Receipt of a Final Report and Recommendations on the Analysis of Police Call Data Conducted by the University of Texas at San Antonio — Piedmont City Council, September 19, 2022

City of Piedmont

COUNCIL AGENDA REPORT
DATE: September 19, 2022
TO: Mayor and Council
FROM: Sara Lillevand, City Administrator
SUBJECT: Receive a Report Summarizing Findings and Recommendations by the University of Texas, San Antonio on Data Analysis of Calls for Service Made to the Piedmont Police Department

RECOMMENDATION
Receive this report which summarizes the research, findings and recommendations by researchers from the University of Texas, San Antonio who examined calls for service made to the Piedmont Police Department.

BACKGROUND
On July 19, 2021, the City Council approved an agreement with the University of Texas, San Antonio to conduct an analysis of calls for service to which police officers respond. This effort was initiated by Chief Bowers and supported by City Council in response to national conversations about policing and bias in the aftermath of the killing of George Floyd. The Piedmont Police Department prides itself as an agency that exemplifies the best of policing. The organization strives to continually improve its service delivery to the public, support and develop employees, and engage in benchmarking against industry best practices for improvement. The City of Piedmont is committed to examining its practices, policies and procedures through an equity lens and this work is in alignment with that commitment. The Department appreciates a high level of community trust and support, and in keeping with that relationship, sought to examine how police are utilized within the City. Two overarching areas guided the scope of services for the project:
• The types of calls for service which police officers respond to and the question of if there are alternative responses based on the call types.
• The frequency of biased-based calls for police intervention from the public that lack specific criminal-related behavior.

The scope of work contained within the agreement included the following:
• Conduct data analysis of calls for service for the last five years which police officers responded to.
o Analyze CAD data to examine the frequency of call types, police resources utilized across call types (e.g. number of officers dispatched, time on scene, reports written, etc.), and, if feasible, how calls change from initial dispatch through disposition.
o Categorize call types and make data-driven and/or research-based recommendations on the need for a police officer versus another type of response such as a mental health clinician, Community Service Officer (CSO), or other resources.
o Prepare a comprehensive report detailing findings from the analysis and summarizing recommendations for sworn police vs. non-sworn police (or copolice) response. • Conduct data analysis to determine the frequency of potentially biased-based calls from the public which lack specific criminal-related behavior as a basis for police intervention. o Benchmark calls received against comparison of callers’ descriptions of the race/ethnicity of suspected offenders to:
 (1) reported crime suspects from official police reports, (2) arrestees, (3) field interrogation entries (if available), or other indicators of field contacts by the police.
o Conduct 3-4 focus group interviews of neighborhood
residents/associations/groups to identify concerns and perceptions regarding suspected offenders in Piedmont and analyze resulting data for patterns and trends.
o Prepare a comprehensive report detailing the results of the administrative data and focus group analysis, including recommendations for community engagement/education on the appropriate legal and ethical role of the police in contacting or stopping crime suspects.

The original project timeline and actual dates are reflected in the chart below.

Project Timeline and Deliverables
Deliverable Description Target
Completion Date
Actual
Completion
Date
Notes
Execution of agreement 19-Jul-21 19-Jul-21
Brief project and methodology to
Piedmont stakeholders
August 26, 2021,
(Available on
City’s YouTube
Page)
26-Aug-21 Virtual
Meeting
Conduct focus groups Sep-21 29,30-Sep-21
1-10-21
Virtual
Meetings
Progress report and updates NovemberDecember 2021, 2-Mar-22 Virtual Meeting
DRAFT final report 14-Jan-22 26-Apr-22
FINAL report 4-Feb-22 11-May-22
Agenda Report Page 2
Presentation on project findings and
recommendations to Piedmont
stakeholders
7-18-February-22 23-May-22 Hybrid Meeting in
Council Chambers
The delay in the overall project was due to challenges in extracting data from the Department’s computer software in a form needed to conduct the analysis. The UTSA report, An Assessment of Requests for Police Services in Piedmont, California, may be found on the Department’s Transparency Portal and via the following URL:
https://piedmont.ca.gov/services___departments/police/transparency_portal

Report Recommendations
In addition to the report’s findings, several recommendations were also made for the Department’s consideration. What follows are the Department’s actions, policies, practices, and procedures relative to the corresponding recommendations from the report.

RECOMMENDATION:

Mental and Behavioral Health Training & Response
Some cities are currently experimenting with alternative response models both to mental healthrelated and domestic violence calls (Puntis et al., 2018; Seo et al., 2021; Shapiro et al., 2015).

These models allow non-police professionals to respond to such calls either alone or with police officers in an effort to reduce conflict and provide better outcomes. However, a recent paper from Lum and colleagues (2021) analyzed 911 data from nine jurisdictions ranging in size from 75,000 to over 1,000,000 people and found that, on average, only 1.3% of calls involved a mental health concern. As Lum et al. pointed out, “shifting the small amount of resources currently spent on these calls by the police will likely not be enough to support an effective response to mental health crises (p. 18).

While experiments certainly are underway to replace police officers with mental health and/or emergency medical professionals as responders to non-violent mental health-related calls, these models require substantial new staffing investments to develop or expand availability during non-standard work hours. Given the relatively few mental health and domestic violence-related calls received by the PPD, providing PPD officers with specialized training in how to more effectively handle conflict between intimate partners or safely manage persons with mental or behavioral health problems may be more cost-effective than hiring full-time social workers or other mental health professionals to take the lead or to co-respond on these calls with PPD officers. That said, a few respondents in the focus groups stated that the police should not be called upon to deal with behavioral health problems at the schools absent a clear threat of violence. Yet with such low call volumes for these types of incidents in Piedmont, it may not be necessary or even possible to hire permanent non-police staff as alternative first responders. If rapid, on-call arrangements with local social workers or mental health professionals could be set up, this might provide a viable co-response capability that Piedmont could adopt for mental health or domestic violence-related calls for service.

RESPONSE: Piedmont Police Policy Sections 414 Mental Illness Commitments and 441 Crisis Intervention Incidents provide guidance on how Piedmont Police Department personnel are to Agenda Report Page 3
respond to events which contain mental and behavioral health dynamics. The Department requires police officers to attend Crisis Intervention Training (CIT) within the first year of employment. CIT training which the Department requires is approved by California Police Officer Standards and Training (POST). The overarching goal of a CIT training course is to provide law enforcement officers with the cognition, information, resources, and skills that allow effective problem-solving and promote positive outcomes when responding to incidents involving mental health issues.

The Department has partnered with Alameda County Behavioral Health Care Services by utilizing two components within its Acute & Crisis Services area, the Mobile Crisis Team (MCT) and the Community Assessment and Transport Team (CATT). The MCT operates with two licensed clinicians who respond to mental health crisis calls from the general community, providers, as well as 911 mental health crisis call-outs. MCT services are available Monday – Friday 8am-6pm. The CATT team pairs an emergency medical technician with a licensed clinician. They focus on community-based crisis intervention and medical clearance. This pilot program is a partnership between Bonita House, Alameda County Behavioral Health Care Services, Alameda County Emergency Medical Services, Falck Northern California Emergency Medical Service, and Alameda County Whole Person Care. CATT services are available seven days a week, 7 am-11 pm and is currently attempting to expand its availability.

The Department is sensitive to the complexities relative to a law enforcement presence on school campuses and maintains an excellent working relationship with the Piedmont Unified School District administration and with the staff at each site. The Juvenile Officer has provided practical information to staff from the Wellness Center on the Department’s policies, and how it works with county resources when responding to calls related to mental health crisis. The Juvenile Officer maintains a collaborative working relationship with the Wellness Center staff which has resulted in a nuanced approach which balances general safety considerations with individual student needs.

RECOMMENDATION: Response to Calls for Service
Apart from mental health calls, there are a variety of call types that may be amenable and/or appropriate for alternative responses by non-sworn officers either within or outside of the PPD.

For example, parking complaints, welfare checks, roadway obstructions, lost and found property, and vandalism calls may be addressed without the need for a sworn officer to be dispatched. It may also be efficient and effective to have non-sworn officers take reports on completed property crimes. Any of these adjustments must be considered with safety of the resident(s) and responding entity as a key consideration. While appearing rather innocuous and unlikely to require a sworn officer in response, some situations change rapidly and may unexpectedly require the presence of a PPD officer to assist. If changes to the current PPD response model are contemplated, alternative and/or co-responders must be provided with appropriate resources and training to ensure their safety and success. Sworn officers have the benefit of extensive training needed to handle many different situations. Piedmont should likewise equip alternative responders with the tools needed to improve outcomes over the current model and should track those outcomes carefully to evaluate their safety and efficacy.

The CSOs regularly conduct vacation home checks and assist police officers in procedures such as impounding vehicles associated with criminal investigations. Aside from found property reports, CSOs do not respond to take police reports for other matters or crimes. Duties of the CSO position, particularly parking enforcement, and park patrol, are frequently requested by the public and other City departments. Increased functions of the CSO position would likely require additional staffing, perhaps a change to full-time staff, and certainly additional training.

RECOMMENDATION: A key theme of the focus groups was to avoid having police officers respond to mental health situations within the schools. Based on resident comments, the PPD should review its current response approach to these calls and ensure that officers are only dispatched when there is clear evidence of criminal activity.

RESPONSE: The Juvenile Officer position was established in 2019 after the City of Piedmont secured a grant to fund the position through the State of California, Department of Justice Tobacco Grant. PPD and PUSD collaboratively established the practices for when the Juvenile Officer responds to school campus, which is generally when called. The Juvenile Officer will respond proactively if there is a specific public safety need or threat to the campuses.

RECOMMENDATION: Several focus group members opined on the importance of collecting information provided by residents on suspicious persons to disentangle behaviorally-based descriptions of ‘suspects’ from perceptions of ‘risk’ that may not accurately reflect true risk.

Some suggested offering residents community education courses or curriculum to improve the ability of residents to disentangle perception from intention when assessing individuals on the street and to prepare community members for the questions they will be asked by PPD calltakers if they call to report suspicious persons or activity.

RESPONSE: The Department has worked hard to increase the public’s understanding of the services it provides and how to best access those services. One of the mechanisms to deliver this information is through a social media production entitled, “Inside PPD”. The Department is in the process of developing content for an educational piece about the appropriate use of 911.

Included in the production will be tips for the public to consider when reporting suspicious person’s, specifically, informing the public that 911 staff will want to know the activity or actions the person is engaged in that are specific for general safety, legal and officer safety reasons. The Inside PPD production will be disseminated through the Department and City’s social media, KCOM, and local media.

RECOMMENDATIONS: Data Collection and Storage
• PPD should consider a review of current call types and how they are categorized into priority categories based on analyses of the calls for service data. Some call types currently listed as ‘medium’ priority may be better located in the ‘high’ category and vice versa.

Agenda Report Page 5
• Current data collection and storage processes do not allow for an assessment of how calls may change over time. This is an important piece of information to fully understand the nature of calls received by the PPD and how often they change during their lifecycles.

RESPONSE: Unfortunately, as the recommendations in the section above highlight, there were technical challenges with the Department’s computer software which made extracting essential information problematic in some cases and impossible in other areas. The Department will share the impressions and recommendations with its CAD vendor to address the areas mentioned. The Department views the ease of access to its data as an essential component of its commitment to transparency. If the CAD vendor is unable or unwilling to make improvements and changes then the Department may consider a new vendor.
• RECOMMENDATION: The codes used for indicating that the call has been resolved should be reviewed and updated. Currently, the most frequently used call disposition category is ‘other,’ which presumably represents a number of different actions or resolutions but lacks specificity. It is not clear what type of action was taken in these situations, and further training on appropriate fields to select and/or the addition of new fields to accurately categorize resolutions would benefit the PPD.

RESPONSE: The Department has addressed the “other” disposition code observation. The “other” category was typically used when officers did not take formal action, such as issuing a citation, making an arrest, or writing a report. The Department analyzed a sample size of the calls for service with the “other” disposition and determined that other existing disposition codes were available and provided a more accurate description of the call resolution. As a result of the analysis, the Department removed the “other” category as an available option. The category of “courtesy assist” was added as an available option.

• RECOMMENDATION: Given California’s adoption of AB953, it is recommended that the PPD review its current RMS and CAD systems to ensure compliance with RIPA and to maximize the value of this effort for the PPD and the City of Piedmont. Given its size, PPD has until April 2023 to become RIPA compliant but began reporting RIPA data to Agenda Report Page 6
the California Department of Justice in July 2021. There are a number of scientifically accepted approaches to analyzing stop data. These should be reviewed and preparations made by the PPD and City to conduct scientifically appropriate and useful analyses of the new data as it becomes available and not rely solely on state-provided reports, which may not be helpful to the Piedmont community given the City’s unique make-up and context.

RESPONSE: The Department is researching opportunities for useful and appropriate analyses of its RIPA data once it becomes available. Such an analysis will likely require the Department and City to retain the services of researchers and/or subject matter experts in these efforts.

Work in the areas incorporated in the report’s recommendations began far in advance of the report and will continue long after. The Department’s strategic plan will be updated to reflect areas the report identified which were not readily known by the Department such as the limitations of the software systems. From the outset, the Piedmont Police Department and City have viewed the final report from UTSA not as a resolution to the topics associated with it, but as one stop along the path of continual assessment and improvement.

ATTACHMENTS
Attachment 1: An Assessment of Requests for Police Services in Piedmont, California: Dr.

