
Rear
below
waist/
buttocks
Arms/
hands
Front
legs/ feet Rear legs
Type of Force
Other control hold/takedown 1 1 1 1 1 1 1 1 1 1 1
Other physical contact (fists, feet, etc.) 2 2 1 1 1 1 1 1 1 1 1 Chemical spray (e.g. OC/CS) 2 2 2 2 2 2 2 2 2 2 2
Electronic control device 2 2 2 2 2 2 2 2 2 2 2
Taser drive stun 2 2 2 2 2 2 2 2 2 2 2
Taser probes deployed 2 2 2 2 2 2 2 2 2 2 2
Taser probes deployed (miss) 2 2 2 2 2 2 2 2 2 2 2
Blunt/impact weapon 3 3 2 2 2 2 2 2 2 2 2
Other dangerous weapon 3 3 2 2 2 2 2 2 2 2 2
Carotid restraint control hold 3 3 3 3 3 3 3 3 3 3 3
Impact projectile 3 3 3 3 3 3 3 3 3 3 3
K-9 contact 3 3 3 3 3 3 3 3 3 3 3
Officer vehicle contact 4 4 4 4 4 4 4 4 4 4 4
Threat of firearm 4 4 4 4 4 4 4 4 4 4 4
Discharge of firearm (miss) 4 4 4 4 4 4 4 4 4 4 4
Discharge of firearm (hit) 4 4 4 4 4 4 4 4 4 4 4
Other Applied Use of Force Levels
Force resulting in bone fracture 3
Force resulting in suspect's loss of consciousness 3
Hospital admission as a direct result of the force 3
76 Appendix E: Group Balance and Regression
Tables
Black compared with white tables
Figure A.8: Post-propensity score matching group balance, Black vs. white White Black t d
N 382 382
Community Member Demographic Characteristics
Community member age - 0 to 25 0.23 (0.02) 0.25 (0.02) -0.42 -0.03 Community member age - 26 to 35 0.36 (0.02) 0.36 (0.02) -0.08 -0.01 Community member age - 36+ 0.41 (0.03) 0.39 (0.03) 0.44 0.03 Community member sex - Female 0.23 (0.02) 0.23 (0.02) 0.00 0.00 Event Characteristics
Shift - 1st 0.27 (0.02) 0.26 (0.02) 0.24 0.02
Shift - 2nd 0.44 (0.03) 0.43 (0.03) 0.36 0.03
Shift - 3rd 0.20 (0.02) 0.19 (0.02) 0.18 0.01
Shift - Other/Unknown 0.09 (0.01) 0.12 (0.02) -1.18 -0.09 Initial contact reason - Crime / CFS 0.79 (0.02) 0.76 (0.02) 0.86 0.06 Initial contact reason - Apprehension 0.03 (0.01) 0.04 (0.01) -0.83 -0.06 Initial contact reason - Traffic/Pedestrian stop 0.16 (0.02) 0.16 (0.02) -0.10 -0.01 Initial contact reason - Other Reason 0.03 (0.01) 0.04 (0.01) -0.95 -0.07 Number of community members - Two or more 0.04 (0.01) 0.06 (0.01) -1.48 -0.11 Ordinal of number of officers present at scene 2.09 (0.06) 2.14 (0.06) -0.64 -0.05 Count of Officers w/ UoF - Age - 18 to 30 a 0.71 (0.04) 0.73 (0.04) -0.35 -0.03 Count of Officers w/ UoF - Age - 31 to 40 a 0.56 (0.04) 0.63 (0.04) -1.26 -0.09 Count of Officers w/ UoF - Age - 41 years or older a 0.47 (0.04) 0.42 (0.04) 0.92 0.07 Count of Officers w/ UoF - Sex - Female 0.17 (0.02) 0.18 (0.02) -0.24 -0.02 Count of Officers w/ UoF - Sex - Male a 1.57 (0.05) 1.60 (0.05) -0.42 -0.03 Count of Officers w/ UoF - Race - White 0.85 (0.04) 0.86 (0.04) -0.09 -0.01 Count of Officers w/ UoF - Race - Black 0.07 (0.01) 0.08 (0.02) -0.75 -0.05 Count of Officers w/ UoF - Race - Hispanic 0.49 (0.03) 0.47 (0.04) 0.42 0.03 Count of Officers w/ UoF - Race - Asian 0.27 (0.03) 0.31 (0.03) -1.02 -0.07 Count of Officers w/ UoF - Race - Other 0.05 (0.01) 0.05 (0.01) 0.16 0.01 Count of Officers w/ UoF - Dress - Uniform a 1.35 (0.05) 1.34 (0.05) 0.11 0.01 Count of Officers w/ UoF - Dress - Plain Clothes a 0.06 (0.02) 0.10 (0.02) -1.47 -0.11 Count of Officers w/ UoF - Dress - Utility a 0.32 (0.03) 0.33 (0.03) -0.19 -0.01 Count of Officers w/ UoF - Tenure - 1 year or less 0.42 (0.03) 0.43 (0.04) -0.11 -0.01
77 Count of Officers w/ UoF - Tenure - 2 to 5 years 0.62 (0.04) 0.64 (0.04) -0.36 -0.03 Count of Officers w/ UoF - Tenure - 6 years or more a 0.70 (0.04) 0.71 (0.04) -0.17 -0.01 Count of Officers w/ UoF - Assignment - Patrol a 1.54 (0.06) 1.51 (0.06) 0.40 0.03 Count of Officers w/ UoF - Assignment - Other a 0.19 (0.03) 0.26 (0.03) -1.67 -0.12 Community Member Demeanor Characteristics
Did officer perceive the community member to be armed? 0.50 (0.03) 0.52 (0.03) -0.58 -0.04 Degree of community member resistance 3.07 (0.03) 3.04 (0.04) 0.58 0.04 Degree of irregular behavior from community member 2.14 (0.05) 2.10 (0.05) 0.56 0.04 Outcomes (not used to balance groups)
Amount of Different Use of Force Activities Used 1.56 (0.05) 1.70 (0.05) -1.97 * -0.14 Taser Discharged 0.12 (0.02) 0.14 (0.02) -0.76 -0.05
Firearm Discharged 0.00 (0.00) 0.01 (0.00) -1.74 -0.13 Weapon Discharged 0.12 (0.02) 0.14 (0.02) -0.97 -0.07
Community member injured to any extent 0.77 (0.02) 0.76 (0.02) 0.43 0.03 Amount of different injuries community member sustained 1.05 (0.04) 1.01 (0.04) 0.77 0.06 Type of force: Carotid restraint control hold 0.00 (0.00) 0.00 (0.00) -1.00 -0.07 Type of force: Threat of firearm 0.01 (0.00) 0.02 (0.01) -1.75 -0.13 Type of force: Other control hold/takedown 0.76 (0.02) 0.74 (0.02) 0.92 0.07 Type of force: Electronic control device 0.13 (0.02) 0.15 (0.02) -0.83 -0.06 Type of force: Other physical contact (fists, feet, etc.) 0.27 (0.02) 0.35 (0.02) -2.35 * -0.17 Type of force: Discharge of firearm (miss) 0.00 (0.00) 0.00 (0.00) -- -- Type of force: Blunt/impact weapon 0.09 (0.01) 0.09 (0.01) 0.00 0.00 Type of force: Discharge of firearm (hit) 0.00 (0.00) 0.00 (0.00) -- -- Type of force: Chemical spray (e.g. OC/CS) 0.03 (0.01) 0.05 (0.01) -1.54 -0.11 Type of force: Officer vehicle contact 0.00 (0.00) 0.00 (0.00) -- -- Type of force: Impact projectile 0.07 (0.01) 0.05 (0.01) 0.77 0.06 Type of force: K-9 contact 0.04 (0.01) 0.03 (0.01) 0.77 0.06 Type of force: Other dangerous weapon 0.00 (0.00) 0.00 (0.00) -- -- Type of force: Taser drive stun 0.03 (0.01) 0.05 (0.01) -1.33 -0.10 Type of force: Taser probes deployed 0.10 (0.02) 0.12 (0.02) -1.03 -0.07 Type of force: Taser probes deployed (miss) 0.03 (0.01) 0.02 (0.01) 0.95 0.07 * p < .05, ** p < .01, *** p < .001
a Variable not included in propensity score matching
78 Figure A.9: Amount of different use of force activities used, Black vs. white San José community members Model A Model B Model C Model D
IRR 95% CI IRR 95% CI IRR 95% CI IRR 95% CI
Constant 1.56 *** 1.44 | 1.69 1.74 *** 1.56 | 1.94 1.21 * 1.02 | 1.45 0.60 ** 0.42 | 0.84 Black (vs. white) 1.09 0.97 | 1.22 1.09 0.97 | 1.22 1.08 0.97 | 1.21 1.09 0.97 | 1.22 CM Age 0 to 25 (vs. 26 to 35) 0.91 0.78 | 1.06 0.93 0.80 | 1.09 0.98 0.84 | 1.14 CM Age 36 and older (vs. 26 to 35) 0.96 0.84 | 1.09 0.96 0.85 | 1.10 0.98 0.87 | 1.12 CM is Female (vs. male) 0.72 *** 0.62 | 0.83 0.76 *** 0.65 | 0.88 0.78 ** 0.67 | 0.91 Event Initiated as Apprehension (vs. CFS/Crime) 1.00 0.70 | 1.43 1.13 0.79 | 1.62 Event Initiated as Stop (vs. CFS/Crime) 0.99 0.84 | 1.17 1.05 0.89 | 1.24 Event Initiated as Other (vs. CFS/Crime) 0.83 0.61 | 1.14 0.87 0.63 | 1.19 Event occur during 1st Shift (vs. 2nd Shift) 0.97 0.84 | 1.11 0.99 0.86 | 1.14 Event occur during 3rd Shift (vs. 2nd Shift) 0.98 0.83 | 1.15 0.99 0.84 | 1.16 Event occur during Unknown Shift (vs. 2nd Shift) 0.90 0.73 | 1.12 0.89 0.72 | 1.10 Two or more Community Members at Event 1.03 0.79 | 1.34 1.08 0.83 | 1.41 Number of Officers at Event 1.02 0.95 | 1.08 1.00 0.94 | 1.07 Number of female officers 0.95 0.84 | 1.08 0.97 0.85 | 1.10 Number of white officers 1.19 *** 1.09 | 1.30 1.16 ** 1.06 | 1.27 Number of Black officers 1.34 ** 1.11 | 1.62 1.27 * 1.05 | 1.54 Number of Hispanic officers 1.24 *** 1.13 | 1.36 1.21 *** 1.10 | 1.33 Number of Asian officers 1.17 * 1.04 | 1.33 1.14 * 1.01 | 1.29 Number of Other Race officers 1.26 0.99 | 1.58 1.16 0.91 | 1.46 Number of officers with 1 year or less tenure 0.98 0.89 | 1.08 0.98 0.89 | 1.07 Number of officers with 2 to 5 years tenure 1.00 0.92 | 1.08 0.98 0.91 | 1.06 Officer perceived the community member to be armed 1.17 * 1.04 | 1.31 Level of community member resistance 1.21 *** 1.11 | 1.33 Level of irregular behavior from community member 1.03 0.97 | 1.10
Number of Observations 764 764 764 764
Chi-squared 2.25 26.27 *** 87.20 *** 112.37 ***
Pseudo R2 .00 .01 .04 .05
Pearson Chi-Squared goodness-of-fit 442.58 401.17 313.16 275.65 * p < .05, ** p < .01, *** p < .001
Poisson Regression, Incident Rate Ratios (IRR) reported with 95% Confidence Intervals
79 Figure A.10: Most severe use of force activity used, Black vs. white San José community members Model A Model B Model C Model D
b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI Constant 1.87 (0.05) *** 1.78 | 1.97 2.02 (0.07) *** 1.89 | 2.15 1.89 (0.11) *** 1.68 | 2.10 1.27 (0.17) *** 0.94 | 1.59 Black (vs. white) -0.00 (0.07) -0.13 | 0.13 -0.00 (0.06) -0.13 | 0.13 -0.01 (0.06) -0.13 | 0.12 -0.00 (0.06) -0.12 | 0.12 CM Age 0 to 25 (vs. 26 to 35) -0.11 (0.09) -0.27 | 0.06 -0.12 (0.08) -0.28 | 0.05 -0.04 (0.08) -0.20 | 0.12 CM Age 36 and older (vs. 26 to 35) -0.04 (0.08) -0.19 | 0.11 -0.06 (0.07) -0.20 | 0.09 -0.05 (0.07) -0.19 | 0.10 CM is Female (vs. male) -0.45 (0.07) *** -0.59 | -0.30 -0.41 (0.07) *** -0.55 | -0.27 -0.34 (0.07) *** -0.49 | -0.20 Event Initiated as Apprehension (vs. CFS/Crime) -0.05 (0.19) -0.43 | 0.33 0.08 (0.18) -0.27 | 0.44 Event Initiated as Stop (vs. CFS/Crime) -0.45 (0.08) *** -0.61 | -0.30 -0.39 (0.08) *** -0.55 | -0.23 Event Initiated as Other (vs. CFS/Crime) -0.38 (0.16) * -0.70 | -0.07 -0.32 (0.16) -0.65 | -0.00 Event occur during 1st Shift (vs. 2nd Shift) -0.09 (0.08) -0.25 | 0.06 -0.08 (0.08) -0.23 | 0.08 Event occur during 3rd Shift (vs. 2nd Shift) 0.08 (0.09) -0.10 | 0.26 0.10 (0.09) -0.07 | 0.28 Event occur during Unknown Shift (vs. 2nd Shift) -0.11 (0.11) -0.32 | 0.11 -0.13 (0.11) -0.35 | 0.08 Two or more Community Members at Event 0.17 (0.14) -0.11 | 0.45 0.20 (0.14) -0.08 | 0.48 Number of Officers at Event -0.08 (0.04) * -0.15 | -0.01 -0.09 (0.03) ** -0.16 | -0.02 Number of female officers 0.02 (0.08) -0.13 | 0.18 0.05 (0.08) -0.10 | 0.20 Number of white officers 0.22 (0.06) *** 0.11 | 0.33 0.17 (0.06) ** 0.06 | 0.29 Number of Black officers 0.47 (0.12) *** 0.24 | 0.70 0.37 (0.11) ** 0.14 | 0.59 Number of Hispanic officers 0.21 (0.06) ** 0.09 | 0.33 0.17 (0.06) ** 0.05 | 0.29 Number of Asian officers 0.25 (0.07) ** 0.10 | 0.39 0.19 (0.07) * 0.05 | 0.34 Number of Other Race officers 0.35 (0.16) * 0.04 | 0.65 0.24 (0.15) -0.06 | 0.54 Number of officers with 1 year or less tenure 0.01 (0.06) -0.11 | 0.13 0.00 (0.06) -0.12 | 0.11 Number of officers with 2 to 5 years tenure -0.05 (0.05) -0.14 | 0.05 -0.06 (0.05) -0.15 | 0.03 Officer perceived the community member to be armed 0.33 (0.06) *** 0.20 | 0.46 Level of community member resistance 0.10 (0.04) * 0.02 | 0.18 Level of irregular behavior from community member 0.10 (0.04) ** 0.03 | 0.17
Number of Observations 764 764 764 764
Chi-squared 0.00 10.33 *** 6.87 *** 9.36 ***
R2 .00 .05 .13 .17
Root MSE 0.91 0.89 0.86 0.84
* p < .05, ** p < .01, *** p < .001
Linear Regression, Unstandardized Beta Values and Robust Standard Errors reported with 95% Confidence Intervals
80 Figure A.11: Weapon discharged, Black vs. white San José community members Model A Model B Model C Model D
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Constant 0.13 *** 0.10 | 0.18 0.16 *** 0.11 | 0.25 0.21 *** 0.11 | 0.42 0.01 *** 0.00 | 0.04 Black (vs. white) 1.23 0.81 | 1.88 1.24 0.81 | 1.91 1.32 0.84 | 2.05 1.32 0.84 | 2.08 CM Age 0 to 25 (vs. 26 to 35) 1.22 0.69 | 2.14 1.29 0.72 | 2.32 1.57 0.84 | 2.91 CM Age 36 and older (vs. 26 to 35) 1.03 0.63 | 1.70 0.97 0.58 | 1.62 1.02 0.61 | 1.73 CM is Female (vs. male) 0.12 *** 0.04 | 0.33 0.12 *** 0.04 | 0.33 0.13 *** 0.05 | 0.37 Event Initiated as Apprehension (vs. CFS/Crime) 0.37 0.05 | 2.98 0.58 0.07 | 4.89 Event Initiated as Stop (vs. CFS/Crime) 0.77 0.40 | 1.48 0.94 0.47 | 1.88 Event Initiated as Other (vs. CFS/Crime) 0.42 0.09 | 1.87 0.47 0.10 | 2.17 Event occur during 1st Shift (vs. 2nd Shift) 1.04 0.61 | 1.76 1.17 0.68 | 2.02 Event occur during 3rd Shift (vs. 2nd Shift) 1.26 0.68 | 2.30 1.36 0.72 | 2.57 Event occur during Unknown Shift (vs. 2nd Shift) 0.23 * 0.07 | 0.81 0.21 * 0.06 | 0.73 Two or more Community Members at Event 0.72 0.20 | 2.57 0.84 0.23 | 3.08 Number of Officers at Event 0.73 * 0.54 | 0.99 0.65 ** 0.47 | 0.89 Number of female officers 0.73 0.41 | 1.30 0.86 0.47 | 1.56 Number of white officers 1.32 0.90 | 1.94 1.18 0.79 | 1.75 Number of Black officers 2.90 ** 1.42 | 5.89 2.18 * 1.03 | 4.60 Number of Hispanic officers 1.68 * 1.12 | 2.54 1.56 * 1.02 | 2.40 Number of Asian officers 1.41 0.85 | 2.36 1.21 0.71 | 2.05 Number of Other Race officers 0.81 0.27 | 2.43 0.56 0.18 | 1.71 Number of officers with 1 year or less tenure 0.76 0.51 | 1.13 0.73 0.48 | 1.10 Number of officers with 2 to 5 years tenure 0.92 0.67 | 1.28 0.87 0.61 | 1.22 Officer perceived the community member to be armed 2.71 *** 1.63 | 4.50 Level of community member resistance 2.27 *** 1.49 | 3.45 Level of irregular behavior from community member 1.18 0.91 | 1.52
Number of Observations 764 764 764 764
Chi-squared 0.94 31.78 *** 60.85 *** 92.59 ***
Pseudo R2 .00 .05 .10 .16
Pearson Chi-Squared goodness-of-fit 0.00 3.35 617.05 653.68 * p < .05, ** p < .01, *** p < .001
Logistic Regression, Odds Ratios (OR) reported with 95% Confidence Intervals
81 Figure A.12: Community member injured to any extent, Black vs. white San José community members Model A Model B Model C Model D
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Constant 3.44 *** 2.71 | 4.38 3.58 *** 2.55 | 5.02 1.61 0.89 | 2.92 0.27 * 0.10 | 0.77 Black (vs. white) 0.93 0.66 | 1.30 0.93 0.66 | 1.31 0.92 0.65 | 1.32 0.93 0.65 | 1.33 CM Age 0 to 25 (vs. 26 to 35) 1.16 0.75 | 1.81 1.25 0.78 | 2.00 1.49 0.92 | 2.41 CM Age 36 and older (vs. 26 to 35) 1.27 0.86 | 1.87 1.19 0.79 | 1.79 1.24 0.82 | 1.89 CM is Female (vs. male) 0.52 ** 0.36 | 0.76 0.56 ** 0.38 | 0.83 0.61 * 0.40 | 0.92 Event Initiated as Apprehension (vs. CFS/Crime) 1.60 0.54 | 4.70 2.36 0.77 | 7.23 Event Initiated as Stop (vs. CFS/Crime) 0.43 *** 0.27 | 0.69 0.51 ** 0.31 | 0.84 Event Initiated as Other (vs. CFS/Crime) 0.62 0.25 | 1.55 0.72 0.28 | 1.83 Event occur during 1st Shift (vs. 2nd Shift) 0.88 0.57 | 1.36 0.89 0.57 | 1.39 Event occur during 3rd Shift (vs. 2nd Shift) 1.33 0.78 | 2.26 1.39 0.81 | 2.41 Event occur during Unknown Shift (vs. 2nd Shift) 0.78 0.43 | 1.44 0.76 0.41 | 1.43 Two or more Community Members at Event 0.65 0.30 | 1.40 0.73 0.33 | 1.64 Number of Officers at Event 1.09 0.87 | 1.36 1.04 0.83 | 1.30 Number of female officers 0.81 0.50 | 1.29 0.86 0.53 | 1.39 Number of white officers 1.43 0.99 | 2.07 1.27 0.87 | 1.85 Number of Black officers 3.54 ** 1.37 | 9.15 2.78 * 1.05 | 7.37 Number of Hispanic officers 1.55 * 1.05 | 2.31 1.35 0.91 | 2.02 Number of Asian officers 1.44 0.89 | 2.33 1.29 0.79 | 2.12 Number of Other Race officers 2.99 0.98 | 9.09 2.11 0.68 | 6.52 Number of officers with 1 year or less tenure 0.97 0.69 | 1.38 0.92 0.64 | 1.31 Number of officers with 2 to 5 years tenure 1.36 * 1.00 | 1.84 1.27 0.93 | 1.74 Officer perceived the community member to be armed 1.88 ** 1.28 | 2.76 Level of community member resistance 1.4 * 1.07 | 1.84 Level of irregular behavior from community member 1.4 ** 1.11 | 1.76