Michael R. Smith, Dr. Rob Tillyer, University of Texas at San Antonio By: Jeremy Bowers, Chief of Police
Agenda Report Page 7
An Assessment of Requests for Police Services
in Piedmont, California
Dr. Michael R. Smith
Dr. Rob Tillyer
University of Texas at San Antonio
May 2022
Agenda Report Page 8
i

TABLE OF CONTENTS
TABLE OF CONTENTS..............................................................................................................................................I TABLES.......................................................................................................................................................................II FIGURES.....................................................................................................................................................................II EXECUTIVE SUMMARY........................................................................................................................................IV RACIAL/ETHNIC COMPARISONS ................................................................................................................................. V RECOMMENDATIONS.................................................................................................................................................VI BACKGROUND .......................................................................................................................................................... 1 RESEARCH GOALS................................................................................................................................................... 1 METHODOLOGY....................................................................................................................................................... 2 VARIABLES ................................................................................................................................................................ 4 ANALYTIC STRATEGY ................................................................................................................................................ 5 QUANTITATIVE RESULTS..................................................................................................................................... 6 CALLS FOR SERVICE .................................................................................................................................................. 6 Call Type .............................................................................................................................................................. 6 Units Dispatched................................................................................................................................................ 10 Time.................................................................................................................................................................... 13 Dispositions........................................................................................................................................................ 15 RACIAL/ETHNIC COMPARISONS ............................................................................................................................... 18 Descriptives........................................................................................................................................................ 19 Disproportionality Indices.................................................................................................................................. 23 FOCUS GROUPS....................................................................................................................................................... 28 COMMUNITY FOCUS GROUPS................................................................................................................................... 28 Types of Calls Made to the Piedmont Police Department (PPD) ...................................................................... 28 Source of Calls to the PPD................................................................................................................................. 29 Calls the PPD Should Not Respond To .............................................................................................................. 29 Suspicious Persons............................................................................................................................................................ 30 Mental Health-Related Calls........................................................................................................................................... 31 Nuisance Calls ................................................................................................................................................................. 31 Calls from Schools .......................................................................................................................................................... 31 PERCEPTION OF CRIME AND PUBLIC SAFETY IN PIEDMONT................................................................................. 32 Home Invasions................................................................................................................................................. 32 Drug Use............................................................................................................................................................ 32 Public Safety-Related........................................................................................................................................ 32 DISPROPORTIONATE CONTACTS WITH NON-WHITE CIVILIANS AND BIAS BY PROXY ....................................... 33 Perceptions of “Typical” Offenders................................................................................................................ 33 Calling the Police About Non-Residents......................................................................................................... 33 Police Response to Calls about Persons of Color.......................................................................................... 33 A Role for Community Education .................................................................................................................... 34 RACIAL BIAS BY THE PPD ..................................................................................................................................... 34 SCHOOL-BASED FOCUS GROUP ............................................................................................................................. 34 Types of Calls to the PPD from the Schools ................................................................................................... 35 Call Types PPD Should Not Be Responsible For........................................................................................... 35 Agenda Report Page 9
ii
Police Officers in the Schools .......................................................................................................................... 35 Perceptions of Crime in and Around the Schools........................................................................................... 35 Perceptions of Unauthorized Persons................................................................................................................ 36 CONCLUSION........................................................................................................................................................... 36 SUMMARY & RECOMMENDATIONS............................................................................................................. 36 RECOMMENDATIONS................................................................................................................................................ 38 Mental and Behavioral Health Training & Response........................................................................................ 38 Response to Calls for Service ............................................................................................................................. 38 Data Collection and Storage .............................................................................................................................. 39 REFERENCES........................................................................................................................................................... 41 APPENDIX A ............................................................................................................................................................. 43 CALL TYPE FREQUENCIES BY PRIORITY LEVEL ....................................................................................................... 43 APPENDIX B ............................................................................................................................................................. 46 NARRATIVE ABBREVIATIONS USED BY THE PPD..................................................................................................... 46 APPENDIX C ............................................................................................................................................................. 47 SUMMARY OF RACE/ETHNICITY FROM ALL DATA SOURCES ................................................................................... 47 APPENDIX D ............................................................................................................................................................. 48 FOCUS GROUP INTERVIEW SCRIPT ........................................................................................................................... 48 TABLES
TABLE 1: DATA VARIABLES ........................................................................................................................................... 4 TABLE 2: ACCURACY CHECK........................................................................................................................................ 19 TABLE 3:SUMMARY OF RECORDS ................................................................................................................................ 20 TABLE 4: HIGHEST PRIORITY CALL TYPES FREQUENCIES ............................................................................................ 43 TABLE 5: MEDIUM PRIORITY CALL TYPES FREQUENCIES............................................................................................. 44 TABLE 6: LOWEST PRIORITY CALL TYPES FREQUENCIES............................................................................................. 45 TABLE 7:SUMMARY OF RACE/ETHNICITY FROM ALL DATA SOURCES ........................................................................ 47

FIGURES
FIGURE 1: CAD SYSTEM SOURCE OF RECORD................................................................................................................ 3 FIGURE 2:PRIORITY LEVEL ............................................................................................................................................ 6 FIGURE 3: INCIDENT CALL TYPES................................................................................................................................... 8 FIGURE 4: CALL TYPES BY PRIORITY LEVEL .................................................................................................................. 9 FIGURE 5: UNITS DISPATCHED...................................................................................................................................... 10 FIGURE 6:SINGLE UNIT DISPATCH BY CALL TYPE....................................................................................................... 11 FIGURE 7: MULTIPLE UNITS DISPATCHED BY CALL TYPE ............................................................................................ 12 FIGURE 8: TIME TO ARRIVAL........................................................................................................................................ 13 FIGURE 9: TIME ON SCENE ........................................................................................................................................... 14 FIGURE 10:PRIORITY CALL TYPE BY AVERAGE ARRIVAL TIME & TIME ON SCENE.................................................... 15 Agenda Report Page 10
iii
FIGURE 11: DISPOSITIONS............................................................................................................................................. 16 FIGURE 12: CALL PRIORITY LEVELS BY DISPOSITION................................................................................................... 17 FIGURE 13: COMPARISON OF RACE/ETHNICITY ACROSS DATA SOURCES .................................................................... 21 FIGURE 14:SPECIFIC CFS TYPES.................................................................................................................................. 22 FIGURE 15: DI - COMPARISON OF PUBLICLY GENERATED INFORMATION TO POLICE ACTIVITY................................... 24 FIGURE 16: DI - COMPARISON OF SPECIFIC CFS TO POLICE ACTIVITY......................................................................... 25 Agenda Report Page 11
iv

EXECUTIVE SUMMARY
In the summer of 2021, the City of Piedmont contracted with Dr. Michael R. Smith and Dr. Rob Tillyer from the University of Texas at San Antonio (UTSA) to analyze five years of computeraided dispatch (CAD) and related data from the Piedmont Police Department (PPD) to help inform decision-making about police response to certain types of calls and whether an alternative response might be feasible or even preferrable. In addition, the City and PPD Chief Jeremy Bowers were interested in an analysis that would examine the potential problem of “bias by proxy” in Piedmont, which refers to disproportionate contacts of non-White civilians1 by the police as the result of calls from the public.

The Piedmont-UTSA research agreement included the following two goals and associated deliverables:

1. Analyze calls for service for the last five years for the purpose of categorizing call types, PPD resource allocation, and providing data-driven and/or research-based recommendations on the need for a police officer versus another type of response such as a mental health clinician, Community Service Officer (CSO), or other resource.

2. Analyze the frequency of potentially biased-based calls from the public that lacked specific criminal-related behavior as a basis for police intervention and compare the racial and/or ethnic composition of police field interrogation stops and arrests to callers’ descriptions of the race/ethnicity of suspected offenders.

Several types of data were obtained from the PPD to facilitate this research study. These data covered a five-year period from January 1, 2017 through December 26, 2021:
• CAD records (all calls to the police)
• Civilian reports to the police regarding suspected or actual criminal activity • Data on self-initiated police activities, including stops of civilians that resulted in a field interrogation (FI) report and/or PPD arrest records

KEY FINDINGS
Calls for Service
The majority of calls received were categorized by the PPD as ‘highest’ priority (71.4%), while the ‘medium’ (20.8%) and ‘lowest’ (7.8%) priority categories comprised a notably smaller 1 Throughout this report, the term “civilian,” rather than “citizen,” is used to describe non-police individuals in the community with whom the police come into contact. The term civilian is preferred over citizen because some people contacted by the police are not “citizens,” either of the United States or the City of Piedmont. They may not permanently reside in the U.S., may not have official citizenship status, or simply may not live in Piedmont as “citizens” of the city. Thus, we use the term “civilian” to describe all members of the public who are not police officers.

Agenda Report Page 12
v
amount of the overall calls for service. The majority of calls involved ‘checking for evidence of a theft/burglary’ (14.9%), report of a ‘burglar alarm’ (12.7%), responding to ‘suspicious circumstances’ (9.1%), ‘ambulance emergency’ (7.4%), and ‘parking complaints’ (5.5%).

Notably, mental health-related calls comprised only .3% (n=85) of all calls received. A widerange of other call types made up the remainder of calls to the PPD with most representing under 1% of all calls received.

The PPD dispatched a single officer to about 47% of calls received and two officers to an additional 35% of calls. When multiple officers were dispatched, they were most often sent to calls involving burglar alarms, suspicious circumstances, ambulance emergencies, or to check for evidence of a burglary or theft. The average time it took PPD officers to respond to a call was 4 minutes, but nearly 30% of calls resulted in the arrival of an officer in under a minute. On average, officers spent 13 minutes on the scene of a call, and most calls (34%) resulted in an ‘other’ disposition, which is often used as a catch-all category by the PPD when other disposition types do not apply. The next most frequent disposition was ‘report taken’ (11.3%) followed by ‘no report taken’ (11.2%) and ‘false alarm’ (10.3%).

Racial/Ethnic Comparisons
To address the question of potential “bias by proxy,” or officers disproportionately contacting non-White civilians based on calls received by the PPD, the UTSA research team produced disproportionality indices (DI), which indicate the degree of alignment between the rates of contact with racial/ethnic groups as a result of PPD-initiated activities (FI card and arrest data) and the rates of contact with those groups based on publicly-generated data – suspects descriptions provided by callers to the PPD and suspect descriptions provided by persons reporting a crime to the PPD. A DI value above 1.0 indicates that the PPD engaged with the group of interest at higher levels than might be expected based on the publicly-generated data, whereas a value below 1.0 denotes that the PPD contacted that racial/ethnic group less frequently than what appears in the publicly-generated data.

With some variation, most comparisons hovered at or near 1.0, suggesting a high degree of alignment between self-initiated PDD activities and suspect representation by race/ethnicity in calls for service and crime reports. Six of the 16 comparisons were above 1.0, indicating that the rate of race/ethnicity in the police activity data exceeded the representation in the publiclygenerated data. For example, stops of White civilians were slightly higher in the FI card data compared to the crime suspect data (1.1), and arrests of Black civilians slightly exceeded the CFS (1.1) and suspect data (1.1). Other variances were evident for Hispanic (arrests vs. CFS) and Asian civilians (FI cards and arrests vs. crime suspects). Overall, the data showed that the PDD stopped (as represented by FI cards) and arrested racial groups at rates that closely matched the racial composition of suspects described by callers and in official police reports, especially for Whites, Blacks, and Hispanics.

Agenda Report Page 13
vi
The findings for Asians showed somewhat greater variability in the DIs than among the other minority groups, but they were not markedly different except for with some specific call-type comparisons (suspicious circumstances & check for theft/burglary calls). Within these call types, Asians were consistently, and in a few cases, significantly over-represented among persons stopped and arrested by the PPD when compared to suspect descriptions provided by civilian callers reporting ‘suspicious circumstances’ or requesting the PPD to ‘check for evidence of theft/burglary’. Additional research and more comprehensive data2are needed to better understand the stop and arrest patterns among all racial and ethnic groups in Piedmont, but especially among Asian civilians.

Focus Groups
Focus group participants consistently told the UTSA research team that crime is generally low in Piedmont. At the same time, some participants expressed concerns about young people from outside the community coming into Piedmont to commit property crimes (e.g., breaking into vehicles) and were appreciative that PPD is responsive to concerns about suspicious persons when called. Other respondents were clearly sensitive to being perceived as racially biased, and some expressed strongly held views that PPD should not respond to calls about suspicious-looking “outsiders” unless callers can specifically describe behaviors or mannerisms that suggest the potential for criminal activity. Finally, there was consensus among the respondents about the need for community engagement and education by the PPD on when residents should call the police about suspicious persons in their neighborhoods, what behaviors to watch out for, and questions to expect from PPD calltakers when residents call to report suspicious behavior.

Recommendations
• Given the relatively few mental health and domestic violence-related calls received by the PPD, providing PPD officers with specialized training in how to more effectively handle conflict between intimate partners or safely manage persons with mental or behavioral health problems may be more cost-effective than hiring full-time social workers or other mental health professionals to take the lead or to co-respond on these calls with PPD officers.
• A key theme of the focus groups was to avoid having police officers respond to mental health situations within the schools. Based on resident comments, the PPD should review its current response approach to these calls and ensure that officers are only dispatched when there is clear evidence of criminal activity.
2 PPD will be required to comply with the data collection requirements of the California Racial and Identify Profiling Act (AB 953) by April 1, 2023. https://post.ca.gov/Racial-and-Identity-Profiling-Act. Stop data collected pursuant to this Act will, over time, provide Piedmont with the information needed to further investigate possible racial or ethnic disparities in officer-initiated stops conducted by the PPD.

Agenda Report Page 14
vii
• PPD should consider offering residents community education courses or curriculum to improve the ability of residents to disentangle perception from intention when assessing individuals on the street and to prepare community members for the questions they will be asked by PPD call-takers if they call to report suspicious persons or activity. • PPD should consider a review of current call types and how they are categorized into priority categories based on analyses of the calls for service data. • The PPD should review its current CAD system and seek to improve its capability for tracking how calls change and are sometimes re-classified both by type and priority during their lifecycles.
• The PPD would benefit from working with its CAD vendor to develop discrete fields that call-takers could use to record descriptions of suspects provided by callers. • Further refinement of disposition codes would allow PPD officers and dispatchers to more accurately record call dispositions rather than relying so heavily on the ‘other’ disposition code that currently comprises the bulk of call dispositions recorded in the CAD system.
• While PPD has been reporting Racial Identify and Profiling Act (RIPA) data to CALDOJ since July 2021, the PPD should review its current RMS and CAD systems to maximize the value of this effort for the PPD and the City of Piedmont and lay the groundwork to conduct its own scientifically appropriate analyses of the new data rather than relying solely on state-provided reports, which may not be helpful to the Piedmont community given the City’s unique make-up and context.
• Piedmont should evaluate the capabilities of its current CAD system to easily and accurately extract data for analysis.

Agenda Report Page 15
1 BACKGROUND
In late spring 2021, Chief Jeremy Bowers from the Piedmont Police Department (PPD) contacted Dr. Michael Smith at the University of Texas at San Antonio (UTSA) to discuss collaborating on a research project with PPD and the City of Piedmont. Chief Bowers and other city leaders were interested in having an analysis done of the city’s 911 and non-emergency calls to the police to help inform decision-making about police response to certain types of calls and whether an alternative response might be feasible or even preferrable. In addition, Chief Bowers also was interested in an analysis that would examine the potential problem of “bias by proxy” in Piedmont, which refers to disproportionate contacts of non-White civilians3 by the police as the result of calls from the public.