Number of Observations 764 764 764 764
Chi-squared 0.18 12.66 * 64.93 *** 91.45 ***
Pseudo R2 .00 .02 .08 .11
Pearson Chi-Squared goodness-of-fit 0.00 7.64 687.87 737.24 * p < .05, ** p < .01, *** p < .001
Logistic Regression, Odds Ratios (OR) reported with 95% Confidence Intervals
82 Figure A.13: Amount of different injuries community members sustained, Black vs. white San José community members Model A Model B Model C Model D
IRR 95% CI IRR 95% CI IRR 95% CI IRR 95% CI
Constant 1.05 0.96 | 1.16 1.15 * 1.00 | 1.32 0.89 0.72 | 1.12 0.42 *** 0.27 | 0.65 Black (vs. white) 0.96 0.83 | 1.10 0.96 0.83 | 1.10 0.95 0.83 | 1.09 0.95 0.83 | 1.10 CM Age 0 to 25 (vs. 26 to 35) 0.89 0.74 | 1.08 0.92 0.75 | 1.11 0.97 0.79 | 1.17 CM Age 36 and older (vs. 26 to 35) 1.04 0.89 | 1.21 1.03 0.87 | 1.20 1.05 0.89 | 1.23 CM is Female (vs. male) 0.68 *** 0.56 | 0.82 0.72 ** 0.60 | 0.88 0.76 ** 0.63 | 0.92 Event Initiated as Apprehension (vs. CFS/Crime) 1.18 0.79 | 1.77 1.35 0.89 | 2.04 Event Initiated as Stop (vs. CFS/Crime) 0.83 0.67 | 1.02 0.87 0.70 | 1.09 Event Initiated as Other (vs. CFS/Crime) 0.87 0.6 0| 1.27 0.91 0.63 | 1.33 Event occur during 1st Shift (vs. 2nd Shift) 0.95 0.80 | 1.14 0.97 0.81 | 1.16 Event occur during 3rd Shift (vs. 2nd Shift) 1.04 0.86 | 1.27 1.06 0.87 | 1.30 Event occur during Unknown Shift (vs. 2nd Shift) 1.05 0.81 | 1.35 1.02 0.79 | 1.32 Two or more Community Members at Event 0.86 0.60 | 1.24 0.90 0.63 | 1.29 Number of Officers at Event 1.00 0.91 | 1.08 0.98 0.90 | 1.07 Number of female officers 0.94 0.80 | 1.10 0.96 0.82 | 1.12 Number of white officers 1.09 0.97 | 1.23 1.06 0.94 | 1.19 Number of Black officers 1.34 * 1.06 | 1.70 1.26 0.99 | 1.59 Number of Hispanic officers 1.14 * 1.00 | 1.29 1.11 0.98 | 1.26 Number of Asian officers 1.13 0.96 | 1.32 1.08 0.93 | 1.27 Number of Other Race officers 1.26 0.95 | 1.67 1.15 0.86 | 1.53 Number of officers with 1 year or less tenure 1.03 0.91 | 1.16 1.02 0.91 | 1.16 Number of officers with 2 to 5 years tenure 1.10 1.00 | 1.21 1.08 0.98 | 1.20 Officer perceived the community member to be armed 1.24 ** 1.07 | 1.44 Level of community member resistance 1.22 ** 1.09 | 1.37 Level of irregular behavior from community member 1.03 0.95 | 1.12
Number of Observations 764 764 764 764
Chi-squared 0.37 21.35 *** 53.42 *** 73.87 ***
Pseudo R2 .00 .01 .03 .04
Pearson Chi-Squared goodness-of-fit 473.71 444.12 416.73 397.76 * p < .05, ** p < .01, *** p < .001
Poisson Regression, Incident Rate Ratios (IRR) reported with 95% Confidence Intervals
83 Figure A.14: Most severe sustained injury, Black vs. white San José community members Model A Model B Model C Model D
b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI Constant 2.21 (0.04) *** 2.13 | 2.29 2.27 (0.06) *** 2.15 | 2.39 2.04 (0.09) *** 1.85 | 2.22 1.38 (0.16) *** 1.06 | 1.70 Black (vs. white) -0.05 (0.06) -0.17 | 0.07 -0.05 (0.06) -0.16 | 0.06 -0.05 (0.06) -0.16 | 0.06 -0.05 (0.06) -0.16 | 0.06 CM Age 0 to 25 (vs. 26 to 35) 0.03 (0.08) -0.12 | 0.18 0.05 (0.08) -0.10 | 0.20 0.11 (0.07) -0.04 | 0.25 CM Age 36 and older (vs. 26 to 35) 0.06 (0.07) -0.07 | 0.19 0.04 (0.07) -0.09 | 0.18 0.06 (0.07) -0.07 | 0.19 CM is Female (vs. male) -0.41 (0.06) *** -0.53 | -0.28 -0.36 (0.06) *** -0.49 | -0.23 -0.32 (0.07) *** -0.45 | -0.19 Event Initiated as Apprehension (vs. CFS/Crime) 0.13 (0.16) -0.19 | 0.46 0.27 (0.17) -0.06 | 0.60 Event Initiated as Stop (vs. CFS/Crime) -0.19 (0.09) * -0.36 | -0.02 -0.13 (0.09) -0.30 | 0.04 Event Initiated as Other (vs. CFS/Crime) -0.13 (0.16) -0.44 | 0.19 -0.08 (0.16) -0.39 | 0.24 Event occur during 1st Shift (vs. 2nd Shift) -0.05 (0.07) -0.18 | 0.09 -0.03 (0.07) -0.17 | 0.11 Event occur during 3rd Shift (vs. 2nd Shift) 0.04 (0.08) -0.11 | 0.20 0.06 (0.08) -0.09 | 0.21 Event occur during Unknown Shift (vs. 2nd Shift) -0.06 (0.11) -0.28 | 0.16 -0.07 (0.12) -0.30 | 0.16 Two or more Community Members at Event -0.16 (0.15) -0.45 | 0.13 -0.11 (0.15) -0.40 | 0.18 Number of Officers at Event -0.02 (0.03) -0.08 | 0.04 -0.04 (0.03) -0.10 | 0.02 Number of female officers -0.08 (0.06) -0.21 | 0.04 -0.06 (0.06) -0.18 | 0.06 Number of white officers 0.17 (0.05) ** 0.07 | 0.26 0.13 (0.05) ** 0.04 | 0.23 Number of Black officers 0.30 (0.10) ** 0.10 | 0.50 0.22 (0.10) * 0.02 | 0.43 Number of Hispanic officers 0.19 (0.06) ** 0.08 | 0.30 0.16 (0.05) ** 0.05 | 0.26 Number of Asian officers 0.21 (0.06) ** 0.08 | 0.33 0.17 (0.06) ** 0.04 | 0.29 Number of Other Race officers 0.18 (0.10) -0.02 | 0.39 0.09 (0.10) -0.11 | 0.29 Number of officers with 1 year or less tenure -0.03 (0.05) -0.14 | 0.07 -0.05 (0.05) -0.15 | 0.06 Number of officers with 2 to 5 years tenure 0.02 (0.04) -0.06 | 0.11 0.01 (0.04) -0.07 | 0.09 Officer perceived the community member to be armed 0.20 (0.06) ** 0.09 | 0.32 Level of community member resistance 0.15 (0.05) ** 0.06 | 0.24 Level of irregular behavior from community member 0.08 (0.03) * 0.01 | 0.14
Number of Observations 764 764 764 764
Chi-squared 0.72 10.34 *** 5.03 *** 6.34 ***
R2 .00 .05 .10 .13
Root MSE 0.81 0.79 0.78 0.77
* p < .05, ** p < .01, *** p < .001
Linear Regression, Unstandardized Beta Values and Robust Standard Errors reported with 95% Confidence Intervals
84 Asian compared with white tables
Figure A.15: Post-propensity score matching group balance, Asian vs. white White Asian t d
N 201 201
Community Member Demographic Characteristics
Community member age - 0 to 25 0.18 (0.03) 0.20 (0.03) -0.51 -0.05 Community member age - 26 to 35 0.34 (0.03) 0.36 (0.03) -0.42 -0.04 Community member age - 36+ 0.48 (0.04) 0.44 (0.04) 0.80 0.08 Community member sex - Female 0.22 (0.03) 0.18 (0.03) 0.87 0.09 Event Characteristics
Shift - 1st 0.31 (0.03) 0.27 (0.03) 0.99 0.10
Shift - 2nd 0.43 (0.04) 0.45 (0.04) -0.40 -0.04
Shift - 3rd 0.20 (0.03) 0.22 (0.03) -0.37 -0.04
Shift - Other/Unknown 0.05 (0.02) 0.06 (0.02) -0.44 -0.04 Initial contact reason - Crime / CFS 0.83 (0.03) 0.81 (0.03) 0.39 0.04 Initial contact reason - Apprehension 0.02 (0.01) 0.02 (0.01) 0.00 0.00 Initial contact reason - Traffic/Pedestrian stop 0.13 (0.02) 0.14 (0.02) -0.43 -0.04 Initial contact reason - Other Reason 0.02 (0.01) 0.02 (0.01) 0.00 0.00 Number of community members - Two or more 0.07 (0.02) 0.06 (0.02) 0.40 0.04 Ordinal of number of officers present at scene 2.03 (0.08) 2.15 (0.09) -1.01 -0.10 Count of Officers w/ UoF - Age - 18 to 30 a 0.72 (0.06) 0.85 (0.06) -1.47 -0.15 Count of Officers w/ UoF - Age - 31 to 40 a 0.57 (0.05) 0.60 (0.05) -0.32 -0.03 Count of Officers w/ UoF - Age - 41 years or older a 0.46 (0.05) 0.41 (0.05) 0.64 0.06 Count of Officers w/ UoF - Sex - Female 0.15 (0.03) 0.16 (0.03) -0.23 -0.02 Count of Officers w/ UoF - Sex - Male a 1.59 (0.07) 1.69 (0.08) -0.90 -0.09 Count of Officers w/ UoF - Race - White 0.89 (0.06) 0.94 (0.07) -0.58 -0.06 Count of Officers w/ UoF - Race - Black 0.02 (0.01) 0.04 (0.01) -1.17 -0.12 Count of Officers w/ UoF - Race - Hispanic 0.42 (0.04) 0.47 (0.05) -0.71 -0.07 Count of Officers w/ UoF - Race - Asian 0.28 (0.03) 0.27 (0.04) 0.10 0.01 Count of Officers w/ UoF - Race - Other 0.14 (0.03) 0.13 (0.02) 0.14 0.01 Count of Officers w/ UoF - Dress - Uniform a 1.39 (0.07) 1.42 (0.07) -0.33 -0.03 Count of Officers w/ UoF - Dress - Plain Clothes a 0.04 (0.02) 0.04 (0.02) -0.21 -0.02 Count of Officers w/ UoF - Dress - Utility a 0.32 (0.04) 0.39 (0.05) -1.08 -0.11 Count of Officers w/ UoF - Tenure - 1 year or less 0.49 (0.06) 0.54 (0.05) -0.65 -0.07 Count of Officers w/ UoF - Tenure - 2 to 5 years 0.60 (0.06) 0.61 (0.06) -0.18 -0.02 Count of Officers w/ UoF - Tenure - 6 years or more a 0.66 (0.06) 0.71 (0.06) -0.52 -0.05 Count of Officers w/ UoF - Assignment - Patrol a 1.55 (0.08) 1.61 (0.09) -0.50 -0.05 Count of Officers w/ UoF - Assignment - Other a 0.20 (0.04) 0.25 (0.05) -0.83 -0.08