Dr. Smith, and his colleague at UTSA, Dr. Rob Tillyer, are experienced criminologists and police researchers who have conducted many applied research projects with police agencies, including a recent, large-scale analysis of 911 call data for the City of San Antonio and numerous projects involving racial disparities in police-civilian contacts. After preliminary discussions between Dr. Smith and Chief Bowers, and the exchange of draft proposals and scopes of work, the City of Piedmont and UTSA entered into a research contract in July 2021 to perform the agreed-upon work, which is described more fully below, with Dr. Smith and Dr.

Tillyer as the lead investigators on the project.

We begin this report by outlining the research goals agreed-upon with the City in the research contract. These goals framed the research methodology that follows and informed the quantitative analyses and focus group results that we present next. We conclude the report with a summary and set of recommendations.

RESEARCH GOALS
The project scope entailed addressing two research goals and associated deliverables:

1. Conduct data analysis of calls for service for the last five years to which police officers responded
o Analyze CAD data to examine frequency of call types, police resources utilized across call types (e.g., number of officers dispatched, time on scene, reports written, etc.), and, if feasible, how calls change from initial dispatch through disposition.
3 Throughout this report, the term “civilian,” rather than “citizen,” is used to describe non-police individuals in the community with whom the police come into contact. The term civilian is preferred over citizen because some people contacted by the police are not “citizens,” either of the United States or the City of Piedmont. They may not permanently reside in the U.S., may not have official citizenship status, or simply may not live in Piedmont as “citizens” of the city. Thus, we use the term “civilian” to describe all members of the public who are not police officers.

Agenda Report Page 16
2 o Categorize call types and make data-driven and/or research-based recommendations on the need for a police officer versus another type of response such as a mental health clinician, Community Service Officer (CSO), or other resource.
o Prepare a comprehensive report detailing findings from the analysis and summarizing recommendations for sworn police vs. non-sworn police (or copolice) response.

2. Conduct data analysis to determine the frequency of potentially biased-based calls from the public which lack specific criminal-related behavior as a basis for police intervention
o Benchmark calls received against comparison of callers’ descriptions of the race/ethnicity of suspected offenders to:
▪ (1) reported crime suspects from official police reports, (2) arrestees, (3) field interrogation entries (if available) or other indicators of field contacts by the police.
o Conduct 3-4 focus group interviews of neighborhood
residents/associations/groups to identify concerns and perceptions regarding suspected offenders in Piedmont and analyze resulting data for patterns and trends
o Prepare a comprehensive report detailing the results of the administrative data and focus group analysis, including recommendations for community engagement/education on the appropriate legal and ethical role of the police in contacting or stopping crime suspects

METHODOLOGY
Achievement of the project’s research goals required access to four data sources, all of which were provided by the Piedmont Police Department (PPD). These data were examined in their original form, or the original records were used to create new measures (hereafter referred to as variables) to complete relevant analyses. Critical to this effort was the accessibility of civilian race/ethnicity information; all four data sources contained some information on this field of interest.

One source of data were records extracted from the computer-aided dispatch (CAD) system, which captures two primary sources of information:

1. Calls for service from the public received by Piedmont’s central call system. These data are generated by recording relevant information from the caller (i.e., the public) and/or additional information added by the dispatcher to provide a comprehensive record of the call for service. For example, calls made to the PPD to request Agenda Report Page 17
3 assistance from the police using the 911 system appear in these data, as do calls received from the PPD’s non-emergency number.

2. Officer-initiated actions undertaken while on duty. For example, if an officer initiated a traffic stop, a record of this activity is generated in the CAD system.

The remaining records in the CAD system (grouped as “miscellaneous”) were calls from other agencies for assistance (e.g., animal control) or internal communications. Collectively, 60,153 records represented activity between January 1, 2017 and December 26, 2021. As displayed in Figure 1, the majority of CAD records were generated from calls for service from the public (50.1% or 30,114 records) with 45.4% of the records created by an officer-initiated action (27,316 records). The remaining records were grouped as miscellaneous records (4.5% or 2,578 records), the majority of which were related to animal control.

Figure 1: CAD System Source of Record
Given the research goals of this project, calls for service (CFS) received by Piedmont’s central call system and recorded in the CAD system were the primary data source used in subsequent analyses. Information contained in these records included the date and time of the call, categorization of the call based on the PPD’s call types and definitions (e.g., suspicious circumstances, theft, etc.), and a narrative summary of the call content. In some records, the narrative field contained information about the race/ethnicity of a civilian(s) about which the call was made. For example, a caller may request assistance from the police because a suspicious person was seen walking through the neighborhood, and the call-taker may obtain a description of the race/ethnicity of the suspicious individual from the caller.

A second source of data came from civilian reports to the police regarding suspected or actual criminal activity. These police reports also may include information on the suspect(s) believed to

50.1%

45.4%

4.5%
0%
10%
20%
30%
40%
50%
60%
Citizen Call for Service Officer Initiated Record Miscellaneous Percent
Source of Record
Percent of Records
Agenda Report Page 18
4 be responsible for the event. For example, an officer may create an official report about a car theft based on a complaint received from a resident whose car was stolen. In additional to basic information about the time and place of the crime, some records also include a description of the suspect’s race/ethnicity as provided to the police. Collectively, these two data sources represent publicly-generated records of requests for police assistance (i.e., calls for service) and suspected criminal activity from police reports (i.e., suspects). Estimates of racial/ethnic composition from these two data sources reflect citizen perspectives, but not police agency activity.

The third and fourth sources of data were generated by self-initiated police activities and measure street contacts with civilians for the purposes of public safety, criminal investigations, and/or intelligence gathering (i.e., field interview cards; FI cards), and arrests of individuals. For the purposes of this project, these data contained the race/ethnicity of the civilian involved whenever possible, although this information was not always available depending on the specific incident.

In sum, the first and second data sources (i.e., CFS and suspects) reflect publicly-generated information, while the third and fourth data sources (i.e., FI cards and arrests) represent selfinitiated actions taken by the PPD. This distinction is critical and lays the foundation for the creation of variables in each data set and the subsequent analyses undertaken to address the key research questions.

Variables
All data contained relevant pieces of information that were either directly analyzed or slightly modified for analysis purposes. For example, fields that contained date and time information were combined to create new variables that measured the amount of time between when the call was received and when the initial officer arrived at the scene (i.e., Time to Arrival) and a measure of how long an officer was present at the scene (i.e., Time on Scene). No substantive information was altered from the original fields during this process. Table 1 summarizes all variables from the four data sources and provides a definition for each field.

Table 1: Data Variables
Data Source Definition
Call For Service Data
Call Type Categorization of the call based on the description provided to the dispatcher by the caller; over 100 categories of call types are available Priority Type A hierarchical categorization of call types; defined by the PPD Categories: Highest, Medium, and Lowest
Units Dispatched The number of police units sent in response to a call for service Time to Arrival The difference in minutes between when the call was received and when the initial officer arrived at the location of the problem.

Agenda Report Page 19
5 Time on Scene The difference in minutes between when the initial officer arrived at the location of the problem and when the last officer left the scene.

Disposition The final resolution to the call for service (e.g., report taken, a citation written, a verbal warning issued, a field interview card completed, etc.) Civilian
Race/Ethnicity
Each narrative record for a CFS was examined to identify civilian race/ethnicity. This characteristic was coded as White, Black, Hispanic, or Asian.

Suspect, FI Card, &
Arrest Data
Civilian
Race/Ethnicity
This characteristic was coded as White, Black, Hispanic, Asian, or Other (any identified race/ethnicity that did not fit within the four listed categories).

Analytic Strategy
Several analytic approaches were utilized to address the research goals. For the first research goal, descriptive statistics were calculated to summarize the patterns within the CFS data. This approach produces a single statistic (i.e., number) that represents how often any characteristic of interest appears within the data. Further analyses were also conducted by intersecting one field of interest with another field of interest to identify how patterns in one variable align with the patterns in another variable (bivariate analyses).

Descriptive statistics were also used to address the second research goal by summarizing the composition of civilian race/ethnicity across all data sources. These statistics reflect the frequency with which the different racial/ethnic groups appear in the various datasets. In sum, percentages were calculated for each racial/ethnic group that appeared in the CFS, suspect, FI card, and arrest databases. Note that the CFS data were further sub-divided into specific call types and analyzed in the following ways:
a) All calls for service that included a reference to the race/ethnicity of the civilian of interest,
b) Calls for service regarding suspicious circumstances that included a reference to the race/ethnicity of the civilian of interest
c) Calls for service regarding checks for evidence of a theft/burglary that included a reference to the race/ethnicity of the civilian of interest Beyond frequencies, comparisons of racial/ethnic representations across databases were also conducted. This step involved the use of the publicly-generated data as a comparison against the actions undertaken by the PPD. In this case, the data of interest are the actions taken by the PPD (i.e., FI card and arrest data) in relation to the CFS and suspect data.

To conduct this comparison, disproportionality indices (DI) were calculated by comparing the percentage of each racial/ethnic group in the publicly-generated data to their representation in the self-initiated PPD data. This allows for an evaluation of whether a particular racial or ethnic Agenda Report Page 20
6 group was over or under-represented in PPD-initiated activities compared to their representation among those whom civilians called about or identified as criminal suspects. The DI statistic varies around a value of 1.0, which indicates alignment between the rate of contact with a racial/ethnic group as a result of PPD actions (i.e., FI card and arrest data) and the rate of representation of the group in the publicly-generated data (i.e., CFS and suspect data). A value above 1.0 indicates that the PPD contacted or arrested the group of interest at higher levels than what would be expected based on the publicly-generated data, whereas a value below 1.0 denotes that the PPD contacted or arrested that racial/ethnic group less frequently than they appear in the publicly-generated data.

QUANTITATIVE RESULTS
Calls For Service
Data were accessed for all calls for service between Jan 1, 2017 and Dec 26, 2021. During this time period, 30,114 calls for service were received and resulted in the dispatch of at least one officer to the scene.

Call Type
As noted above, calls for service were categorized in two ways using pre-existing categories defined by the PPD. Initially, each call was placed into a specific category, and these call types were also categorized as ‘highest’, ‘medium’, and ‘lowest’ priority. Figure 1 below displays the percentage of all calls that appeared in each of the Priority categories. The majority of calls corresponded to call types that were categorized as ‘highest’ priority (71.4%), while the ‘medium’ (20.8%) and ‘lowest’ (7.8%) priority categories comprised a notably smaller amount of the overall calls for service.

Figure 2: Priority Level

71.4%

20.8%

7.8%
0%
10%
20%
30%
40%
50%
60%
70%
80%
Highest Medium Lowest
Priority Level
Percent of Records
Agenda Report Page 21
7 Figure 3 below summarizes the most frequently occurring call types. There were a total of 93 distinct call types used during this five year period; Figure 3 shows call types that accounted for at least 1% (each) of the overall calls for service. The majority of calls involved ‘checking for evidence of a theft/burglary’ (14.9%), report of a ‘burglar alarm’ (12.7%), or responding to ‘suspicious circumstances’ (9.1%). Additional calls for service not displayed include ‘burglary of a residence’ (0.7%), ‘vehicle stolen’ (0.6%), and ‘mental health hold’ (0.3%). Calls that represent less than 0.1% of all calls received include ‘assault,’ ‘robbery,’ and ‘burglary of a business’. A complete list of all call types, the number of records for each call type, and the priority level for each call type appears in Appendix A.

Figure 4 below intersects call types with priority levels in two ways. Within each priority level, the orange bars report the most frequently occurring call types, while the blue bars demonstrate the overall percentage of call types. For example, ‘checking for evidence of theft/burglary’ represented 19.9% of the high priority calls and 14.9% of all calls, whereas ‘animal control piedmont call for service’ represented 32.2% of all medium priority calls but only 7.0% of all calls for service. Likewise, calls for ‘vehicles parked in violation of the 72-hour city ordinance’ comprised 39.4% of the lowest priority calls, but just 3.1% of all calls for service.

Agenda Report Page 22
8 Figure 3: Incident Call Types

14.9%

12.7%

9.1%

7.4%

7.0%

5.5%

4.9%

3.4%

3.1%

2.8%

2.6%

2.0%

1.7%

1.6%

1.2%

1.1%

1.0%

1.0%
0% 2% 4% 6% 8% 10% 12% 14% 16%

CHECK FOR EVIDENCE OF THEFT/BURGLARY

BURGLAR ALARM

SUSPICIOUS CIRCUMSTANCES

AMBULANCE EMERGENCY

ANIMAL CONTROL PIEDMONT CALL FOR SERVICE

PARKING COMPLAINT

PIEDMONT CITY ORDINANCE

WELFARE CHECK
VEHICLE PARKED IN VIOLATION OF 72 HOUR CITY ORDINANCE
THEFT

AMBULANCE DISPATCHED TO OUTSIDE JURISDICTION

FIRE ALARM RESIDENTIAL

OBSTRUCTION IN ROADWAY

ARGUMENT/YELLING ETC (UNKNOWN RELATIONSHIP)

PROERTY LOST AND FOUND

HANG UP/NO RESPONSE

VANDALISM

AUTO BURGLARY

PANIC OR HOLD UP ALARM
Call Type
Percent of Records
Agenda Report Page 23
9 Figure 4: Call Types by Priority Level

14.9%

12.7%

9.1%

7.4%

5.5%

3.4%

2.6%

7.0%

4.9%

2.8%

0.9%

0.8%

0.7%

0.6%

3.1%

1.6%

0.5%

19.9%

17.8%

12.8%

10.3%

7.7%

4.5%

3.6%

32.2%

23.3%

13.3%

4.4%

3.6%

3.2%

3.0%

39.4%

20.6%

6.2%

6.1%
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%

CHECK FOR EVIDENCE OF THEFT/BURGLARY

BURGLAR ALARM

SUSPICIOUS CIRCUMSTANCES

AMBULANCE EMERGENCY

PARKING COMPLAINT

WELFARE CHECK

AMBULANCE DISPATCHED TO OUTSIDE JURISDICTION

ANIMAL CONTROL PIEDMONT CALL FOR SERVICE

PIEDMONT CITY ORDINANCE
THEFT

ACCIDENT WITH NO INJURIES

ACCIDENT HIT AND RUN

GIVING GUIDANCE MEDIATE NON-CRIMINAL MATTERS

ANIMAL CONTROL BARKING COMPLAINT

VEHICLE STOLEN
VEHICLE PARKED IN VIOLATION OF 72 HOUR CITY ORDINANCE

PROPERTY LOST AND FOUND

BROADCAST INFORMATION (NO UNIT DISPATCH REQUIRED)

IDENTITY THEFT
Priority Level by Call Type
Percent of All Records Percent of Records (within Priority) Highest Priority
Calls
Medium Priority Calls
Lowest Priority Calls
Agenda Report Page 24
10 Units Dispatched
Most calls (46.4%) resulted in the dispatch of a single officer, while two units were dispatched in

34.9% of all incidents. A third unit was assigned in 14.0% of all calls for service from the public.

See Figure 5 for additional information.

Figure 5: Units Dispatched
Information displayed accounts for 99.2% of all records.

Figures 6 & 7 below display the call types by number of units dispatched. 4 The most common
call types with a single unit dispatched were ‘check for evidence of theft/burglary’ (20.4%), ‘animal control call for service’ (11.5%), and ‘parking complaint’ (9.9%). Figure 7 summarizes the calls for service that resulted in more than one unit dispatched and include ‘burglar alarm’ (22.3%), ‘suspicious circumstances’ (14.0%), and ‘ambulance emergency’ (13.2%). 4 Call types accounting for less than 1% of the overall calls for service are not displayed.

46.4%

34.9%

14.0%

2.9% 1.1%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
1 2 3 4 5
Units Dispatched
Percent of Records
Agenda Report Page 25
11 Figure 6: Single Unit Dispatch by Call Type

20.4%

11.5%

9.9%

6.8%

5.4%

3.8%

3.5%

2.8%

2.3%

1.6%

1.5%

1.4%

1.3%

1.1%

1.0%
0% 5% 10% 15% 20% 25%

CHECK FOR EVIDENCE OF THEFT/BURGLARY

ANIMAL CONTROL PIEDMONT CALL FOR SERVICE

PARKING COMPLAINT

PIEDMONT CITY ORDINANCE

AMBULANCE DISPATCHED TO OUTSIDE JURISDICTION
VEHICLE PARKED IN VIOLATION OF 72 HOUR CITY ORDINANCE
THEFT

SUSPICIOUS CIRCUMSTANCES

FIRE ALARM RESIDENTIAL

PROPERTY LOST AND FOUND

OBSTRUCTION IN ROADWAY

VANDALISM

BURGLAR ALARM

SMELL OF GAS/ELECTRICAL BURNING/UNKNOWN ODOR

AUTO BURGLARY

WELFARE CHECK

ANIMAL CONTROL BARKING COMPLAINT

WIRE DOWN (FIRE CALL TYPE)

IDENTITY THEFT
Single Unit Dispatched by Call Type
Percent of Records
Agenda Report Page 26
12 Figure 7: Multiple Units Dispatched by Call Type

22.3%

14.0%

13.2%

10.1%

5.2%

3.2%

3.1%

2.3%

1.9%

1.8%

1.6%

1.3%

1.2%

1.0%

1.0%
0% 5% 10% 15% 20% 25%

BURGLAR ALARM

SUSPICIOUS CIRCUMSTANCES

AMBULANCE EMERGENCY
CHECKING RESIDENCE OR VEHICLE FOR EVIDENCE OF…

WELFARE CHECK

PIEDMONT CITY ORDINANCE

ANIMAL CONTROL PIEDMONT CALL FOR SERVICE

ARGUMENT/YELLING ETC (UNKNOWN RELATIONSHIP)
THEFT

PARKING COMPLAINT

PANIC OR HOLD UP ALARM
HANG UP/NO RESPONSE. NEVER BECAME ANOTHER…

ACCIDENT WITH NO INJURIES

OBSTRUCTION IN ROADWAY
VEHICLE PARKED IN VIOLATION OF 72 HOUR CITY ORDINANCE

BURGLARY OF A RESIDENCE
Mulitple Units Dispatched by Call Type
Percent of Records
Agenda Report Page 27
13 Time
An important component of analyzing PPD’s response to calls for service from the public is assessing the time it takes from receipt of the call until an officer arrives on the scene (time to arrival) and the time spent on scene as calculated from the time an initial officer arrives until the final officer leaves (time on-scene). These two measurements were analyzed in minutes. Figures 8 & 9 below summarize the available data5for time to arrival and time on-scene.

The average time to arrival was 4.1 minutes after removing outliers. 6 Figure 8 reports the
percentage of calls that received a response within 1 minute (28.8%), within 2 minutes (7.6%), and within 3 minutes (10.3%), etc. The blue line reports the cumulative percentage of all calls within each minute. For example, 77.0% of all calls for service had a unit on scene within six minutes.

Figure 8: Time to Arrival
Average time to arrival is based on an upper limit of 1 hour. 5 Fields required to calculate time to arrival were missing date and/or time in 7.9% of all cases. Fields required to create time on scene were missing date and/or time in 5.3% of all cases. 6An outlier was defined as a response time greater than one hour; these incidents accounted for less than one half of one percent of calls. Including those calls resulted in an average time to arrival of 9.2 minutes.

28.8%

7.6% 10.3% 11.7% 10.6% 7.9% 5.1% 3.3% 2.5% 1.9% 1.4%

36.4%

46.7%

58.4%

69.0%

77.0%

82.1% 85.4% 87.9% 89.8% 91.2%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Under 1
minute
1 2 3 4 5 6 7 8 9 10
Time to Arrival in Minutes
Percent of Records Cumulative Percent
Average Time to Arrival:

4.1 minutes
Agenda Report Page 28
14 The average amount of time spent on-scene was 13.1 minutes after removing outliers.7 Figure 9 summarizes the percentage of calls that resulted in one minute or less (7.9%), within two minutes (6.7%), and within three minutes (7.1%), etc. of time on scene. The blue line is a measure of the cumulative records as time on-scene increased. For example, 73.8% of all incidents resulted in between 0 and 16 minutes of on-scene activity by PPD personnel.

Figure 9: Time On Scene
Average time on scene is based on an upper limit of 2 hours.

Finally, Figure 10 reports the average time to arrival and time on-scene based on the highest, medium, and lowest priority calls. The time to arrival analysis reveals that the highest priority calls had a response time of under 4 minutes, while the medium and lowest priority calls experienced longer response times. Recall that the highest priority calls comprised nearly threequarters of all calls for service. As a result, the average response time for high priority calls (3.9 minutes) is similar to the average for all calls (4.1 minutes). Interestingly, the average time onscene was relatively similar across the highest (19.2 minutes), medium (16.4 minutes), and lowest (18.0 minutes) priority levels.
7An outlier was defined as a response time greater than two hours; these incidents accounted roughly one half of one percent of calls. Including those calls resulted in an average time to arrival of 18.5 minutes.

7.9% 6.7% 7.1% 6.9% 6.8% 6.5% 5.5% 4.7% 4.0% 3.6% 3.1% 2.7% 2.5% 2.1% 1.9% 1.8%

14.6%

21.7%

28.7%

35.4%

42.0%

47.5%

52.2%

56.2%

59.7%62.8%65.5%68.0%70.1%72.1%73.8%
0%
10%
20%
30%
40%
50%
60%
70%
80%
Time On Scene in Minutes
Percent of Records Cumulative Percent
Average Time On Scene:

13.1 minutes
Agenda Report Page 29
15 Figure 10: Priority Call Type by Average Arrival Time & Time On Scene Dispositions
Calls for service data were also analyzed to determine the most common methods of completing or clearing the incident (i.e., disposition) (Figure 11). Slightly more than one-third (34.1%) of all incidents were resolved in the category of ‘other’. The next most frequent categories included ‘report taken’ (11.3%), ‘no report taken’ (11.2%), ‘false alarm’ (10.3%), ‘gone on arrival/unable to locate (GOA/UTL)’ (9.0%), and ‘fire incident’ (7.9%). Of note, warnings were issued in roughly 1.8% of all incidents (1.1% - ‘verbal’; 0.7% - ‘written’), a ‘citation’ was issued in 1.7% of all incidents, a ‘field interview card’ was completed in 1.6% of all incidents, and an arrest resolved the incident in 0.3% of all calls for service.

Dispositions were also analyzed in relation to priority level of the call. Figure 12 below summarizes the most frequent dispositions within each priority level. The highest priority calls were most frequently concluded with an ‘other’ designation (33.1%), a ‘false alarm’ (14.4%), ‘no report taken’ (10.9%), or ‘fire incident’ (10.8%) disposition. Similarly, medium priority calls were resolved with an ‘other’ designation (40.8%) or ‘report taken’ (22.4%) disposition. Finally, lowest priority calls for service were most frequently concluded with a ‘marked for tow’ (27.1%), ‘other’ (25.5%), or ‘report taken’ (24.8%) disposition.

3.85 19.19

24.75 16.41 14.83

17.97 0
5 10
15 20
25 30
Time to Arrival Time On Scene
Priority by Time to Arrival & On Scene (in minutes)
Highest Medium Lowest
Agenda Report Page 30
16 Figure 11: Dispositions

34.1%

11.3%

11.2%

10.3%

9.0%

7.9%

2.5%

2.2%

2.1%

1.7%

1.6%

1.4%

1.1%

1.0%

0.7%

0.4%

0.3%

0.2%

0.1%
0% 5% 10% 15% 20% 25% 30% 35% 40%
Other
Report Taken
No Report Taken
False Alarm

GOA/UTL
Fire Incident
Marked for Tow
Patient Transport
Call Unfounded
Citation Issued
Field Interview Card
False Alarm (Fire)
Verbal Warning
Patient Assist
Written Warning
Patient Refusal
Beat Information Only
Parental Refusal
Arrest
Supplement Report
Civil Matter
Dispositions
Percent of Records
Agenda Report Page 31
17 Figure 12: Call Priority Levels by Disposition

33.1%

14.4%

10.9%

10.8%

7.9%

6.6%

3.1%

2.2%

2.1%

40.8%

22.4%

14.6%

11.1%

3.0%

2.9%

27.1%

25.5%

24.8%

13.4%
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%
Other
False Alarm
No Report Taken
Fire Incident

GOA/UTL
Report Taken
Patient Transport
Citation Issued
Field Interview Card
Other
Report Taken

GOA/UTL
No Report Taken
Call Unfounded
Verbal Warning
Marked for Tow
Other
Report Taken
No Report Taken
Priority Level by Disposition
Percent of Records
Highest Priority Calls
Medium Priority Calls
Lowest Priority Calls
Agenda Report Page 32
18 Discussion
A primary goal of this analysis8 was to help inform city and PPD decision-making regarding when to dispatch an officer and/or whether certain call types might be better handled by nonpolice or other professional personnel. An examination of the most frequent call types (Figure 3) reveals how police resources are currently being deployed across different types of calls for service. Some call types are not high priority level calls (Figure 4) and likely could be handled by non-sworn personnel such as the current community service officers, while others seem misplaced as high or medium priority calls (e.g., parking complaints; guidance on non-criminal matters). Below is a list of call types that might be amenable to a non-sworn police response:
• Parking complaints
• Welfare checks
• Roadway obstructions
• Lost and found property
• Vandalism
In addition, reports of completed property crimes (auto burglary, theft, etc.) with no suspects nearby are often taken over the phone or in-person by non-sworn personnel in many cities.

Whether such an approach is appropriate for Piedmont is, of course, a local decision, but the nature and distribution of calls suggest options to reduce costs (non-sworn personnel are typically less expensive to train and maintain) to the city and to lessen the footprint of sworn police officers in the community. In the wake of George Floyd’s death and efforts in cities across the U.S to re-imagine policing, some communities have actively sought to reduce the number of contacts between police officers and civilians, which may or may not be desirable in Piedmont.

Additional background and recommendations for handling mental health-related calls and domestic violence incidents can be found in the Summary & Recommendations section of the report below. For now, we simply note the PPD handled only 85 mental health calls and 27 domestic violence calls (collectively, these calls represent less than 0.4% of total calls) over the five-year analysis period. Thus, it may make little fiscal or operational sense to hire permanent personnel to respond or co-respond to these calls with the PPD.

Racial/Ethnic Comparisons
The second research goal was to assess the alignment of civilian race/ethnicity appearing in publicly-generated police records (i.e., CFS and crime reports) with the race/ethnicity recorded in police-generated activity (i.e., FI Cards and arrests). These analyses required all four data sources and the identification of the racial/ethnic composition of each.

Identifying race/ethnicity in the CFS narratives required an intricate process of extracting specific words and phrases used by call-takers to record the physical descriptions of suspects 8 Due to data limitations, no analyses were performed on how calls change from initial dispatch through disposition.

Agenda Report Page 33
19 when known or observed by callers. Appendix B provides examples of the codes commonly used by call-takers to record race/ethnicity. We developed an automated script that we applied to all narratives to flag the mention of suspect race/ethnicity in the call narratives for subsequent analysis.

To check for the accuracy of this script in correctly identifying suspect race/ethnicity in the CFS narratives, two checks were applied. First, a false positive analysis examined the first 100 records that were flagged by the script as possibly containing references to suspect race/ethnicity.

A second check involved an analysis for false negatives which involved a random check of all narratives that were not identified as containing race/ethnicity to confirm they did not contain any relevant reference to a specific racial or ethnic group. Table 2 reports the results of these two checks, which showed a false positive rate at less than 3% and a false negative rate at less than 4%; a normal statistical threshold for identifying statistical significance is 5%. As a result, the automated script accurately identified the presence (or lack thereof) of a reference to race/ethnicity at a highly robust level, and we are confident that our procedure correctly identified caller descriptions of suspect race/ethnicity as recorded by PPD call-takers.

Table 2: Accuracy Check
False Positive False Negative
White 0.0% 0.0%
Black 2.0% 0.7%
Hispanic 0.0% 4.0%
Asian 3.0% 0.2%
Descriptives
Using the two publicly-generated data sources (i.e., CFS and Suspects) and the two data sources reflecting police activity (i.e., FI Cards and Arrests), comparisons of race/ethnicity within these data were undertaken. Table 3 reports the original number of records in each data set and the associated number of records that included race/ethnicity. The high percentage of CFS that do not contain a reference to a specific race/ethnicity is not surprising as many calls were unrelated to a specific person(s). Similarly, the suspect data did not always contain a reference to a suspect’s race/ethnicity. As expected, the police activity data contain very limited missing information on race/ethnicity. Note that all subsequent analyses are based on records that contain a reference to race/ethnicity.