85 Community Member Demeanor Characteristics
Did officer perceive the community member to be armed? 0.62 (0.03) 0.57 (0.03) 0.91 0.09 Degree of community member resistance 3.03 (0.05) 3.11 (0.05) -1.08 -0.11 Degree of irregular behavior from community member 2.28 (0.07) 2.27 (0.06) 0.11 0.01 Outcomes (not used to balance groups)
Amount of Different Use of Force Activities Used 1.57 (0.06) 1.66 (0.07) -1.01 -0.10 Taser Discharged 0.13 (0.02) 0.13 (0.02) 0.15 0.01
Firearm Discharged 0.00 (0.00) 0.03 (0.01) -1.91 -0.19
Weapon Discharged 0.14 (0.02) 0.15 (0.03) -0.28 -0.03
Community member injured to any extent 0.81 (0.03) 0.79 (0.03) 0.50 0.05 Amount of different injuries community member sustained 1.08 (0.05) 1.05 (0.05) 0.39 0.04 Type of force: Carotid restraint control hold 0.00 (0.00) 0.00 (0.00) -1.00 -0.10 Type of force: Threat of firearm 0.00 (0.00) 0.02 (0.01) -2.26 * -0.23 Type of force: Other control hold/takedown 0.70 (0.03) 0.70 (0.03) 0.00 0.00 Type of force: Electronic control device 0.13 (0.02) 0.14 (0.02) -0.14 -0.01 Type of force: Other physical contact (fists, feet, etc.) 0.31 (0.03) 0.33 (0.03) -0.43 -0.04 Type of force: Discharge of firearm (miss) 0.00 (0.00) 0.00 (0.00) -1.00 -0.10 Type of force: Blunt/impact weapon 0.07 (0.02) 0.04 (0.01) 1.07 0.11 Type of force: Discharge of firearm (hit) 0.00 (0.00) 0.02 (0.01) -2.02 * -0.20 Type of force: Chemical spray (e.g. OC/CS) 0.03 (0.01) 0.06 (0.02) -1.38 -0.14 Type of force: Officer vehicle contact 0.00 (0.00) 0.00 (0.00) -- -- Type of force: Impact projectile 0.07 (0.02) 0.07 (0.02) 0.19 0.02 Type of force: K-9 contact 0.07 (0.02) 0.06 (0.02) 0.40 0.04 Type of force: Other dangerous weapon 0.00 (0.00) 0.00 (0.00) -1.00 -0.10 Type of force: Taser drive stun 0.03 (0.01) 0.04 (0.01) -0.54 -0.05 Type of force: Taser probes deployed 0.10 (0.02) 0.11 (0.02) -0.32 -0.03 Type of force: Taser probes deployed (miss) 0.03 (0.01) 0.03 (0.01) -0.28 -0.03 * p < .05, ** p < .01, *** p < .001
a Variable not included in propensity score matching
86 Figure A.16: Amount of different use of force activities used, Asian vs. white San José community members Model A Model B Model C Model D
IRR 95% CI IRR 95% CI IRR 95% CI IRR 95% CI
Constant 1.57 *** 1.40 | 1.75 1.73 *** 1.48 | 2.01 1.19 0.93 | 1.51 0.68 0.42 | 1.10 Asian (vs. white) 1.06 0.91 | 1.24 1.05 0.90 | 1.22 1.02 0.87 | 1.19 1.02 0.87 | 1.19 CM Age 0 to 25 (vs. 26 to 35) 0.99 0.80 | 1.23 0.93 0.75 | 1.17 0.97 0.77 | 1.21 CM Age 36 and older (vs. 26 to 35) 0.93 0.78 | 1.10 0.96 0.80 | 1.15 0.95 0.79 | 1.14 CM is Female (vs. male) 0.74 ** 0.59 | 0.91 0.78 * 0.63 | 0.97 0.79 * 0.64 | 0.99 Event Initiated as Apprehension (vs. CFS/Crime) 1.34 0.82 | 2.18 1.33 0.81 | 2.19 Event Initiated as Stop (vs. CFS/Crime) 1.00 0.79 | 1.27 1.04 0.82 | 1.33 Event Initiated as Other (vs. CFS/Crime) 0.95 0.56 | 1.60 0.98 0.58 | 1.66 Event occur during 1st Shift (vs. 2nd Shift) 0.99 0.82 | 1.19 1.00 0.83 | 1.21 Event occur during 3rd Shift (vs. 2nd Shift) 0.98 0.79 | 1.22 0.99 0.80 | 1.24 Event occur during Unknown Shift (vs. 2nd Shift) 0.90 0.62 | 1.32 0.93 0.63 | 1.37 Two or more Community Members at Event 1.04 0.75 | 1.44 1.03 0.74 | 1.44 Number of Officers at Event 1.05 0.96 | 1.15 1.04 0.95 | 1.14 Number of female officers 1.10 0.92 | 1.30 1.08 0.91 | 1.29 Number of white officers 1.12 1.00 | 1.26 1.11 0.98 | 1.24 Number of Black officers 0.90 0.56 | 1.46 0.90 0.55 | 1.46 Number of Hispanic officers 1.18 * 1.02 | 1.37 1.16 1.00 | 1.35 Number of Asian officers 1.17 0.99 | 1.39 1.16 0.98 | 1.38 Number of Other Race officers 1.12 0.89 | 1.41 1.06 0.84 | 1.35 Number of officers with 1 year or less tenure 0.92 0.81 | 1.05 0.91 0.80 | 1.04 Number of officers with 2 to 5 years tenure 1.05 0.94 | 1.16 1.04 0.93 | 1.15 Officer perceived the community member to be armed 1.11 0.93 | 1.32 Level of community member resistance 1.16 * 1.02 | 1.32 Level of irregular behavior from community member 1.04 0.94 | 1.15
Number of Observations 402 402 402 402
Chi-squared 0.56 10.32 * 49.70 *** 57.25 ***
Pseudo R2 .00 .01 .04 .05
Pearson Chi-Squared goodness-of-fit 217.46 201.42 144.13 133.48 * p < .05, ** p < .01, *** p < .001
Poisson Regression, Incident Rate Ratios (IRR) reported with 95% Confidence Intervals
87 Figure A.17: Most severe use of force activity used, Asian vs. white San José community members Model A Model B Model C Model D
b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI Constant 2.01 (0.06) *** 1.89 | 2.14 2.20 (0.09) *** 2.02 | 2.38 1.92 (0.14) *** 1.65 | 2.20 1.30 (0.25) *** 0.81 | 1.78 Asian (vs. white) 0.00 (0.09) -0.18 | 0.19 -0.02 (0.09) -0.20 | 0.16 -0.03 (0.09) -0.20 | 0.14 -0.04 (0.09) -0.21 | 0.13 CM Age 0 to 25 (vs. 26 to 35) 0.07 (0.13) -0.19 | 0.34 0.03 (0.13) -0.22 | 0.28 0.08 (0.12) -0.16 | 0.32 CM Age 36 and older (vs. 26 to 35) -0.24 (0.10) * -0.45 | -0.04 -0.22 (0.1) * -0.42 | -0.02 -0.21 (0.10) * -0.41 | -0.00 CM is Female (vs. male) -0.36 (0.12) ** -0.60 | -0.13 -0.33 (0.12) ** -0.56 | -0.11 -0.32 (0.12) ** -0.55 | -0.09 Event Initiated as Apprehension (vs. CFS/Crime) 0.49 (0.28) -0.07 | 1.05 0.44 (0.29) -0.13 | 1.02 Event Initiated as Stop (vs. CFS/Crime) -0.48 (0.12) *** -0.72 | -0.23 -0.44 (0.13) ** -0.69 | -0.19 Event Initiated as Other (vs. CFS/Crime) -0.64 (0.26) * -1.16 | -0.13 -0.58 (0.26) * -1.09 | -0.08 Event occur during 1st Shift (vs. 2nd Shift) -0.03 (0.11) -0.24 | 0.19 -0.02 (0.11) -0.23 | 0.19 Event occur during 3rd Shift (vs. 2nd Shift) 0.06 (0.12) -0.17 | 0.30 0.06 (0.12) -0.17 | 0.30 Event occur during Unknown Shift (vs. 2nd Shift) 0.19 (0.22) -0.25 | 0.62 0.21 (0.22) -0.21 | 0.64 Two or more Community Members at Event -0.28 (0.16) -0.59 | 0.03 -0.30 (0.15) -0.60 | -0.00 Number of Officers at Event 0.05 (0.05) -0.06 | 0.15 0.03 (0.05) -0.08 | 0.13 Number of female officers -0.01 (0.10) -0.21 | 0.19 -0.02 (0.10) -0.23 | 0.18 Number of white officers 0.26 (0.07) *** 0.12 | 0.39 0.26 (0.07) *** 0.12 | 0.39 Number of Black officers -0.31 (0.37) -1.03 | 0.41 -0.30 (0.34) -0.97 | 0.36 Number of Hispanic officers 0.29 (0.08) ** 0.12 | 0.45 0.27 (0.08) ** 0.11 | 0.44 Number of Asian officers 0.11 (0.12) -0.13 | 0.35 0.10 (0.12) -0.13 | 0.34 Number of Other Race officers 0.23 (0.12) -0.01 | 0.47 0.20 (0.13) -0.05 | 0.45 Number of officers with 1 year or less tenure -0.21 (0.08) ** -0.36 | -0.06 -0.23 (0.08) ** -0.38 | -0.08 Number of officers with 2 to 5 years tenure -0.1 (0.07) -0.23 | 0.03 -0.12 (0.06) -0.25 | 0.01 Officer perceived the community member to be armed 0.11 (0.09) -0.08 | 0.29 Level of community member resistance 0.21 (0.06) ** 0.08 | 0.34 Level of irregular behavior from community member -0.01 (0.05) -0.12 | 0.10