Agenda Report Page 34
20 Table 3: Summary of Records
Data Period All Records
Records with
Race/Ethnicity
% Missing
Race/Ethnicity
CFS Jan 2017-Dec 2021 23,5479 1,842 92.2%
Suspects Jan 2016-Oct 2021 875 558 36.2%
FI Cards Jan 2016-Oct 2021 1,232 1,195 3.0%
Arrests Jan 2016-Oct 2021 1,482 1,423 4.0%
Figures 13 & 14 below summarize the percentage of records in the four data sets that contained descriptions of White, Black, Hispanic, or Asian civilians; also see Appendix C for a table summary. The figures are organized to report the percentage in each racial/ethnic group across each data source. For example, White civilians were mentioned or appear in 45.5% of calls for service records (red bar), 33.7% of suspect records (blue bar), 38.2% of FI card records (green bar), and 30.8% of all arrestees (gray bar). Likewise, Black civilians were mentioned or appeared in 38.8% of CFS records and 43.7% of arrest records. Hispanic civilian representation ranged from a low of 15.1% in FI card records to a high of 22.8% in suspect records. Finally, Asian civilians most frequently appeared in the FI card records (6.4%) and least often in suspect records (4.2%).
9 Narratives were not available for 6,567 of the original 30,114 calls for service.

Agenda Report Page 35
21 Figure 13: Comparison of Race/Ethnicity Across Data Sources

45.5%

38.8%

16.3%

6.1%

33.7%

39.2%

22.8%

4.2%

38.2%

40.3%

15.1%

6.4%

30.8%

43.7%

21.0%

4.5%
0%
10%
20%
30%
40%
50%
60%
White Black Hispanic Asian
Comparison of Publicly Generated Information to Police Activity Calls For Service (N=1,842) Suspects (N=543) FI Cards (N= 1,370) Arrests (N=1,166) Agenda Report Page 36
22 Figure 14 further examines these data by analyzing the racial/ethnic composition of civilians described by callers in ‘suspicious circumstances’ or ‘check for evidence of theft/burglary’ calls. These two call types were chosen for further analysis because they often contained suspect descriptions provided by callers, and they represent frequent call types received by the PPD. White, Black, and Hispanic groups were represented in these call types at rates equal to or higher than their representation across all call types. In contrast, Asian civilians were represented at lower rates among these calls when compared to all other types of calls.

The green and gray bars within each racial/ethnic group provide a visual means for comparing the racial composition of suspects identified by callers in these two types of calls (red shaded bars) to the racial composition of those stopped by the PDD where an FI card was completed (green bars) and to those arrested by the PPD (gray bars). Further comparisons are described below.

Figure 14: Specific CFS Types

45.5%

42.0%

16.9%

2.3%

47.7%

39.3%

18.7%

3.7%

38.2% 40.3%

15.1%

6.4%

30.8%

43.7%

21.0%

4.5%
0%
10%
20%
30%
40%
50%
60%
White Black Hispanic Asian
Comparison of Specific CFS to Police Activity
CFS: ‘Suspicious Circumstances’ (N=686) CFS: ‘Checks for theft/burglary’ (N=214) FI Cards (N= 1,370) Arrests (N=1,166) Agenda Report Page 37
23 Disproportionality Indices
A Disproportionality Index (DI) statistic was calculated to directly compare the four data sources. It varies around a value of 1.0, which indicates the degree of alignment between the rates of contact with the racial/ethnic groups as a result of PPD-initiated activities (FI card and arrest data) and the rates of contact with those groups based on the publicly-generated data (CFS and suspect data). A value above 1.0 indicates that the PPD engaged with the group of interest at higher levels than might be expected based on the publicly-generated data, whereas a value below 1.0 denotes that the PPD contacted that racial/ethnic group less frequently than what appears in the publicly-generated data.

Figure 15 summarizes the DI rates for each group. The dark green bar compares FI card records to CFS cases, the light green bar assesses FI cards in relation to suspect records, the black bar compares arrestees to CFS cases, and the gray bar examines arrestees in relation to suspect records. With some variation, most comparisons hovered at or near 1.0, suggesting a high degree of alignment between self-initiated PDD activities and suspect representation by race/ethnicity in calls for service and crime reports. Six of the 16 comparisons were above 1.0, indicating that the rate of race/ethnicity in the police activity data exceeded the representation in the publicly-generated data. For example, stops of White civilians were slightly higher in the FI card data compared to the crime suspect data (1.1), and arrests of Black civilians slightly exceeded the CFS (1.1) and suspect data (1.1). Other variances were evident for Hispanic (arrests vs. CFS) and Asian civilians (FI cards and arrests vs. crime suspects). Overall, the data showed that the PDD stopped (as represented by FI cards) and arrested racial groups at rates that closely matched the racial composition of suspects described by callers and in official police reports.

Agenda Report Page 38
24 Figure 15: DI - Comparison of Publicly Generated Information to Police Activity

0.8 1.0

0.9 1.0

1.1 1.0

0.7 1.5

0.7 1.1

1.3 0.7

0.9 1.1

0.0 0.5

1.0 1.5

2.0 White Black Hispanic Asian
Comparison of Publicly Generated Information to Police Activity:

Disproportionality Indices
FI Cards-CFS DI FI Cards-Suspects DI Arrests-CFS DI Arrest-Suspects DI Agenda Report Page 39
25 Figure 16 extends the DI analysis by examining the ‘suspicious circumstances’ and ‘check for evidence of theft/burglary’ CFS subtypes. The dark red bar compares FI card records to CFS-suspicion cases, the light red bar assesses FI card records in relation to CFStheft/burglary cases, the dark blue bar compares arrest records to CFS-suspicion cases, and the light blue bar examines arrest records in relation to CFS-theft/burglary cases. These comparisons show moderate under-representation of Whites in police-initiated actions (FI card stops and arrests) compared to suspect descriptions provided by civilians and relatively close alignment among the data source comparisons for Blacks and Hispanics. Asian civilians were significantly over-represented among those stopped (i.e., FI cards) and arrested compared to caller-identified suspects in suspicion and burglary/theft calls. Overall, seven of the 16 comparisons were above 1.0, indicating that the rate of contacts (i.e., FI cards) and/or arrests in the data for these groups exceeded what might be expected based on the publicly-generated CFS data for specific calls. Again, these variances were mostly concentrated among Asian civilians.

Figure 16: DI - Comparison of Specific CFS to Police Activity

0.8 1.0 0.9

2.8 0.8

1.0 0.8

1.7 0.7

1.0 1.2

2.0 0.6

1.1 1.1 1.2

0.0 0.5

1.0 1.5

2.0 2.5

3.0 White Black Hispanic Asian
Comparison of Specific CFS to Police Activity:

Disproportionality Indices
FI Cards-CFS: Suspicion DI FI Cards-CFS: Theft/Burglary DI Arrests-CFS: Suspicion DI Arrests-CFS: Theft/Burglary DI Agenda Report Page 40
26 Discussion
The analyses represented in this section compare the rates at which the PPD stopped (based on FI card entries) and arrested individuals by race/ethnicity to the racial composition of suspects identified by callers or persons who made official crime reports to the PPD. They are designed to address the question of whether the PPD disproportionately stopped or arrested certain racial or ethnic groups relative to the descriptions of suspects they received from callers or victims of crime. These analyses should be interpreted with caution and should not be read as definitive evidence for the over or under-representation of various racial or ethnic groups in policeinitiated activities (stops and arrests) compared to those groups’ involvement in suspected criminal activity. Simple comparisons such as these do not take into account a wide range of individual, legal, situational, and environmental factors that may contribute to why the PPD stops and arrests certain individuals or groups of people at higher rates than others (Smith et al, 2021; Smith et al., 2019). The rigorous evaluation of disproportionate stops and stop outcomes (e.g., arrests) requires data that were unavailable to the UTSA research team, and such an evaluation was outside the scope of their research engagement.10
With those caveats in mind, a key question the UTSA researchers were asked to address is whether the PPD contacts people of color more often than might be expected given their representation in the Piedmont community, and if so, whether those stops are the result of policeinitiated activities involving the exercise of discretion or in response to requests for service from the community. According to 2021 Census estimates, Piedmont’s population is approximately 70% White, 1.8% Black, 3.7% Hispanic, and 19.5% Asian. While Figure 13 above indicates that Black and Hispanic civilians were significantly over-represented among those stopped and arrested by the PPD compared to their residential populations in Piedmont, their representation in these police-initiated activities closely matches their representation among suspects reported by callers and crime victims to the PPD. Thus, PPD stops and arrests are closely tied to community reports and do not show a pattern of disproportionate, police-initiated activities targeting minority groups in Piedmont. By themselves, and with the previously stated caveats in mind, the analyses shown above do not provide evidence that the PPD disproportionately stopped or arrested Black or Hispanic individuals compared to the community-generated comparison data.

Furthermore, while the findings for Asians show somewhat greater variability in the DIs than among the other minority groups (Figure 15), they are not markedly different except for in the specific call-type comparisons (Figure 16). Here, Asians are consistently, and in a few cases, 10 There is a robust scientific literature on racial disparities in police stops (traffic or pedestrian stops) of civilians and the enforcement outcomes (e.g. arrests or searches) that sometimes take place during or after those stops. For recent reviews, see Pryor et al., 2020 and Smith et al., 2017. The benchmarking dilemma, which refers to identification of an appropriate at-risk population to compare against police stop data, has been the subject of much research and scholarly writing. Today, most social scientists who regularly work with police stop data and publish in the peer-reviewed literature reject the use of census data as a scientifically valid estimate of the population at-risk for contacts with or stops by the police (Smith et al., 2021; Smith et al., 2019; Tillyer et al., 2010).

Agenda Report Page 41
27 significantly over-represented among persons stopped and arrested by the PPD when compared to suspect descriptions provided by civilian callers reporting ‘suspicious circumstances’ or requesting the PPD to ‘check for evidence of theft/burglary’. Additional research and more comprehensive data11 are needed to better understand the stop and arrest patterns among all racial and ethnic groups in Piedmont, but especially among Asian civilians. 11 PPD will be required to comply with the data collection requirements of the California Racial and Identify Profiling Act (AB 953) by April 1, 2023. https://post.ca.gov/Racial-and-Identity-Profiling-Act. Stop data collected pursuant to this Act will, over time, provide Piedmont with the information needed to further investigate possible racial or ethnic disparities in officer-initiated stops conducted by the PPD.

Agenda Report Page 42
28 FOCUS GROUPS
In late September 2021, the UTSA research team led three focus groups of Piedmont residents to help gauge community sentiment on topics and issues related to the overall goals of the research project. These groups were comprised of 5-7 residents who were selected by City leadership to broadly represent the community and its various constituencies. One of the groups was specifically focused on Piedmont schools and was made up of parents and school employees.

The focus group interview guide can be found in Appendix D. While these questions served as a guide for the researchers, the scripts varied somewhat between groups based on the flow of the conversations and the input received as community members shared their insights and concerns.

The focus groups were conducted via Zoom, which, with the agreement of all participants, allowed the research team to record the conversations and subsequently transcribe them for analysis. Participants were randomly assigned participant numbers and were asked to refer to themselves and other participants by their numbers rather than by their names. This allowed for the anonymous transcription of the conversations, which typically ran more than 20 pages in length for each group. Once transcribed, the focus group data were content analyzed using an inductive thematic coding approach (Guest et al., 2012). Because the focus groups loosely followed the structure provided by the interview guide, the resulting data were naturally grouped and nested within the questions posed. At the same time, focus groups tend to be conversations that ebb and flow and sometimes diverge, as they should, from more structured interviews.

Careful inductive analysis allows for the identification of key themes across focus groups even when expressed thoughts are not directly tied to questions posed by the researchers. Several such organic themes emerged from the data. Below, we summarize the results from our analysis of the focus group data and, where appropriate, we provide quotes from respondents that help illustrate the themes that we identified. When needed to allow for a clearer understanding of the ideas being expressed, we occasionally add contextual words or phrases in brackets to the quotations.

Otherwise, all quotes are taken directly from the transcribed recordings.

Following the inductive methodology we used for the analysis, the results below are organized thematically rather than being presented in a question-response format. This approach better reflects the flow of the conversations and allows for the identification of cross-cutting themes that were not necessarily tied to discrete questions posed by the researchers.

Community Focus Groups
Types of Calls Made to the Piedmont Police Department (PPD) The research team was interested in community perceptions about the types of calls made to and received by the PPD. Three themes emerged. Participants consistently identified noise Agenda Report Page 43
29 complaints, and particularly noise caused by gas-powered leaf blowers, as a frequent call made to the PPD. One participant put it like this:
“I think we get a lot of complaints around noise in general, so if somebody is talking past 9pm you'll get some people that are calling and complaining about that. The gas blowers, we do have a lot of complaints [about them], so those types of things.” Calls about suspicious people were another consistent theme identified in the data. Most respondents spoke in general terms about persons who appear out of place as a source of frequent calls to the police. This theme is represented in the responses from these two participants:
“I think that it's mostly calls about suspicious people, maybe not even suspicious people [but people] that quote, ‘don't belong’".
“Often in the paper its somebody sitting in their car, so again someone who kind of doesn't belong there, but they're sitting in their car outside of a house and they've been sitting there a long time.”
One respondent, though, specifically identified people of color as frequent subjects of calls to the PPD:
“My impression is that people of color, I mean black and brown people that come through Piedmont, are considered outsiders. People call the police on them, whether they're walking through, whether they're in their cars,sitting in their cars at the park.

But I think people know not to be explicit about it. I think we've learned to be, you know, very subtle in what we say when we call the police, but I think that happens a lot.”
The perceived social undesirability of such calls is explicit in this participant’s statement, and that theme repeated itself in other contexts. Participants clearly felt that calling the police about “suspicious” minority group members was undesirable, yet some felt it is commonplace in Piedmont.

Finally, a number of respondents identified property crimes, especially those involving stolen cars or car break-ins, are common reasons for calling the police in Piedmont.

Source of Calls to the PPD
When asked whether most people place calls to the PPD using 911 or the PPD non-emergency number, most participants either were not sure or believed that most Piedmont residents know the non-emergency number and likely use it to call the police when needed.