Number of Observations 402 402 402 402
Chi-squared 0.00 4.58 ** 6.12 *** 6.29
R2 .00 .05 .18 .21
Root MSE 0.94 0.92 0.87 0.86
* p < .05, ** p < .01, *** p < .001
Linear Regression, Unstandardized Beta Values and Robust Standard Errors reported with 95% Confidence Intervals
88 Figure A.18: Weapon discharged, Asian vs. white San José community members Model A Model B Model C Model D
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Constant 0.16 *** 0.11 | 0.24 0.26 *** 0.15 | 0.43 0.23 ** 0.10 | 0.56 0.02 *** 0.00 | 0.13 Asian (vs. white) 1.08 0.62 | 1.89 1.05 0.59 | 1.84 1.11 0.62 | 1.98 1.08 0.59 | 1.97 CM Age 0 to 25 (vs. 26 to 35) 0.54 0.23 | 1.27 0.48 0.19 | 1.18 0.50 0.20 | 1.28 CM Age 36 and older (vs. 26 to 35) 0.70 0.38 | 1.29 0.72 0.38 | 1.38 0.67 0.34 | 1.33 CM is Female (vs. male) 0.26 * 0.09 | 0.75 0.25 * 0.09 | 0.73 0.29 * 0.10 | 0.88 Event Initiated as Apprehension (vs. CFS/Crime) 2.61 0.38 | 18.14 3.50 0.50 | 24.38 Event Initiated as Stop (vs. CFS/Crime) 0.89 0.36 | 2.22 1.11 0.43 | 2.88 Event Initiated as Other (vs. CFS/Crime) 0.91 0.10 | 8.12 0.80 0.09 | 7.39 Event occur during 1st Shift (vs. 2nd Shift) 0.88 0.44 | 1.76 0.97 0.48 | 1.98 Event occur during 3rd Shift (vs. 2nd Shift) 0.99 0.45 | 2.17 1.17 0.52 | 2.65 Event occur during Unknown Shift (vs. 2nd Shift) -- -- -- -- Two or more Community Members at Event 0.96 0.24 | 3.87 0.95 0.22 | 4.02 Number of Officers at Event 1.16 0.83 | 1.62 1.08 0.76 | 1.52 Number of female officers 1.15 0.57 | 2.33 1.17 0.56 | 2.43 Number of white officers 1.04 0.66 | 1.63 0.99 0.61 | 1.60 Number of Black officers 1.03 0.19 | 5.50 0.96 0.17 | 5.36 Number of Hispanic officers 0.87 0.46 | 1.65 0.82 0.42 | 1.59 Number of Asian officers 1.63 0.87 | 3.06 1.51 0.78 | 2.92 Number of Other Race officers 0.61 0.24 | 1.57 0.45 0.16 | 1.23 Number of officers with 1 year or less tenure 0.71 0.41 | 1.23 0.68 0.38 | 1.20 Number of officers with 2 to 5 years tenure 0.91 0.60 | 1.38 0.84 0.54 | 1.31 Officer perceived the community member to be armed 2.89 ** 1.41 | 5.89 Level of community member resistance 1.82 * 1.07 | 3.08 Level of irregular behavior from community member 1.12 0.77 | 1.64
Number of Observations 402 402 380 380
Chi-squared 0.08 11.58 * 20.90 35.82 *
Pseudo R2 .00 .03 .06 .11
Pearson Chi-Squared goodness-of-fit 0.00 9.70 322.83 437.12 ** * p < .05, ** p < .01, *** p < .001
Logistic Regression, Odds Ratios (OR) reported with 95% Confidence Intervals
89 Figure A.19: Community member injured to any extent, Asian vs. white San José community members Model A Model B Model C Model D
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Constant 4.29 *** 3.01 | 6.11 4.37 *** 2.65 | 7.20 1.71 0.72 | 4.07 0.32 0.08 | 1.38 Asian (vs. white) 0.88 0.54 | 1.44 0.87 0.53 | 1.43 0.82 0.49 | 1.38 0.82 0.48 | 1.40 CM Age 0 to 25 (vs. 26 to 35) 1.03 0.52 | 2.04 1.19 0.55 | 2.56 1.43 0.65 | 3.16 CM Age 36 and older (vs. 26 to 35) 1.21 0.70 | 2.10 1.22 0.68 | 2.20 1.27 0.69 | 2.32 CM is Female (vs. male) 0.63 0.35 | 1.12 0.70 0.38 | 1.30 0.77 0.41 | 1.45 Event Initiated as Apprehension (vs. CFS/Crime) 2.05 0.19 | 22.31 1.93 0.18 | 20.75 Event Initiated as Stop (vs. CFS/Crime) 0.75 0.36 | 1.56 0.86 0.41 | 1.84 Event Initiated as Other (vs. CFS/Crime) 0.23 * 0.06 | 0.92 0.25 0.06 | 1.08 Event occur during 1st Shift (vs. 2nd Shift) 1.12 0.59 | 2.11 1.12 0.58 | 2.14 Event occur during 3rd Shift (vs. 2nd Shift) 1.30 0.63 | 2.67 1.37 0.65 | 2.89 Event occur during Unknown Shift (vs. 2nd Shift) 0.80 0.26 | 2.42 0.80 0.25 | 2.56 Two or more Community Members at Event 0.27 * 0.10 | 0.75 0.27 * 0.09 | 0.76 Number of Officers at Event 1.10 0.80 | 1.51 1.05 0.76 | 1.47 Number of female officers 0.65 0.33 | 1.28 0.64 0.32 | 1.26 Number of white officers 2.04 * 1.13 | 3.70 1.89 * 1.03 | 3.46 Number of Black officers 2.82 0.31 | 25.58 2.71 0.29 | 25.53 Number of Hispanic officers 2.56 * 1.25 | 5.23 2.23 * 1.08 | 4.63 Number of Asian officers 1.38 0.67 | 2.86 1.31 0.63 | 2.74 Number of Other Race officers 1.91 0.75 | 4.85 1.46 0.56 | 3.80 Number of officers with 1 year or less tenure 0.75 0.43 | 1.28 0.73 0.42 | 1.26 Number of officers with 2 to 5 years tenure 0.89 0.55 | 1.45 0.83 0.51 | 1.35 Officer perceived the community member to be armed 1.78 * 1.01 | 3.14 Level of community member resistance 1.55 * 1.06 | 2.26 Level of irregular behavior from community member 1.11 0.78 | 1.57
Number of Observations 402 402 402 402
Chi-squared 0.25 3.04 33.98 * 44.13 **
Pseudo R2 .00 .01 .08 .11
Pearson Chi-Squared goodness-of-fit 0.00 3.22 346.84 377.69 * p < .05, ** p < .01, *** p < .001
Logistic Regression, Odds Ratios (OR) reported with 95% Confidence Intervals
90 Figure A.20: Amount of different injuries community member sustained, Asian vs. white San José community members Model A Model B Model C Model D
IRR 95% CI IRR 95% CI IRR 95% CI IRR 95% CI
Constant 1.08 0.95 | 1.23 1.17 0.97 | 1.41 0.87 0.65 | 1.17 0.45 ** 0.25 | 0.82 Asian (vs. white) 0.97 0.80 | 1.18 0.97 0.80 | 1.17 0.94 0.78 | 1.14 0.94 0.77 | 1.14 CM Age 0 to 25 (vs. 26 to 35) 0.88 0.67 | 1.16 0.87 0.65 | 1.16 0.90 0.68 | 1.21 CM Age 36 and older (vs. 26 to 35) 0.99 0.80 | 1.22 1.01 0.81 | 1.25 0.99 0.79 | 1.24 CM is Female (vs. male) 0.76 * 0.59 | 0.99 0.81 0.62 | 1.06 0.83 0.63 | 1.08 Event Initiated as Apprehension (vs. CFS/Crime) 1.18 0.62 | 2.26 1.19 0.62 | 2.29 Event Initiated as Stop (vs. CFS/Crime) 1.04 0.78 | 1.38 1.10 0.82 | 1.47 Event Initiated as Other (vs. CFS/Crime) 0.66 0.31 | 1.42 0.69 0.32 | 1.48 Event occur during 1st Shift (vs. 2nd Shift) 0.99 0.79 | 1.25 1.01 0.80 | 1.27 Event occur during 3rd Shift (vs. 2nd Shift) 0.97 0.74 | 1.26 0.99 0.76 | 1.29 Event occur during Unknown Shift (vs. 2nd Shift) 0.83 0.51 | 1.36 0.86 0.52 | 1.41 Two or more Community Members at Event 0.79 0.50 | 1.24 0.79 0.50 | 1.25 Number of Officers at Event 1.03 0.92 | 1.16 1.01 0.90 | 1.14 Number of female officers 0.96 0.77 | 1.20 0.94 0.75 | 1.17 Number of white officers 1.16 * 1.01 | 1.34 1.14 0.98 | 1.32 Number of Black officers 0.96 0.55 | 1.70 0.96 0.54 | 1.68 Number of Hispanic officers 1.18 0.98 | 1.42 1.15 0.96 | 1.39 Number of Asian officers 1.06 0.85 | 1.32 1.05 0.84 | 1.32 Number of Other Race officers 1.12 0.85 | 1.49 1.06 0.79 | 1.42 Number of officers with 1 year or less tenure 0.97 0.83 | 1.13 0.96 0.82 | 1.12 Number of officers with 2 to 5 years tenure 1.02 0.90 | 1.17 1.01 0.89 | 1.16 Officer perceived the community member to be armed 1.13 0.91 | 1.39 Level of community member resistance 1.19 * 1.01 | 1.39 Level of irregular behavior from community member 1.06 0.93 | 1.20
Number of Observations 402 402 402 402
Chi-squared 0.08 5.64 29.05 35.77 *
Pseudo R2 .00 .01 .03 .04
Pearson Chi-Squared goodness-of-fit 221.85 214.03 192.60 186.92 * p < .05, ** p < .01, *** p < .001
Poisson Regression, Incident Rate Ratios (IRR) reported with 95% Confidence Intervals
91 Figure A.21: Most severe sustained injury, Asian vs. white San José community members Model A Model B Model C Model D
b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI Constant 2.27 (0.06) *** 2.16 | 2.38 2.38 (0.08) *** 2.22 | 2.55 2.17 (0.13) *** 1.92 | 2.41 1.58 (0.25) *** 1.09 | 2.08 Asian (vs. white) 0.00 (0.08) -0.16 | 0.16 -0.01 (0.08) -0.16 | 0.15 -0.02 (0.08) -0.18 | 0.13 -0.02 (0.08) -0.18 | 0.13 CM Age 0 to 25 (vs. 26 to 35) -0.13 (0.11) -0.34 | 0.09 -0.08 (0.11) -0.29 | 0.14 -0.03 (0.11) -0.24 | 0.19 CM Age 36 and older (vs. 26 to 35) -0.03 (0.09) -0.21 | 0.15 -0.02 (0.09) -0.20 | 0.17 -0.01 (0.09) -0.19 | 0.18 CM is Female (vs. male) -0.33 (0.10) ** -0.52 | -0.14 -0.29 (0.10) ** -0.48 | -0.09 -0.25 (0.10) * -0.45 | -0.06 Event Initiated as Apprehension (vs. CFS/Crime) 0.07 (0.26) -0.44 | 0.59 0.06 (0.27) -0.47 | 0.59 Event Initiated as Stop (vs. CFS/Crime) -0.01 (0.12) -0.25 | 0.23 0.02 (0.12) -0.22 | 0.27 Event Initiated as Other (vs. CFS/Crime) -0.57 (0.24) * -1.04 | -0.10 -0.53 (0.23) * -0.97 | -0.09 Event occur during 1st Shift (vs. 2nd Shift) 0.01 (0.10) -0.18 | 0.21 0.02 (0.10) -0.17 | 0.22 Event occur during 3rd Shift (vs. 2nd Shift) 0.02 (0.10) -0.18 | 0.22 0.04 (0.10) -0.16 | 0.24 Event occur during Unknown Shift (vs. 2nd Shift) -0.14 (0.21) -0.55 | 0.27 -0.14 (0.2) -0.53 | 0.25 Two or more Community Members at Event -0.54 (0.15) *** -0.83 | -0.24 -0.55 (0.14) *** -0.84 | -0.27 Number of Officers at Event 0.04 (0.05) -0.05 | 0.13 0.02 (0.05) -0.07 | 0.11 Number of female officers -0.16 (0.09) -0.33 | 0.02 -0.16 (0.09) -0.34 | 0.01 Number of white officers 0.20 (0.06) ** 0.07 | 0.32 0.18 (0.06) ** 0.06 | 0.31 Number of Black officers 0.06 (0.26) -0.46 | 0.58 0.06 (0.24) -0.41 | 0.53 Number of Hispanic officers 0.14 (0.08) -0.02 | 0.30 0.11 (0.08) -0.04 | 0.27 Number of Asian officers 0.08 (0.10) -0.11 | 0.28 0.07 (0.10) -0.12 | 0.26 Number of Other Race officers 0.14 (0.12) -0.09 | 0.37 0.08 (0.12) -0.16 | 0.32 Number of officers with 1 year or less tenure -0.09 (0.07) -0.23 | 0.04 -0.10 (0.07) -0.23 | 0.03 Number of officers with 2 to 5 years tenure -0.08 (0.06) -0.19 | 0.04 -0.09 (0.06) -0.20 | 0.02 Officer perceived the community member to be armed 0.18 (0.08) * 0.02 | 0.35 Level of community member resistance 0.17 (0.07) * 0.03 | 0.30 Level of irregular behavior from community member 0.01 (0.05) -0.10 | 0.11
Number of Observations 402 402 402 402
Chi-squared 0.00 3.41 ** 3.56 *** 3.91 ***
R2 .00 .03 .12 .14
Root MSE 0.82 0.81 0.79 0.78
* p < .05, ** p < .01, *** p < .001
Linear Regression, Unstandardized Beta Values and Robust Standard Errors reported with 95% Confidence Intervals
92 Hispanic compared with white tables
Figure A.22: Post-propensity score matching group balance, Hispanic vs. white White Hispanic t d
N
428 -
Unweighted
1,522 -
Weighted
1,522
Community Member Demographic Characteristics
Community member age - 0 to 25 0.32 (0.01) 0.35 (0.01) -1.54 -0.06 Community member age - 26 to 35 0.37 (0.01) 0.35 (0.01) 1.13 0.04 Community member age - 36+ 0.31 (0.01) 0.31 (0.01) 0.39 0.01 Community member sex - Female 0.18 (0.01) 0.16 (0.01) 0.77 0.03 Event Characteristics
Shift - 1st 0.20 (0.01) 0.21 (0.01) -0.90 -0.03
Shift - 2nd 0.45 (0.01) 0.43 (0.01) 1.13 0.04
Shift - 3rd 0.27 (0.01) 0.26 (0.01) 0.45 0.02
Shift - Other/Unknown 0.09 (0.01) 0.10 (0.01) -1.35 -0.05 Initial contact reason - Crime / CFS 0.75 (0.01) 0.76 (0.01) -1.22 -0.04 Initial contact reason - Apprehension 0.04 (0.00) 0.04 (0.01) -0.28 -0.01 Initial contact reason - Traffic/Pedestrian stop 0.16 (0.01) 0.16 (0.01) 0.50 0.02 Initial contact reason - Other Reason 0.05 (0.01) 0.04 (0.00) 1.89 0.07 Number of community members - Two or more 0.08 (0.01) 0.10 (0.01) -1.71 -0.06 Ordinal of number of officers present at scene 2.00 (0.03) 2.12 (0.03) -2.67 ** -0.10 Count of Officers w/ UoF - Age - 18 to 30 a 0.73 (0.02) 0.74 (0.02) -0.10 0.00 Count of Officers w/ UoF - Age - 31 to 40 a 0.56 (0.02) 0.61 (0.02) -1.94 -0.07 Count of Officers w/ UoF - Age - 41 years or older a 0.41 (0.02) 0.38 (0.02) 1.29 0.05 Count of Officers w/ UoF - Sex - Female 0.08 (0.01) 0.12 (0.01) -3.06 ** -0.11 Count of Officers w/ UoF - Sex - Male a 1.62 (0.02) 1.61 (0.03) 0.38 0.01 Count of Officers w/ UoF - Race - White 0.78 (0.02) 0.77 (0.02) 0.26 0.01 Count of Officers w/ UoF - Race - Black 0.03 (0.00) 0.04 (0.00) -1.55 -0.06 Count of Officers w/ UoF - Race - Hispanic 0.56 (0.02) 0.54 (0.02) 0.95 0.03 Count of Officers w/ UoF - Race - Asian 0.27 (0.01) 0.29 (0.01) -0.88 -0.03 Count of Officers w/ UoF - Race - Other 0.07 (0.01) 0.10 (0.01) -2.55 * -0.09 Count of Officers w/ UoF - Dress - Uniform a 1.30 (0.02) 1.29 (0.03) 0.15 0.01 Count of Officers w/ UoF - Dress - Plain Clothes a 0.08 (0.01) 0.06 (0.01) 2.22 * 0.08 Count of Officers w/ UoF - Dress - Utility a 0.32 (0.01) 0.38 (0.02) -2.52 * -0.09 Count of Officers w/ UoF - Tenure - 1 year or less 0.42 (0.02) 0.41 (0.02) 0.31 0.01 Count of Officers w/ UoF - Tenure - 2 to 5 years 0.63 (0.02) 0.62 (0.02) 0.22 0.01 Count of Officers w/ UoF - Tenure - 6 years or more a 0.66 (0.02) 0.70 (0.02) -1.24 -0.05
93 0.66 (0.02) 0.7 (0.02) -1.24 -0.05
Count of Officers w/ UoF - Assignment - Patrol a 1.50 (0.03) 1.44 (0.03) 1.58 0.06 Count of Officers w/ UoF - Assignment - Other a 0.20 (0.01) 0.29 (0.02) -3.72 *** -0.13 Community Member Demeanor Characteristics
Did officer perceive the community member to be armed? 0.55 (0.01) 0.54 (0.01) 0.55 0.02 Degree of community member resistance 3.03 (0.02) 3.02 (0.02) 0.57 0.02 Degree of irregular behavior from community member 2.03 (0.02) 2.06 (0.02) -1.02 -0.04 Outcomes (not used to balance groups)
Amount of Different Use of Force Activities Used 1.55 (0.02) 1.66 (0.03) -3.18 ** -0.12 Taser Discharged 0.13 (0.01) 0.11 (0.01) 1.45 0.05
Firearm Discharged 0.00 (0.00) 0.01 (0.00) -1.16 -0.04
Weapon Discharged 0.13 (0.01) 0.12 (0.01) 1.21 0.04
Community member injured to any extent 0.77 (0.01) 0.79 (0.01) -1.93 -0.07 Amount of different injuries community member sustained 1.02 (0.02) 1.06 (0.02) -1.36 -0.05 Type of force: Carotid restraint control hold 0.00 (0.00) 0.00 (0.00) -2.45 * -0.09 Type of force: Threat of firearm 0.01 (0.00) 0.03 (0.00) -4.19 *** -0.15 Type of force: Other control hold/takedown 0.71 (0.01) 0.76 (0.01) -3.38 *** -0.12 Type of force: Electronic control device 0.13 (0.01) 0.13 (0.01) 0.48 0.02 Type of force: Other physical contact (fists, feet, etc.) 0.28 (0.01) 0.32 (0.01) -2.54 * -0.09 Type of force: Discharge of firearm (miss) 0.00 (0.00) 0.00 (0.00) -1.41 -0.05 Type of force: Blunt/impact weapon 0.08 (0.01) 0.11 (0.01) -3.51 *** -0.13 Type of force: Discharge of firearm (hit) 0.00 (0.00) 0.00 (0.00) 0.00 0.00 Type of force: Chemical spray (e.g. OC/CS) 0.03 (0.00) 0.02 (0.00) 2.07 * 0.07 Type of force: Officer vehicle contact 0.00 (0.00) 0.01 (0.00) -2.84 ** -0.10 Type of force: Impact projectile 0.08 (0.01) 0.06 (0.01) 1.63 0.06 Type of force: K-9 contact 0.06 (0.01) 0.04 (0.01) 2.67 ** 0.10 Type of force: Other dangerous weapon 0.00 (0.00) 0.00 (0.00) -2.65 ** -0.10 Type of force: Taser drive stun 0.02 (0.00) 0.04 (0.01) -2.82 ** -0.10 Type of force: Taser probes deployed 0.11 (0.01) 0.10 (0.01) 1.08 0.04 Type of force: Taser probes deployed (miss) 0.03 (0.00) 0.03 (0.00) -0.11 0.00 * p < .05, ** p < .01, *** p < .001
a Variable not included in propensity score matching
94 Figure A.23: Amount of different use of force activities used, Hispanic vs. white San José community members Model A Model B Model C Model D
IRR 95% CI IRR 95% CI IRR 95% CI IRR 95% CI