Calls the PPD Should Not Respond To
A primary purpose of this research study is to help the City of Piedmont identify calls or call types for which a police response may not be desirable or for which other agencies or Agenda Report Page 44
30 professionals may be better suited to respond. We asked the focus group participants a series of initial and follow-up questions designed to probe community sentiment about this important subject, one that many cities across the United States are grappling with in the wake of George Floyd’s death in May 2020.

Suspicious Persons
Interestingly, while most respondents believed that suspicious person calls are among the most frequently made to the PDD, a number of participants also felt that the PPD should not respond to such calls, or alternatively, should require clear evidence of observed behavior indicative of a crime or public safety concern before responding. This point of view is reflected in the answers we received from these two respondents:
“Calls that are about people that don't belong; suspicious people where the caller cannot articulate a reason for them being suspicious. Yes, I think those should not be responded to by a sworn police officer [or] by anyone.” “I think that there have been articles in the paper about people just seeing an individual who is a person of color and calling the police when they're just, you know, sitting doing nothing. So those type of what we call biased by proxy type calls forservice [should not be responded to] by the police.” That point of view was not universal, though, and some participants expressed a countervailing view:
“There are going to . be situations where it's not a matter of profiling or where there's a genuine risk perceived regardless of color. And so those, you know, those may be few and far between versus just people reacting inappropriately, but that's going to be a problem if, you know, they don'trespond and someone said, ‘I called the police’ [but no one came].”
“I mean, you know, ‘see something, say something,’ you know? I mean we have to be the eyes and ears of the town. And, you know, things happen quickly, and Piedmont is a little different than Oakland because, quite honestly, they are not a lot of criminals who live in this town. The problem is the criminals coming into the town. That's the problem. And, so you know, we need to help the police filter those people out before there's a problem.”
While some community members felt that their neighbors are too quick to call the police about people who look “out of place” in Piedmont, others expressed gratitude for living in a city with a responsive police department that has the resources to respond quickly to signs of potential trouble. They preferred that the PPD err on the side of caution by responding quickly to potential threats from outside the city.

Agenda Report Page 45
31 Mental Health-Related Calls
Especially with the onset of the Covid-19 pandemic, communities nationwide are grappling with how to better respond to those experiencing mental health crises. The default response of sending the police (or solely the police) to mental health-related calls has been called into question, and alternative response models have been developed and adopted in many communities (Puntis et al., 2018; Seo et al., 2021; Shapiro et al., 2015).

The focus group participants we spoke with unanimously agreed that police officers are not the ideal first responders to calls involving persons with mental health problems, yet a number of them also stated that mental health-related calls seemed to be increasing during the Covid-19 pandemic. One respondent stated that in the case of people suffering from mental health problems, it is “clearly obvious . . . the cops don’t really bring the expertise to the situation.” Another participant stated that “A little mental health team would be really helpful.”
On a cautionary note, some participants drew a distinction between sending mental health professionals to respond to non-violent persons in mental health-related distress and those who are aggressive or who may pose a risk to others. In her recent analysis of police dispatch data from nine U.S. cities, Lum and her colleagues (2021) found that while, on average, less than two percent of calls responded to by the police involved mental-related concerns, such calls can result in an increased risk for the use of force and officer injury.

That said, some cities are experimenting with sending mental health professionals and/or medics to most mental health calls instead of sending the police (Dholakia & Gilbert, 2021; Mills, 2021). Most such models have not yet been evaluated to assess, among other things, the risks to the mental health professionals, family members of those in crisis, or others involved in the calls of not sending the police as first or co-responders.

Nuisance Calls
A number of participants felt that the PPD should not routinely respond to calls that can be characterized as “nuisances” – leaf blower noise complaints, barking dogs, or loud music – are examples that were provided.

Calls from Schools
Several respondents expressed reticence at having the PPD respond to most calls at Piedmont schools that involve students. For example, one participant put it this way: “You know, with the exception of a shooter or violence or things like that, they [calls] really should not be handled by a sworn police officer.” Several others invoked the Education Code and stated that police should not respond to a call from a school unless a response is required by the ED code. The overarching theme is summed up by this response from one participant:

Agenda Report Page 46
32 “Most of the school calls can be handled by somebody with a health and wellness perspective and grounded in that framework rather than law enforcement.” In a separate section below, we summarize the findings from the school-based focus group.

We did not hear similar concerns from that group regarding PPD response to student-related matters, although we note that the group contained several employees from various schools in Piedmont.

Finally, one cross-cutting theme we identified dealt with the intersection between students and mental health and is reflected in this statement from one of the focus group participants:
“Ithink what has come through is our kids are in crisisin the middle and high schools here in the city, and itseems especially important to not to call the police and to have somebody who's trained in working with kids and adolescents and kind of deescalating the situation and maybe also with substance use issues.” Here, in this comment, the respondent is expressing concerns about child and adolescent mental health and substance abuse and the preference for a response to such matters in the schools by persons specifically trained to deal with them rather than by the police.

Perception of Crime and Public Safety in Piedmont
All of the focus group participants who spoke about this topic agreed that crime in Piedmont is generally low compared to many communities and certainly lower than in Oakland. A consistent theme among participants was that crime in Piedmont is often committed by people from outside the city. Below are some of the specific crime and public-safety related concerns expressed by the respondents.

Home Invasions
Several respondents recalled some home invasion robberies that occurred in Piedmont a few years ago and expressed concern about them: “I am very concerned about the number of home invasion-type robberies Piedmont has had over the years.” Another respondent described these as a “rash” of home invasions and characterized them as “very scary” and “very serious and violent.”
Drug Use
Two respondents mentioned drug sales or use as a problem in Piedmont. One person stated that “drug use and drug selling is a major problem, [especially] in the schools and parks.” Public Safety-Related
A common theme among respondents was that unsafe driving is a problem in Piedmont due to the nature of the “curvy” roads. And one participant mentioned urban fires as a potential safety concern, especially with the increase in forest fires in the Bay area in recent years.

Agenda Report Page 47
33 Disproportionate Contacts with Non-White Civilians and Bias by Proxy One area of research interest to city leaders is whether calls for service to the PPD may result in disproportionate contacts between PPD officers and non-White civilians. As a community-driven law enforcement agency, the PPD is responsive to calls from community residents. If some calls for service, especially those involving suspicious persons or activity, disproportionately involve non-White subjects, then police-civilian contacts may be skewed in a manner that diverges from the demographic make-up of Piedmont itself, which is largely White. To explore this issue, the UTSA researchers asked focus group participants a series of questions about their perceptions of the residency of “typical” offenders and/or suspicious persons and whether non-residents were more likely to draw the attention of the community than those who appeared to live in the city.

Perceptions of “Typical” Offenders
A majority of the focus group participants believed that most crime problems in Piedmont that come to the attention of the PDD are caused by non-residents. They described these individuals as mainly coming from Oakland and stated that most appear to be young males.

Several mentioned groups of males in their 20s, and one person stated that from what s/he saw in the newspaper and on Nextdoor posts, those who commit most of the crime in Piedmont are young, Black males from Oakland. However, most of the respondents did not identify “typical” offenders as being of a certain race or ethnicity; rather, they identified them as young (20s or adolescents) and male.

Calling the Police About Non-Residents
When respondents were asked whether they would call the PDD about someone walking in their neighborhoods who did not appear to be from Piedmont, everyone who responded said “no.” When the question was posed differently and participants were asked whether they thought some of their neighbors would call the police under those circumstances, one respondent stated “I’ve had neighbors that have called based on somebody’s race,” and another agreed that “I know that’s happened, too.”
When asked if people in Piedmont would be more suspicious of someone looking into cars or knocking on doors if the individual appeared to be a non-resident of the city, all respondents agreed and stated: “Absolutely,” “I second that. Absolutely,” “Same,” and “Same.”
Police Response to Calls about Persons of Color
During one focus group, the UTSA researchers asked participants how the PPD should handle or respond to calls received about non-White persons acting suspiciously. Several respondents stated that PPD call-takers should ask additional questions about specific behaviors or mannerisms that suggested possible criminal intent by the subject of the call.

One respondent flatly stated “I would not want them [PPD] to respond unless they [callers] Agenda Report Page 48
34 could articulate criminal behavior,” while another stated “It’s almost like there needs to be a checklist [that could be used by the call-takers/dispatchers]. However, another respondent identified the inherent tension between being responsive to community members who pay high taxes and expect high levels of service and making callers answer a series of questions designed to identify specific behaviors suggestive of criminal intent:
“I think people live here because they expect a response, and they want to see their police presence in their communities. And if you were to say to an elderly couple, for example, or some other people, ‘You know, oh, let me first ask you these 10 questions on my checklist,’ they would blow a gasket. They would want a response more promptly, I would imagine, and this is a very traditional town with, you know, a police department that has been very present. And I think that we want to maintain that. I’d like to see us maintain that , and I’d like to see the community feel like they're getting something for their tax dollars, frankly.”
A Role for Community Education
Following this exchange, the UTSA researchers asked participants if they felt it would be helpful for the PPD to educate the community about how to better distinguish suspicious persons or behaviors which might warrant a call to the police from mere appearance or actions less likely to suggest criminal intent. Here is how the participants responded:
“I think absolutely it's necessary. That absolutely has to happen. I think it would be. . . it's a type of conversation that community should have. This is engaging the community in, honestly,somewhat of an almost better use of resources.” “Just because we have the community resources doesn't mean we should use them as an abuse of power. We need to use them appropriately.” “And if we, you know, if the threshold is aslow as‘Isaw somebody, you know, walk by my house who is black, young and wearing a hoodie, and you know I need help,’ [then] I think what we're trying to do is find more equity for Piedmont.” Racial Bias by the PPD
When asked directly if participants thought the PPD was racially biased in how it responds to the larger community, one respondent stated that s/he thought the PPD had improved a lot over the many years s/he had lived in Piedmont. Another respondent agreed and stated “Yeah, I agree. I think Piedmont has done a great job over the last few years to hire, you know, officers that come from different backgrounds.”
School-Based Focus Group
As an important constituency in Piedmont, city leadership was interested in hearing the perspectives of school-affiliated individuals – teachers, administrators, and parents – regarding the PPD, public safety, and potential sources of bias in police contacts in and around the schools. Consequently, the UTSA researchers facilitated a focus group of five Agenda Report Page 49
35 people affiliated with various schools in Piedmont. Here we report our findings from this group.

Types of Calls to the PPD from the Schools
The focus group participants mentioned five types of calls to the PPD they believed commonly come from the schools:
• Car break-ins
• Unknown or unwanted persons on campus
• Emotional problems/crises with students
• Stolen property
• Vandalism
Respondents generally felt that most calls come from school employees, rather than students or parents, and that most are made using the non-emergency number rather than 911.

Call Types PPD Should Not Be Responsible For
Several participants felt that PPD should not be responsible for responding to most mental or emotional health-related calls at the schools. There was consensus that mental health professionals or counselors are better equipped than the police to deal with such matters.

Police Officers in the Schools
Respondents were split over whether police officers should be assigned to or have a regular presence in the schools. One person stated that “Having a non-police officer respond to some these incidents, I would be concerned as to an increase of activity that is going to occur or escalation.” This individual seemed to express a preference for having PPD offi cers assigned to the schools. Other respondents, though, felt that PPD officers should only be in the schools as a last resort: “In general, you don't want the police in the schools if you can avoid that. You know, if you need it, then you can have it, but you'd rather not have it.” By the same token, most respondents were aware of and seemed satisfied with the current model being used in Piedmont of having a school liaison officer rather than school resource officers present at the schools at all times. Most did not feel it was necessary to have officers on school campuses “because, again, the police are just two minutes away. You know, they are quite responsive.”
Perceptions of Crime in and Around the Schools
Consistent with perceptions of the types of calls generated to the PPD from the schools, focus group participants mentioned vandalism – specifically “tagging” in the bathrooms and other areas of the schools – as a problem. They also mentioned vaping in the bathrooms and Agenda Report Page 50
36 unauthorized persons on the school campuses as possible crime or public safety/health concerns.

One participant commented that “I think Piedmont police do an amazing job of deterring crime.” Perceptions of Unauthorized Persons
The school focus group participants seemed very sensitive to the issue of “bias by proxy” and were reticent to discuss the types of unauthorized individuals who are sometimes present on school campuses. There was some consistency among participants that most unauthorized persons on campuses are from outside the community, but as one respondent put it, “We make a conscious effort, at least I do personally, to not do that [profile by race], and again, it sort of just comes down to instincts and there's no way that we can keep people off the middle school or high school campus because there's dog parks on the other side of the campus.” The school employees seemed to prefer to handle matters at the schools themselves and only to call PPD when absolutely necessary.

Conclusion
Focus group participants consistently told the UTSA research team that crime is generally low in Piedmont. At the same time, some participants expressed concerns about young people from outside the community coming into Piedmont to commit property crimes (e.g., breaking into vehicles) and were appreciative that PPD is responsive to concerns about suspicious persons when called. Other respondents were clearly sensitive to being perceived as racially biased, and some expressed strongly held views that PPD should not respond to calls about suspicious-looking “outsiders” unless callers can specifically describe behaviors or mannerisms that suggest the potential for criminal activity. Finally, there was consensus among the respondents about the need for community engagement and education by the PPD on when residents should call the police about suspicious persons in their neighborhoods, what behaviors to watch out for, and questions to expect from PPD calltakers when residents call to report suspicious behavior. Overall, focus group participants expressed confidence that the PPD does not regularly engage in bias-based policing, although a few conceded that some residents have in the past called the police when they saw young, non-White persons in their neighborhoods who did not appear to be Piedmont residents.

SUMMARY & RECOMMENDATIONS
This project addressed two primary research goals through an analysis of official data and convening focus groups with Piedmont residents. These data sources provide the foundation for several recommendations for consideration by the Piedmont Police Department (PPD) and the City of Piedmont. The first research goal was to investigate the types of calls received by the PPD and to assess their responses to those requests. Most calls were classified as high priority (~75%) and most frequently involved checks for burglary or theft, burglary alarms, and suspicious circumstances. Single unit responses were most common, with an average response Agenda Report Page 51
37 time of approximately four minutes, and calls were most frequently resolved by actions categorized by the PPD as ‘other’12, report taken, or no report taken.

The second goal was to determine the frequency of potentially biased-based calls from the public which lacked specific criminal-related behavior as a basis for police intervention. While available CAD data were not sufficient to provide a direct assessment of this question, data were available to assess the racial composition of PPD arrests and/or stops resulting in the completion of FI cards and the degree of consistency between those stops/arrests and the racial/ethnic composition of persons believed to be involved in suspicious and/or criminal activity as described by members of the public who called the PPD for assistance. As previously mentioned, these analyses should be interpreted with caution and should not be read as definitive evidence for the over or under-representation of various racial or ethnic groups in police-initiated activities (stops and arrests) compared to those groups’ involvement in suspected criminal activity.

With these caveats in mind, most of the 32 different assessments (overall and call type-specific) conducted showed no evidence of disproportionate non-White contacts or arrests when using calls for service and crime suspect information for comparison purposes, while a few showed some slightly elevated disproportionality indices between 1.1-1.3. A more consistent pattern of over-representation was seen in stops and arrests of Asian civilians when compared against suspect descriptions in specific call types (suspicion and evidence of theft/burglary). Additional research and more comprehensive data are needed to better understand the stop and arrest patterns among all racial and ethnic groups in Piedmont, but especially among Asian civilians.

The focus groups added context and civilian perspective to these quantitative findings and revealed that residents maintain a general feeling of safety in Piedmont and overall positive opinions of the police. Residents commented on some concerns about property crime and nuisance problems (e.g., loud noise) but indicated overall satisfaction with how the PPD handled these problems. Respondents consistently stated that non-Piedmont residents were primarily responsible for safety concerns within the city, but several residents indicated that the presence of non-residents alone, particular those of color, should not generate a different response by the PPD. Others expressed appreciation for the responsiveness of the PPD to calls about suspicious persons or circumstances and believed that community members expected this level of service.

However, a number of the focus group participants believed that the PPD should not respond to generalized concerns about suspicious outsiders unless callers could articulate actual behaviors that suggested possible criminal intent. Finally, there was consistent support for community education by the PPD on what behaviors to look for and the questions to expect from PPD calltakers when reporting suspicious events. 12 The “other” disposition is commonly used by PPD to clear officer-initiated activities or for calls where “no report” or “report” are not appropriate (e.g., vacant home checks).

Agenda Report Page 52
38 Recommendations
Drawing from the results, we offer a number of recommendations for consideration by the PPD and the City of Piedmont.

While experiments certainly are underway to replace police officers with mental health and/or emergency medical professionals as responders to non-violent mental health-related calls, these models require substantial new staffing investments to develop or expand availability during nonstandard work hours. Given the relatively few mental health and domestic violence-related calls received by the PPD, providing PPD officers with specialized training in how to more effectively handle conflict between intimate partners or safely manage persons with mental or behavioral health problems may be more cost-effective than hiring full-time social workers or other mental health professionals to take the lead or to co-respond on these calls with PPD officers. That said, a few respondents in the focus groups stated that the police should not be called upon to deal with behavioral health problems at the schools absent a clear threat of violence. Yet with such low call volumes for these types of incidents in Piedmont, it may not be necessary or even possible to hire permanent non-police staff as alternative first responders. If rapid, on-call arrangements with local social workers or mental health professionals could be set up, this might provide a viable co-response capability that Piedmont could adopt for mental health or domestic violence-related calls for service.

Response to Calls for Service
• Apart from mental health calls, there are a variety of call types that may be amenable and/or appropriate for alternative responses by non-sworn officers either within or outside of the PPD. For example, parking complaints, welfare checks, roadway obstructions, lost and found property, and vandalism calls may be addressed without the need for a sworn officer to be dispatched. It may also be efficient and effective to have non-sworn officers take reports on completed property crimes. Any of these adjustments must be considered with safety of the resident(s) and responding entity as a key consideration. While appearing rather innocuous and unlikely to require a sworn officer in response, some situations change rapidly and may unexpectedly require the presence Agenda Report Page 53
39 of a PPD officer to assist. If changes to the current PPD response model are contemplated, alternative and/or co-responders must be provided with appropriate resources and training to ensure their safety and success. Sworn officers have the benefit of extensive training needed to handle many different situations. Piedmont should likewise equip alternative responders with the tools needed to improve outcomes over the current model and should track those outcomes carefully to evaluate their safety and efficacy.
• A key theme of the focus groups was to avoid having police officers respond to mental health situations within the schools. Based on resident comments, the PPD should review its current response approach to these calls and ensure that officers are only dispatched when there is clear evidence of criminal activity.
• Several focus group members opined on the importance of collecting information provided by residents on suspicious persons to disentangle behaviorally-based descriptions of ‘suspects’ from perceptions of ‘risk’ that may not accurately reflect true risk. Some suggested offering residents community education courses or curriculum to improve the ability of residents to disentangle perception from intention when assessing individuals on the street and to prepare community members for the questions they will be asked by PPD call-takers if they call to report suspicious persons or activity.

Data Collection and Storage
• PPD should consider a review of current call types and how they are categorized into priority categories based on analyses of the calls for service data. Some call types currently listed as ‘medium’ priority may be better located in the ‘high’ category and vice versa.
• Current data collection and storage processes do not allow for an assessment of how calls may change over time. This is an important piece of information to fully understand the nature of calls received by the PPD and how often they change during their lifecycles.

For example, a call may originate as a ‘low’ or ‘medium’ call type but escalate to a ‘high’ priority call type. The PPD should review its current CAD system and seek to improve its capability for tracking how calls change and are sometimes re-classified both by type and priority during their lifecycles.
• The PPD would benefit from working with its CAD vendor to develop discrete fields that call-takers could use to record descriptions of suspects provided by callers. For example, standardized fields for race, ethnicity, gender, age, and other physical descriptors would provide much better data for subsequent analysis. Alternatively, standardizing the language/codes used by call-takers and recorded in the narrative records would allow for easier suspect identification and analyses. In particular, ‘shorthand’ codes for race/ethnicity should be aligned to reduce the number of combinations currently being used to describe similar suspects. The current analyses required the creation of a complex Agenda Report Page 54
40 algorithm to identify the race/ethnicity of persons described by callers; standardization of the codes would vastly streamline this process.
• The codes used for indicating that the call has been resolved should be reviewed and updated. Currently, the most frequently used call disposition category is ‘other,’ which presumably represents a number of different actions or resolutions but lacks specificity. It is not clear what type of action was taken in these situations, and further training on appropriate fields to select and/or the addition of new fields to accurately categorize resolutions would benefit the PPD.
• Given California’s adoption of AB953, it is recommended that the PPD review its current RMS and CAD systems to ensure compliance with RIPA and to maximize the value of this effort for the PPD and the City of Piedmont. Given its size, PPD has until April 2023 to become RIPA compliant but began reporting RIPA data to the California Department of Justice in July 2021. There are a number of scientifically accepted approaches to analyzing stop data. These should be reviewed and preparations made by the PPD and City to conduct scientifically appropriate and useful analyses of the new data as it becomes available and not rely solely on state-provided reports, which may not be helpful to the Piedmont community given the City’s unique make-up and context. • Piedmont should evaluate the capabilities of its current CAD system to easily and accurately extract data for analysis. The UTSA research team encountered multiple challenges and delays in attempting to obtain useable CAD data for this project.

Agenda Report Page 55
41 REFERENCES
Dholakia, N. & Gilbert, D. What happens when we send mental health providers instead of the police. Vera Institute of Justice: Think Justice Blog. https://www.vera.org/blog/whathappens-when-we-send-mental-health-providers-instead-of-police.

Guest, G., MacQueen, K.M., & Namey, E.E. (2012). Applied thematic analysis. Thousand Oaks, CA: Sage.

Lum, C., Koper, C.S., & Wu, X. (2021). Can we really defund the police? A nine-agency study of police response to calls for service. Police Quarterly. https://doi.org/10.1177%2F10986111211035002.

Mills, S. (2021, Aug 27). Professionals with mental health training to respond to some 911 calls instead of Madison police. Wisconsin Public Radio. https://www.wpr.org/professionalsmental-health-training-respond-some-911-calls-instead-madison-police.

Pryor, M., Buchanan, K.S., & Goff, P.A. (2020). Risky situations: Sources of racial disparity in police behavior. Annual Review of Law and Social Science, 16, 343-360.

Puntis, S., Perfect, D., Kirubarajan, A., Bolton, S., Davies, F., Hayes, A., Harriss, E., & Molodynski, A. (2018). A systematic review of co-responder models of police mental health ‘street’ triage. BMC Psychiatry, 18, 256. https://doi.org/10.1186/s12888-018- 1836-2.

Seo, C., Kim, B., & Kruis, N.E. (2021). Variation across police response models for handling encounters with people with mental illness: A systematic review and meta-analysis.

Journal of Criminal Justice, 72. https://doi.org/10.1016/j.jcrimjus.2020.101752.

Shapiro, G.K., Cusi, A., Kirst, M., O’Campo, P., Nakhost, A., & Stergiopoulos, V. (2015). Coresponding police-mental health programs: A review. Administration and policy in Mental Health and Mental Health Services Research, 42, 606-620.

Smith, M.R., Rojek, J., Petrocelli, M., & Withrow, B. (2017). Measuring disparities in police activities: A state of the art review. Policing: An International Journal, 40, 166-183.

Smith, M.R., Tillyer, R., Smith, M., & Lloyd, C. (2021). Assessing police stops of pedestrians:

Toward a new generation of benchmarks. Urban Affairs Review, 1-30. https://

10.1177/10780874211016930.

Agenda Report Page 56
42 Smith, M.R., Tillyer, R., Lloyd, C., & Petrocelli, M. (2019). Benchmarking disparities in police stops: A comparative application of 2nd and 3rd generation techniques. Justice Quarterly. https://doi.org/10.1080/07418825.2019.1660395.

Tillyer, R., Engel, R.S., & Cherkauskas, J.C. (2010). Best practices in vehicle stop data collection and analysis. Policing: An International Journal, 33, 69-92.

Agenda Report Page 57
43 APPENDIX A
Call Type Frequencies by Priority Level
Table 4: Highest Priority Call Types Frequencies
CALL TYPE # OF
CALLS
% OF ALL
CALLS

CHECKING RESIDENCE OR VEHICLE FOR EVIDENCE OF
THEFT/BURGLARY 4488 14.9%
BURGLAR ALARM 3830 12.7%
SUSPICIOUS CIRCUMSTANCES 2744 9.1%
AMBULANCE EMERGENCY 2215 7.4%
PARKING COMPLAINT 1664 5.5%
WELFARE CHECK 1024 3.4%
AMBULANCE DISPATCHED TO OUTSIDE JURISDICTION 772 2.6%
FIRE ALARM RESIDENTIAL 593 2.0%
OBSTRUCTION IN ROADWAY 506 1.7%
ARGUMENT/YELLING ETC (UNKNOWN RELATIONSHIP) 496 1.6%
HANG UP/NO RESPONSE. NEVER BECAME ANOTHER INCIDENT
TYPE 352 1.2%
VANDALISM 344 1.1%
AUTO BURGLARY 298 1.0%
PANIC OR HOLD UP ALARM 292 1.0%
SMELL OF GAS/ELECTRICAL BURNING/UNKNOWN ODOR 241 0.8%
BURGLARY OF A RESIDENCE 219 0.7%
WIRE DOWN (FIRE CALL TYPE) 178 0.6%
OUTDOOR WATER PROBLEM 149 0.5%
RECKLESS DRIVER 144 0.5%

RANDOM RESPONSE THAT DOESNT FIT UNDER OTHER CALL
TYPES 137 0.5%
FIRE ALARM COMMERCIAL 115 0.4%
TRESPASSING 95 0.3%
F SMOKE INVESTIGATION 91 0.3%
MENTAL HEALTH HOLD 85 0.3%
THREATS 76 0.3%
ACCIDENT UNKNOWN IF INJURIES 74 0.2%
INSIDE WATER PROBLEM 65 0.2%
ASSAULT 45 0.1%
DRIVING UNDER INFLUENCE 44 0.1%
RESIDENT LOCKED OUT OF RESIDENCE/RUNNING VEHICLE 38 0.1% HAZMAT CALL 34 0.1%
VEHICLE OR RESIDENTIAL LOCKOUT (POLICE CALL TYPE) 27 0.1% Agenda Report Page 58
44 ELDER ABUSE 27 0.1%
DEATH INVESTIGATION 27 0.1%
FAMILY DISTURBANCE OR CLASSIC DOMESTIC VIOLENCE 27 0.1% SMOKE/CO DETECTOR PROBLEMS 27 0.1%
STRUCTURE FIRE 26 0.1%
CUSTODY ORDER VIOLATIONS/RESTRAINING ORDER VIOLATIONS 24 0.1% ROBBERY 24 0.1%
BURGLARY OF A BUSINESS 21 0.1%
CHILD ABUSE 21 0.1%
AMBULANCE OTHER 20 0.1%

MEDICAL CALL WHERE PATIENT SELF TRANSPORTED TO POLICE
OR FIRE DEPT 18 0.1%
SUICIDE ATTEMPT 16 0.1%

FIRE APPARATUS DISPATCHED TO WORKING EVENT (NOT
STRUCTURE FIRE) OUTSIDE JURISDICTION 15 0.0%
REPORTS OF SHOTS FIRED 15 0.0%
RESIDENT STUCK IN RESIDENCE/GARAGE/VEHICLE 12 0.0%
HARASSING OR ANNOYING PHONE CALLS 11 0.0%
GRASS FIRE 9 0.0%
PROWLER 9 0.0%
GRASS FIRE 8 0.0%

FIRE APPARATUS DISPATCHED TO LARGE FIRE AS PART OF
STRIKE TEAM 8 0.0%

FIRE APPARATUS DISPATCHED FOR STATION COVERAGE OUTSIDE
JURISDICTION 7 0.0%
SEX OFFENSES 7 0.0%
VEHICLE FIRE 6 0.0%
WELFARE & INSTITUTION EVENTS OTHER THAN 5150 6 0.0%
FIRE ALARM-MED ALERT 6 0.0%
WEAPONS 4 0.0%
CARJACKING 3 0.0%

FIRE APPARATUS DISPATCHED TO WORKING STRUCTURE FIRE
OUTSIDE JURISDICTION 3 0.0%
ARSON 2 0.0%
WIRE DOWN (POLICE CALL TYPE) 1 0.0%
KIDNAPPING 1 0.0%
Table 5: Medium Priority Call Types Frequencies
CALL TYPE # OF
CALLS
% OF ALL
CALLS
ANIMAL CONTROL PIEDMONT CALL FOR SERVICE 2113 7.0%
PIEDMONT CITY ORDINANCE 1471 4.9%
Agenda Report Page 59
45 THEFT 836 2.8%
ACCIDENT WITH NO INJURIES 277 0.9%