Constant 1.55 *** 1.49 | 1.61 1.69 *** 1.60 | 1.78 1.17 *** 1.07 | 1.27 0.64 *** 0.54 | 0.76 Hispanic (vs. white) 1.07 * 1.01 | 1.13 1.07 * 1.01 | 1.13 1.05 1.00 | 1.12 1.06 * 1.00 | 1.12 CM Age 0 to 25 (vs. 26 to 35) 0.86 *** 0.81 | 0.93 0.88 *** 0.82 | 0.94 0.89 ** 0.83 | 0.95 CM Age 36 and older (vs. 26 to 35) 1.00 0.94 | 1.07 1.02 0.95 | 1.09 1.02 0.95 | 1.09 CM is Female (vs. male) 0.77 *** 0.71 | 0.83 0.77 *** 0.71 | 0.84 0.79 *** 0.72 | 0.86 Event Initiated as Apprehension (vs. CFS/Crime) 0.80 * 0.67 | 0.96 0.83 0.69 | 1.00 Event Initiated as Stop (vs. CFS/Crime) 1.06 0.97 | 1.14 1.09 * 1.01 | 1.19 Event Initiated as Other (vs. CFS/Crime) 0.97 0.85 | 1.11 0.95 0.83 | 1.09 Event occur during 1st Shift (vs. 2nd Shift) 1.07 0.99 | 1.15 1.06 0.98 | 1.14 Event occur during 3rd Shift (vs. 2nd Shift) 0.98 0.91 | 1.06 1.00 0.93 | 1.07 Event occur during Unknown Shift (vs. 2nd Shift) 0.90 0.81 | 1.01 0.89 0.80 | 1.00 Two or more Community Members at Event 1.09 0.98 | 1.20 1.07 0.97 | 1.19 Number of Officers at Event 1.01 0.98 | 1.05 1.01 0.98 | 1.04 Number of female officers 1.11 * 1.02 | 1.20 1.10 * 1.01 | 1.19 Number of white officers 1.17 *** 1.12 | 1.22 1.16 *** 1.11 | 1.21 Number of Black officers 1.39 *** 1.21 | 1.60 1.35 *** 1.17 | 1.56 Number of Hispanic officers 1.26 *** 1.20 | 1.32 1.25 *** 1.19 | 1.31 Number of Asian officers 1.15 *** 1.09 | 1.23 1.12 *** 1.06 | 1.19 Number of Other Race officers 1.27 *** 1.18 | 1.38 1.23 *** 1.13 | 1.33 Number of officers with 1 year or less tenure 0.93 ** 0.88 | 0.97 0.92 ** 0.88 | 0.97 Number of officers with 2 to 5 years tenure 1.00 0.97 | 1.04 0.99 0.95 | 1.03 Officer perceived the community member to be armed 1.16 *** 1.09 | 1.23 Level of community member resistance 1.19 *** 1.14 | 1.25 Level of irregular behavior from community member 1.00 0.97 | 1.04
Number of Observations 3,044 3,044 3,044 3,044
Chi-squared 5.65 * 76.45 *** 386.46 *** 462.88 ***
Pseudo R2 .00 .01 .05 .06
Pearson Chi-Squared goodness-of-fit 1,689.76 1,575.90 1,139.91 1,034.02 * p < .05, ** p < .01, *** p < .001
Poisson Regression, Incident Rate Ratios (IRR) reported with 95% Confidence Intervals
95 Figure A.24: Most severe use of force activity used, Hispanic vs. white San José community members Model A Model B Model C Model D
b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI Constant 1.94 (0.02) *** 1.89 | 1.98 1.98 (0.03) *** 1.92 | 2.04 1.78 (0.05) *** 1.68 | 1.87 1.12 (0.09) *** 0.95 | 1.30 Hispanic (vs. white) -0.03 (0.03) -0.09 | 0.04 -0.03 (0.03) -0.09 | 0.04 -0.05 (0.03) -0.12 | 0.01 -0.04 (0.03) -0.11 | 0.02 CM Age 0 to 25 (vs. 26 to 35) -0.05 (0.04) -0.13 | 0.03 -0.07 (0.04) -0.14 | 0.01 -0.03 (0.04) -0.11 | 0.05 CM Age 36 and older (vs. 26 to 35) 0.01 (0.04) -0.07 | 0.09 0.02 (0.04) -0.06 | 0.10 0.02 (0.04) -0.05 | 0.10 CM is Female (vs. male) -0.17 (0.05) *** -0.26 | -0.08 -0.22 (0.05) *** -0.30 | -0.13 -0.16 (0.04) *** -0.25 | -0.07 Event Initiated as Apprehension (vs. CFS/Crime) 0.15 (0.10) -0.04 | 0.33 0.21 (0.09) * 0.03 | 0.38 Event Initiated as Stop (vs. CFS/Crime) -0.26 (0.04) *** -0.35 | -0.18 -0.22 (0.04) *** -0.31 | -0.14 Event Initiated as Other (vs. CFS/Crime) -0.35 (0.08) *** -0.50 | -0.20 -0.37 (0.08) *** -0.53 | -0.22 Event occur during 1st Shift (vs. 2nd Shift) -0.10 (0.04) * -0.19 | -0.01 -0.11 (0.04) * -0.20 | -0.03 Event occur during 3rd Shift (vs. 2nd Shift) 0.04 (0.04) -0.05 | 0.12 0.06 (0.04) -0.02 | 0.14 Event occur during Unknown Shift (vs. 2nd Shift) -0.12 (0.06) * -0.23 | -0.01 -0.16 (0.06) ** -0.27 | -0.05 Two or more Community Members at Event 0.21 (0.06) *** 0.10 | 0.32 0.19 (0.06) ** 0.08 | 0.31 Number of Officers at Event -0.00 (0.02) -0.04 | 0.03 -0.01 (0.02) -0.05 | 0.03 Number of female officers 0.20 (0.05) *** 0.10 | 0.30 0.19 (0.05) *** 0.10 | 0.29 Number of white officers 0.21 (0.03) *** 0.16 | 0.27 0.18 (0.03) *** 0.13 | 0.23 Number of Black officers 0.28 (0.10) ** 0.09 | 0.47 0.22 (0.09) * 0.03 | 0.40 Number of Hispanic officers 0.17 (0.03) *** 0.11 | 0.23 0.13 (0.03) *** 0.07 | 0.19 Number of Asian officers 0.18 (0.04) *** 0.11 | 0.25 0.12 (0.04) ** 0.04 | 0.19 Number of Other Race officers 0.36 (0.05) *** 0.25 | 0.46 0.27 (0.05) *** 0.16 | 0.37 Number of officers with 1 year or less tenure -0.09 (0.03) ** -0.14 | -0.03 -0.08 (0.03) ** -0.14 | -0.02 Number of officers with 2 to 5 years tenure -0.07 (0.02) ** -0.12 | -0.02 -0.08 (0.02) *** -0.13 | -0.04 Officer perceived the community member to be armed 0.36 (0.03) *** 0.29 | 0.42 Level of community member resistance 0.14 (0.02) *** 0.09 | 0.19 Level of irregular behavior from community member 0.05 (0.02) * 0.01 | 0.09
Number of Observations 3,044 3,044 3,044 3,044
Chi-squared 0.64 4.63 ** 16.20 *** 23.75 ***
R2 .00 .01 .08 .12
Root MSE 0.93 0.93 0.89 0.87
* p < .05, ** p < .01, *** p < .001
Linear Regression, Unstandardized Beta Values and Robust Standard Errors reported with 95% Confidence Intervals
96 Figure A.25: Weapon discharged, Hispanic vs. white San José community members Model A Model B Model C Model D
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Constant 0.15 *** 0.13 | 0.17 0.22 *** 0.18 | 0.27 0.22 *** 0.16 | 0.30 0.02 *** 0.01 | 0.04 Hispanic (vs. white) 0.87 0.70 | 1.09 0.88 0.71 | 1.10 0.85 0.68 | 1.07 0.82 0.65 | 1.03 CM Age 0 to 25 (vs. 26 to 35) 0.43 *** 0.32 | 0.58 0.46 *** 0.34 | 0.62 0.50 *** 0.37 | 0.68 CM Age 36 and older (vs. 26 to 35) 0.93 0.73 | 1.19 1.02 0.79 | 1.32 1.03 0.79 | 1.35 CM is Female (vs. male) 0.18 *** 0.10 | 0.31 0.16 *** 0.09 | 0.28 0.17 *** 0.10 | 0.29 Event Initiated as Apprehension (vs. CFS/Crime) 0.20 * 0.05 | 0.83 0.22 * 0.05 | 0.94 Event Initiated as Stop (vs. CFS/Crime) 1.28 0.95 | 1.72 1.49 * 1.09 | 2.02 Event Initiated as Other (vs. CFS/Crime) 0.24 ** 0.10 | 0.57 0.22 ** 0.09 | 0.53 Event occur during 1st Shift (vs. 2nd Shift) 1.08 0.81 | 1.45 1.06 0.78 | 1.43 Event occur during 3rd Shift (vs. 2nd Shift) 1.01 0.76 | 1.34 1.14 0.85 | 1.52 Event occur during Unknown Shift (vs. 2nd Shift) 0.26 *** 0.13 | 0.51 0.24 *** 0.12 | 0.48 Two or more Community Members at Event 1.51 * 1.02 | 2.25 1.43 0.94 | 2.17 Number of Officers at Event 0.83 ** 0.73 | 0.95 0.80 ** 0.69 | 0.92 Number of female officers 1.55 ** 1.12 | 2.16 1.61 ** 1.15 | 2.26 Number of white officers 1.05 0.87 | 1.28 0.99 0.81 | 1.21 Number of Black officers 2.90 *** 1.68 | 4.99 2.31 ** 1.31 | 4.09 Number of Hispanic officers 1.57 *** 1.27 | 1.95 1.48 ** 1.19 | 1.86 Number of Asian officers 1.37 * 1.06 | 1.77 1.18 0.90 | 1.54 Number of Other Race officers 1.37 0.96 | 1.95 1.06 0.73 | 1.55 Number of officers with 1 year or less tenure 0.69 ** 0.56 | 0.86 0.67 *** 0.53 | 0.84 Number of officers with 2 to 5 years tenure 1.08 0.92 | 1.28 1.03 0.86 | 1.22 Officer perceived the community member to be armed 2.81 *** 2.16 | 3.66 Level of community member resistance 1.81 *** 1.47 | 2.23 Level of irregular behavior from community member 1.13 0.98 | 1.30
Number of Observations 3,044 3,044 3,044 3,044
Chi-squared 1.48 110.76 *** 228.10 *** 325.04 ***
Pseudo R2 .00 .05 .10 .14
Pearson Chi-Squared goodness-of-fit 0.00 16.02 * 2,495.97 *** 3,185.20 *** * p < .05, ** p < .01, *** p < .001
Logistic Regression, Odds Ratios (OR) reported with 95% Confidence Intervals
97 Figure A.26: Community member injured to any extent, Hispanic vs. white San José community members Model A Model B Model C Model D
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Constant 3.26 *** 2.90 | 3.67 3.33 *** 2.82 | 3.93 1.65 ** 1.22 | 2.23 0.42 ** 0.25 | 0.71 Hispanic (vs. white) 1.18 1.00 | 1.41 1.18 0.99 | 1.40 1.16 0.97 | 1.38 1.18 0.98 | 1.41 CM Age 0 to 25 (vs. 26 to 35) 1.10 0.89 | 1.35 1.15 0.92 | 1.42 1.25 * 1.01 | 1.56 CM Age 36 and older (vs. 26 to 35) 1.28 * 1.03 | 1.59 1.25 1.00 | 1.56 1.26 * 1.00 | 1.58 CM is Female (vs. male) 0.53 *** 0.43 | 0.66 0.49 *** 0.39 | 0.62 0.53 *** 0.42 | 0.66 Event Initiated as Apprehension (vs. CFS/Crime) 1.49 0.88 | 2.53 1.83 * 1.07 | 3.13 Event Initiated as Stop (vs. CFS/Crime) 0.58 *** 0.46 | 0.73 0.66 ** 0.52 | 0.85 Event Initiated as Other (vs. CFS/Crime) 0.39 *** 0.26 | 0.59 0.41 *** 0.27 | 0.62 Event occur during 1st Shift (vs. 2nd Shift) 0.96 0.75 | 1.24 0.92 0.71 | 1.19 Event occur during 3rd Shift (vs. 2nd Shift) 0.89 0.71 | 1.12 0.91 0.72 | 1.15 Event occur during Unknown Shift (vs. 2nd Shift) 0.89 0.64 | 1.24 0.85 0.61 | 1.19 Two or more Community Members at Event 0.49 *** 0.36 | 0.66 0.52 *** 0.38 | 0.70 Number of Officers at Event 1.05 0.95 | 1.16 1.03 0.93 | 1.15 Number of female officers 1.21 0.86 | 1.69 1.22 0.86 | 1.72 Number of white officers 1.76 *** 1.43 | 2.15 1.57 *** 1.27 | 1.92 Number of Black officers 2.76 ** 1.40 | 5.43 2.14 * 1.07 | 4.27 Number of Hispanic officers 2.24 *** 1.78 | 2.83 1.89 *** 1.50 | 2.38 Number of Asian officers 1.51 ** 1.18 | 1.95 1.26 0.98 | 1.63 Number of Other Race officers 2.08 *** 1.43 | 3.04 1.55 * 1.06 | 2.29 Number of officers with 1 year or less tenure 0.80 * 0.66 | 0.97 0.82 * 0.68 | 0.99 Number of officers with 2 to 5 years tenure 0.93 0.80 | 1.09 0.91 0.78 | 1.07 Officer perceived the community member to be armed 1.95 *** 1.61 | 2.36 Level of community member resistance 1.26 ** 1.10 | 1.44 Level of irregular behavior from community member 1.32 *** 1.16 | 1.49
Number of Observations 3,044 3,044 3,044 3,044
Chi-squared 3.71 40.16 *** 213.48 *** 290.45 ***
Pseudo R2 .00 .01 .07 .09
Pearson Chi-Squared goodness-of-fit 0.00 81.66 *** 2,538.42 *** 2,967.31 *** * p < .05, ** p < .01, *** p < .001
Logistic Regression, Odds Ratios (OR) reported with 95% Confidence Intervals
98 Figure A.27: Amount of different injuries community member sustained, Hispanic vs. white San José community members Model A Model B Model C Model D
IRR 95% CI IRR 95% CI IRR 95% CI IRR 95% CI
Constant 1.02 0.98 | 1.08 1.05 0.98 | 1.13 0.81 *** 0.73 | 0.90 0.43 *** 0.34 | 0.53 Hispanic (vs. white) 1.04 0.97 | 1.11 1.04 0.97 | 1.11 1.03 0.96 | 1.10 1.03 0.96 | 1.11 CM Age 0 to 25 (vs. 26 to 35) 0.96 0.89 | 1.05 0.98 0.89 | 1.06 1.01 0.92 | 1.10 CM Age 36 and older (vs. 26 to 35) 1.09 * 1.00 | 1.19 1.09 1.00 | 1.18 1.08 0.99 | 1.17 CM is Female (vs. male) 0.74 *** 0.67 | 0.82 0.75 *** 0.68 | 0.84 0.77 *** 0.70 | 0.86 Event Initiated as Apprehension (vs. CFS/Crime) 1.03 0.85 | 1.26 1.10 0.90 | 1.34 Event Initiated as Stop (vs. CFS/Crime) 0.92 0.83 | 1.02 0.97 0.87 | 1.08 Event Initiated as Other (vs. CFS/Crime) 0.96 0.82 | 1.14 0.96 0.82 | 1.14 Event occur during 1st Shift (vs. 2nd Shift) 0.98 0.89 | 1.08 0.97 0.89 | 1.07 Event occur during 3rd Shift (vs. 2nd Shift) 0.99 0.91 | 1.09 1.01 0.92 | 1.11 Event occur during Unknown Shift (vs. 2nd Shift) 0.95 0.83 | 1.09 0.95 0.83 | 1.09 Two or more Community Members at Event 0.89 0.78 | 1.02 0.91 0.79 | 1.04 Number of Officers at Event 1.01 0.97 | 1.05 0.99 0.96 | 1.03 Number of female officers 1.02 0.92 | 1.13 1.00 0.90 | 1.12 Number of white officers 1.12 *** 1.06 | 1.19 1.10 ** 1.04 | 1.16 Number of Black officers 1.30 ** 1.09 | 1.56 1.23 * 1.03 | 1.48 Number of Hispanic officers 1.23 *** 1.16 | 1.31 1.20 *** 1.13 | 1.28 Number of Asian officers 1.10 * 1.02 | 1.18 1.05 0.97 | 1.14 Number of Other Race officers 1.26 *** 1.14 | 1.40 1.19 ** 1.07 | 1.31 Number of officers with 1 year or less tenure 0.99 0.93 | 1.05 0.98 0.93 | 1.04 Number of officers with 2 to 5 years tenure 1.01 0.96 | 1.06 1.00 0.95 | 1.05 Officer perceived the community member to be armed 1.24 *** 1.15 | 1.34 Level of community member resistance 1.15 *** 1.08 | 1.22 Level of irregular behavior from community member 1.08 ** 1.03 | 1.12
Number of Observations 3,044 3,044 3,044 3,044
Chi-squared 1.10 44.69 *** 177.66 *** 244.12 ***
Pseudo R2 .00 .01 .02 .03
Pearson Chi-Squared goodness-of-fit 1,814.86 1,762.82 1,626.05 1,569.51 * p < .05, ** p < .01, *** p < .001
Poisson Regression, Incident Rate Ratios (IRR) reported with 95% Confidence Intervals
99 Figure A.28: Most severe sustained injury, Hispanic vs. white San José community members Model A Model B Model C Model D
b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI b (SE) 95% CI Constant 2.19 (0.02) *** 2.15 | 2.23 2.23 (0.03) *** 2.17 | 2.29 1.98 (0.04) *** 1.90 | 2.07 1.50 (0.08) *** 1.34 | 1.66 Hispanic (vs. white) 0.06 (0.03) * 0.01 | 0.12 0.06 (0.03) * 0.00 | 0.12 0.06 (0.03) * 0.00 | 0.11 0.06 (0.03) * 0.01 | 0.12 CM Age 0 to 25 (vs. 26 to 35) -0.00 (0.03) -0.07 | 0.07 0.02 (0.03) -0.05 | 0.08 0.04 (0.03) -0.02 | 0.11 CM Age 36 and older (vs. 26 to 35) 0.08 (0.04) * 0.01 | 0.15 0.06 (0.03) -0.00 | 0.13 0.06 (0.03) -0.01 | 0.13 CM is Female (vs. male) -0.36 (0.04) *** -0.43 | -0.28 -0.35 (0.04) *** -0.42 | -0.28 -0.32 (0.04) *** -0.39 | -0.25 Event Initiated as Apprehension (vs. CFS/Crime) 0.20 (0.08) * 0.05 | 0.35 0.25 (0.08) ** 0.10 | 0.40 Event Initiated as Stop (vs. CFS/Crime) -0.08 (0.04) -0.16 | 0.01 -0.03 (0.04) -0.12 | 0.05 Event Initiated as Other (vs. CFS/Crime) -0.12 (0.07) -0.26 | 0.02 -0.12 (0.07) -0.26 | 0.02 Event occur during 1st Shift (vs. 2nd Shift) -0.05 (0.04) -0.12 | 0.02 -0.06 (0.04) -0.13 | 0.01 Event occur during 3rd Shift (vs. 2nd Shift) -0.04 (0.04) -0.11 | 0.04 -0.02 (0.04) -0.10 | 0.05 Event occur during Unknown Shift (vs. 2nd Shift) -0.07 (0.06) -0.18 | 0.04 -0.08 (0.06) -0.19 | 0.03 Two or more Community Members at Event -0.33 (0.05) *** -0.42 | -0.24 -0.32 (0.05) *** -0.41 | -0.22 Number of Officers at Event 0.02 (0.02) -0.01 | 0.05 0.01 (0.02) -0.02 | 0.04 Number of female officers 0.05 (0.04) -0.03 | 0.14 0.05 (0.04) -0.03 | 0.13 Number of white officers 0.17 (0.02) *** 0.12 | 0.22 0.15 (0.02) *** 0.10 | 0.19 Number of Black officers 0.19 (0.07) ** 0.05 | 0.33 0.13 (0.07) -0.00 | 0.27 Number of Hispanic officers 0.26 (0.03) *** 0.21 | 0.32 0.23 (0.03) *** 0.18 | 0.29 Number of Asian officers 0.16 (0.03) *** 0.10 | 0.22 0.12 (0.03) *** 0.05 | 0.18 Number of Other Race officers 0.19 (0.04) *** 0.11 | 0.28 0.12 (0.04) ** 0.04 | 0.21 Number of officers with 1 year or less tenure -0.09 (0.03) *** -0.14 | -0.04 -0.09 (0.03) *** -0.14 | -0.04 Number of officers with 2 to 5 years tenure -0.07 (0.02) *** -0.11 | -0.03 -0.08 (0.02) *** -0.12 | -0.04 Officer perceived the community member to be armed 0.20 (0.03) *** 0.14 | 0.26 Level of community member resistance 0.10 (0.02) *** 0.05 | 0.14 Level of irregular behavior from community member 0.06 (0.02) *** 0.03 | 0.10
Number of Observations 3,044 3,044 3,044 3,044
Chi-squared 4.71 * 27.05 *** 20.78 *** 23.73 ***
R2 .00 .03 .10 .12
Root MSE 0.81 0.80 0.77 0.76
* p < .05, ** p < .01, *** p < .001
Linear Regression, Unstandardized Beta Values and Robust Standard Errors reported with 95% Confidence Intervals
100
This page intentionally left blank.
3003 Washington Blvd., Arlington Virginia 22201