GIVING GUIDANCE OR HELPING MEDIATE NON-CRIMINAL
MATTERS 233 0.8%
ACCIDENT HIT AND RUN 229 0.8%
ANIMAL CONTROL BARKING COMPLAINT 201 0.7%
VEHICLE STOLEN 192 0.6%
FRAUD 162 0.5%
AMBULANCE CALL TO NON EMERGENCY SITUATION 122 0.4%
PARTY 113 0.4%
JUVENILE MATTERS/JUV SEXUAL ASSAULT 97 0.3%
ANIMAL CONTROL DOG BITE 77 0.3%
FIREWORKS 70 0.2%
DRUG RELATED CALLS 33 0.1%
DRUNK IN PUBLIC 16 0.1%
ILLEGAL DUMPING 5 0.0%
JUVENILE CURFEW VIOLATON 2 0.0%
VEHICLE REPOSESSION EVENT 2 0.0%
Table 6: Lowest Priority Call Types Frequencies
CALL TYPE # OF
CALLS
% OF ALL
CALLS
VEHICLE PARKED IN VIOLATION OF 72 HOUR CITY ORDINANCE 923 3.1% PORPERTY LOST AND FOUND 483 1.6%
IDENTITY THEFT 161 0.5%
BROADCAST INFORMATION (NO UNIT DISPATCH REQUIRED) 146 0.5% MISSING PERSON 96 0.3%
VEHICLE IDENTIFICATION NUMBER CHECK 58 0.2%
PROPERTY FOR DESTRUCTION 48 0.2%
NON CRIMINAL POLICE REPORT 37 0.1%
PROPERTY FOUND 21 0.1%
RESTRAINING ORDER 3 0.0%
FORGERY 1 0.0%
Agenda Report Page 60
46 APPENDIX B
Narrative Abbreviations Used by the PPD
Race Abbreviation Definition

B BLACK

W WHITE

H HISPANIC

A ASIAN

O OTHER
Some examples of subject description abbreviations would be:

WMJ = White male juvenile
BFA = Black female adult
AMA = Asian male adult
HMJ = Hispanic male juvenile
Some dispatchers may put sex in front of the race identifier. Some examples of those:

MBA = Male black adult
FWA = female white adult
OMJ = Other male juvenile
Two-character abbreviations may also be used. Some examples below:

BM = Black male
MB = Male black
AF = Asian female
FA = Female Asian
OJ = Other juvenile
These are the most common abbreviation structures our dispatchers use to note subject descriptors. However, we do not have hard policies on abbreviations. You may see other variations or combinations of the above.

Agenda Report Page 61
47 APPENDIX C
Summary of Race/Ethnicity from All Data Sources
Table 7: Summary of Race/Ethnicity from All Data Sources White Black Hispanic Asian Other
Calls for Service with Narratives (N=23,547) 3.6% 3.0% 1.3% 0.5% -- Of Calls with Race/Ethnicity (N=1,842) 45.5% 38.8% 16.3% 6.1% -- Of ‘Suspicious’ with Race/Ethnicity (N=686) 45.5% 42.0% 16.9% 2.3% -- Of ‘Checks for theft/burglary’ with Race/Ethnicity
(N=214)

47.7% 39.3% 18.7% 3.7% --
Suspects with race/ethnicity (N=558) 32.8% 38.2% 22.2% 4.1% 2.7% Suspects with race/ethnicity (w/o Other) (N=543) 33.7% 39.2% 22.8% 4.2% -- FI Cards with race/ethnicity (N=1,423) 36.8% 38.8% 14.5% 6.2% 3.7% FI Cards with race/ethnicity (w/o Other) (N= 1,370) 38.2% 40.3% 15.1% 6.4% -- Arrests with race/ethnicity (N=1,195) 30.0% 42.7% 20.5% 4.4% 2.4% Arrests with race/ethnicity (w/o Other) (N=1,166) 30.8% 43.7% 21.0% 4.5% -- Agenda Report Page 62
48 APPENDIX D
Focus Group Interview Script
Community Focus Group Questions
In spring 2021, Chief Jeremy Bowers approached Dr. Michael Smith from the University of Texas at San Antonio to discuss a research collaboration with the City of Piedmont and Piedmont PD. Based on those discussions, a set of research questions were developed and at the end of July 2021, the City of Piedmont entered into a research agreement with UTSA and Dr.

Smith and Dr. Rob Tillyer. The scope of work for the project involves two primary areas of inquiry:
• An analysis of calls for service made to the PPD over the last five years to help the PDD and City understand
o The frequency of call types, police resources utilized, and changes in call types from initial dispatch through disposition and
o Categorize call types and make data-driven recommendations on the need for police versus non-police or co-response models
• Analysis of calls for service (CFS) to determine the frequency and impact on police operations of calls that lack specific descriptions of crime-related behavior. The UTSA research team will
o Benchmark caller descriptions of individuals against:
▪ Reported crime suspects from official police reports
▪ Arrestees
▪ Field interrogation
As part of their work, the UTSA research team will conduct focus groups with community and school stakeholders to help the researchers better understand the context of the calls for service data and to inform their analysis and final report.

The focus groups were selected by City leadership and will be facilitated by the UTSA researchers. With the permission of all participants, the focus groups will be recorded and written transcript of comments and responses will be created and used for qualitative data analysis purposes. Focus groups participants will be assigned a unique number and will be asked to identify themselves by that number whenever they speak. Once the recordings are transcribed, the recordings themselves will be destroyed. No one will be personally identified in any written report produced by the UTSA research team. Finally, focus group participants may withdraw from the focus group at any time.

Calls for Service Questions

1. What types of calls do you think the PPD responds to most often? Agenda Report Page 63
49 a. Nuisance, noise
b. Suspicious people
c. “people who don’t belong”: could be based on race and income (e.g., vehicles broken down, don’t appear to fit)
d. Schools
e. Additional items: suicide, burglary hiding in yard, etc.

2. Do you think most of the calls come through 911 or the PPD non-emergency number? a. Police would need to answer that question
b. Possibly non-emergency due to small community and awareness

3. Are there types of calls that you think the PPD should not respond to with a sworn police officer?
a. Suspicious person without evidence/”out of place”; it’s a hassle and impacts local residents
b. Mental health – perhaps mental health team needs to respond; complicated issue/no clear line given the overlap between crime and mental health c. Minors (May or may not have mental health involved); in particular on school grounds; pull back on sworn officers responding to schools d. High school kids and parties may not be necessary; disagreement on this point e. If so, how should these types of calls be handled?
i. Should they be handled by non-sworn PPD community service officers? Heard of, but not clear on role and actions
ii. Should some other city or county agency respond to these calls? Potentially a mental health unit

4. Are there types of calls that the PPD should send a sworn officer to handle but also send another responder? Significant differences between Piedmont and Oakland/SFO a. On what types of calls should PPD send someone else along with a police officer? b. What type of individual or agency representative should respond with a PDD officer to these kinds of calls?
Potential Bias-Based Calls to the Police

1. Does Piedmont have a crime problem?
a. What is the nature of that problem?
b. What kinds of crimes or public safety issues most concern you? i. Mixed; COVID has increased nuisance/property crimes
ii. Crime rate is low; particularly for violent crime

1. Why: police responsive; residents call about suspicious people; license plate reader; crime prevention; wealthy community; older Agenda Report Page 64
50 community; low unemployment; lack of commercial property as a target; lack of housing density
iii. Proximity is a factor; Oakland is an influencer (e.g., home invasions) iv. Additional safety: traffic safety

2. Are most crime or public safety problems created by Piedmont residents or by nonresidents? a. Maybe 50/50 in some neighborhoods; serious crimes are driven by non-residents; “big crime” (child molestation, drug dealing)-residents b. Depends on the type of crime (e.g., traffic vs. property crime); Drug use and selling-residents; Traffic-residents; Stealing catalytic converters-non-residents c. If non-residents, where do you think most of the suspects/offenders live? Oakland; greater East Bay; repeat offenders who may come into Piedmont d. Can you describe the “typical” person who causes crime or public-safety-related problems in Piedmont? YBM
i. How do crime or public safety problems vary in Piedmont by offender? In other words, are certain types of individuals (e.g. youth) more likely to cause problems than others?
ii. Do “typical” suspects or offenders in Piedmont come from a certain racial, ethnic, or socioeconomic class?
Perceptions are driven by media and what/who the police catch

3. When are you most likely to call the PDD for a crime or public safety problem? Driving behavior, potential suspicious activity; after the fact (car thefts, construction site, garages) a. Feedback loop: police responsive to calls, residents more likely to respond b. Are you most likely to dial 911 or call the PPD non-emergency number? c. Would you likely describe the individual about whom you called to the call-taker if asked?
i. If you were asked to describe the person’s approximate heigh, weight, and age would you be comfortable doing so?
ii. Would you be comfortable describing the individual’s perceived gender? iii. How about perceived race and/or ethnicity?
d. If you saw a person who did not appear to live in Piedmont walking on your street would you call the PPD?
i. What if the person was looking into cars or knocking on doors to see if anyone was home?

1. Would you be more or less suspicious of someone engaging in those behaviors if they appeared to be a resident versus a nonresident? Good confidence in the police; most see them as responsive, even maybe an over-response Agenda Report Page 65
51 School-Based Focus Group
Calls for Service Questions

1. What types of calls do you think the PPD responds to most often at Piedmont schools?

2. Do you think most of the calls come through 911 or the PPD non-emergency number?

3. Do you think most of the calls related to the schools come from school personnel (teachers, administrators), students, parents, or others?

4. Are there types of school-related calls that you think the PPD should not respond to with a sworn police officer?
a. If so, how should these types of calls be handled?
i. Should they be handled by non-sworn PPD community service officers? ii. Should some other city or county agency respond to these calls?

5. Are there types of school-related calls that the PPD should send a sworn officer to handle but also send another responder?
a. On what types of calls should PPD send someone else along with a police officer? b. What type of individual or agency representative should respond with a PDD officer to these kinds of calls?
Potential Bias-Based Calls to the Police

1. Does Piedmont have a crime problem at its schools?
a. What is the nature of that problem?
b. What kinds of crimes or public safety issues at the schools most concern you?

2. Are most crime or public safety problems at the schools created by Piedmont residents or by non-residents?
a. If non-residents, where do you think most of the suspects/offenders live? b. Can you describe the “typical” person who causes crime or public-safety-related problems at a Piedmont school?
i. How do crime or public safety problems at the schools vary in Piedmont by offender? In other words, are certain types of individuals more likely to cause problems at the schools than others?
ii. Do “typical” suspects or offenders in Piedmont come from a certain racial, ethnic, or socioeconomic class?

3. When would you be most likely to call the PDD for a crime or public safety problem at a school?
Agenda Report Page 66
52 a. Are you most likely to dial 911 or call the PPD non-emergency number? b. Would you likely describe the individual about whom you called to the call-taker if asked?
i. If you were asked to describe the person’s approximate heigh, weight, and age would you be comfortable doing so?
ii. Would you be comfortable describing the individual’s perceived gender? iii. How about perceived race and/or ethnicity?
c. If you saw a person who did not appear to be a parent or student at a Piedmont school would you call the PPD?
i. What if the person was looking into cars or approaching students?

1. Would you be more or less suspicious of someone engaging in those behaviors if they appeared to be a Piedmont resident versus a non-resident?
Agenda Report Page 67
Item # 4 - Recommendations on the Analysis of Police Call Data Correspondence Received before 3:30 p.m. on Monday, September 19, 2022 Hello City Council,
The text below is taken from page 4 of the staff report. The section raises the idea of an "alternative response" for some types of calls for service. I note that the consultants do not actually recommend this change based on any analysis of the response data in the report so you may want to inquire if they feel this alternative model is necessary. Staff responds by noting the creation of a part-time Community Services Officer (CSO) to take on these calls for service, many of which are administrative.

I think data in the report support the creation of a code enforcement officer. Data in table 4 show that 20% of calls for service are for city code violations (parking complaint, animal control, Piedmont city ordinance, 72-hour parking violation). Figure 6 shows that over 30% of single unit dispatches are sent to these types of calls which should tell you that our residents want to see the city code enforced.

A key question for the creation of whatever the position is called is whether that person can do code enforcement - would that person need to be a sworn officer? The staff response seems to suggest no and that only additional training would be required.

I would suggest that before the staff time of the part-time CSO be increased that a targeted implementation be tested for a trial period. Consistent code enforcement could suppress code violations and lessen the calls for such services.

Garrett Keating
RECOMMENDATION: Response to Calls for Service Apart from mental health calls, there are a variety of call types that may be amenable and/or appropriate for alternative responses by non-sworn officers either within or outside of the PPD. For example, parking complaints, welfare checks, roadway obstructions, lost and found property, and vandalism calls may be addressed without the need for a sworn officer to be dispatched. It may also be efficient and effective to have non-sworn officers take reports on completed property crimes.

Any of these adjustments must be considered with safety of the resident(s) and responding entity as a key consideration. While appearing rather innocuous and unlikely to require a sworn officer in response, some situations change rapidly and may unexpectedly require the presence of a PPD officer to assist. If changes to the current PPD response model are contemplated, alternative and/or co-responders must be provided with appropriate resources and training to ensure their safety and success. Sworn officers have the benefit of extensive training needed to handle many different situations. Piedmont should likewise equip alternative responders with the tools needed to improve outcomes over the current model and should track those outcomes carefully to evaluate their safety and efficacy.

RESPONSE: In 2019, the Department transitioned the part-time parking enforcement officer positions to part-time community service officer (CSO) positions. The specialized, part-time CSOs perform a variety of support functions within the police department. These functions include assistance to police officers with a wide range of responsibilities in support of safety and Agenda Report Page 4 community service. Some of the responsibilities for CSOs include: • Property and evidence management • Parking enforcement • Investigation of park and/or athletic field use violations • Fingerprint scans • Traffic control The CSOs regularly conduct vacation home checks and assist police officers in procedures such as impounding vehicles associated with criminal investigations. Aside from found property reports, CSOs do not respond to take police reports for other matters or crimes. Duties of the CSO position, particularly parking enforcement, and park patrol, are frequently requested by the public and other City departments.

Increased functions of the CSO position would likely require additional staffing, perhaps a change to full-time staff, and certainly additional training.

Piedmont City CouncilSeptember 19, 2022

Piedmont City Council
September 19, 2022