April 26, 2022
2021-105

The Governor of California
President pro Tempore of the Senate
Speaker of the Assembly
State Capitol
Sacramento, California 95814

Dear Governor and Legislative Leaders:

Our audit of five law enforcement departments throughout the State uncovered the actions of some officers at each department who engaged in biased conduct, either during their on-duty interactions with individuals or online through their social media posts. Although we did not find officers who were members of hate groups, some officers made statements indicating that they support problematic groups. We have included in this report some disturbing content (with some redactions) because we believe that it is important to accurately reflect the nature of the conduct that we observed.

We depend on law enforcement departments and the peace officers they employ to ensure that they exercise their unique authority without regard for individuals' identity characteristics, such as race, national origin, or mental or physical disability. What we found is that these five departments have not adequately guarded against biased conduct among their officers:

• They have not used sufficient strategies to achieve representative diversity in hiring.
• They have not implemented robust community engagement strategies or employee training practices.
• They have not established sufficient, proactive processes to identify possibly biased behavior.
• They have not consistently conducted adequate investigations of alleged biased behavior.

Departments' internal investigations often considered only the most blatant forms of bias. In one such case, a member of the public filed a complaint about an officer's social media posts. Although the officer's posts endorsed potentially harmful stereotypes about Black parents and Syrian refugees, the department's investigation concluded that it was "unable to find any racially derogatory remarks" and that the allegation of prejudice was "clearly false."

This report makes specific recommendations about steps each department can take to better ensure that Californians receive fair and impartial policing services. We also make several recommendations to the Legislature to better align expectations in state law with best practices for addressing bias in policing, such as by adopting a uniform definition of biased conduct, requiring more frequent and thorough training, and increasing independent oversight.

Respectfully submitted,

MICHAEL S. TILDEN, CPA
Acting California State Auditor

---

Selected Abbreviations Used in This Report

CDCR	California Department of Corrections and Rehabilitation
CPOST	Commission on Correctional Peace Officer Standards and Training
DOJ	California Department of Justice
IACP	International Association of Chiefs of Police
NACOLE	National Association for Civilian Oversight of Law Enforcement
POST	Commission on Peace Officer Standards and Training
RIPA	Racial and Identity Profiling Advisory Board

Summary

Audit Highlights . . .

We conducted an audit of five law enforcement departments throughout the State—the Los Angeles Sheriff; the police departments of San Bernardino, San José, and Stockton; and CDCR—to assess their efforts to guard against officers' biased conduct, and we found the following:

• » Some officers at each department had engaged in biased conduct.
•     Some misconduct occurred during on‑duty interactions, and we identified 17 officers who promoted biased content on social media.
•     We did not definitively identify any officers as members of hate groups.
• » None of the departments had fully implemented best practices to mitigate the effects of officer bias.
•     Each of the departments had struggled to ensure that its officers fully reflect the diversity of the community.
•     Each department's training about bias could be more frequent and include additional content.
•     The local departments could do more to build and strengthen relationships with their communities.
•     None had established adequate systems for proactively identifying and correcting problematic officer performance trends.
• » The local departments did not consistently or adequately investigate their officers' possibly biased conduct.
•     Many investigations were narrowly focused on blatant signs of bias, relied heavily on officers' denials, or did not account for how officers' conduct reasonably appeared.
•     Each department needs a better framework for consistently identifying, investigating, and tracking incidents of bias.
• » State-level intervention would help ensure broader use of practices that address bias.
•     Local departments have not proactively adopted best practices to address the threat that bias poses to fair and impartial law enforcement.
•     External reviews and oversight would be effective measures for increasing departments' adoption of best practices.

Results in Brief

Law enforcement departments play an important role by providing public safety for all members of their communities. To fulfill that responsibility, peace officers (officers) must work in inherently challenging and dangerous environments, and must sometimes make quick decisions about how to uphold public safety. The public has entrusted law enforcement departments and their officers with significant, unique authority, including the right to detain, arrest, and use force against individuals within the community. Officers should exercise that authority with fairness, impartiality, and professionalism. In particular, their conduct should not be improperly influenced by an individual's identity characteristics, including their race, ethnicity, gender, religion, or sexual orientation.

Because bias can take the form of preconceived judgments, opinions, or attitudes about people based on their actual or perceived identity characteristics, it can interfere with officers' abilities to treat members of the public with fairness, impartiality, and respect. Some people possess explicit biases, which are conscious and deliberate. Indications of explicit bias include using racial or ethnic slurs toward individuals and the intentional promotion of stereotypes. Another form of bias, which all people possess, is implicit. Implicit biases are unconscious associations or perceptions about groups of people. Research has shown that implicit biases can cause officers to perceive certain groups of people as more threatening than others. Further, implicit biases can cause even well-intentioned people who outwardly reject prejudice and stereotypes to treat people differently and unfairly based on their identity characteristics.

Law enforcement departments should be vigilant and proactive in identifying and addressing explicit and implicit biases in officers because of the harm that both can cause. However, we identified some officers at each of the five law enforcement departments we reviewed—the California Department of Corrections and Rehabilitation (CDCR), the Los Angeles County Sheriff's Department (Los Angeles Sheriff), the San Bernardino Police Department (San Bernardino Police), the San José Police Department (San José Police), and the Stockton Police Department (Stockton Police)—who had engaged in biased conduct. We reviewed a selection of five internal investigations at each department, and in some of those cases, officers' conduct showed signs of possible bias. We also selected 750 officers from across the five departments and determined whether they had identifiable public social media accounts. Of the about 450 officers who had public social media accounts, 17 officers had posted biased statements or content. The posts we identified either promoted negative stereotypes or contained deliberately hateful and derogatory speech directed at groups of people.

We also reviewed selected internal investigations and public social media accounts to determine whether any officers were members of hate groups. The Federal Bureau of Investigation defines a hate group as "an organization whose primary purpose is to promote animosity, hostility, and malice against persons of or with a race, religion, disability, sexual orientation, ethnicity, gender, or gender identity which differs from that of the members or the organization." There were limitations on our ability to assess whether officers were members of hate groups, including that our statutory access is generally limited to the records held by public entities and that there were privacy settings on some officers' social media accounts. Although we did not identify evidence that any officers were members of hate groups, six officers posted content suggesting that they support groups with problematic principles or activities. For example, one officer posted a statement defending the Proud Boys—a group that has expressed hostility toward women and Muslim individuals—by saying that people who are against the Proud Boys are "in reality just against masculinity."

The biased conduct that we identified at the five law enforcement departments likely occurred in part because the departments have not fully implemented comprehensive strategies for addressing bias within their organizations. Several law enforcement authorities—including the U.S. Department of Justice and the California Commission on Peace Officer Standards and Training (POST)—have published guidance and best practices for creating bias-free policing cultures. These best practices highlight the importance of establishing a formal commitment to preventing bias from interfering with the impartial delivery of law enforcement services. They also stress honoring that commitment by incorporating strategies for addressing bias into key processes, including recruitment, hiring, training, community engagement, and misconduct investigations. Although generally directed at local law enforcement departments, many of these best practices are relevant and critical to addressing bias in correctional facilities as well. However, neither CDCR nor the local law enforcement departments that we reviewed have fully adopted best practices in these key areas.

Further, we found that each of the local departments had not appropriately addressed indications of bias when they occurred. When we reviewed internal investigations at the four local law enforcement departments, we found significant weaknesses in their approaches to investigating whether bias had influenced officers' actions. Departments often limited investigations to examining whether officers had engaged in only the most blatant forms of bias, such as uttering racial slurs. For example, a member of the public submitted to one department a complaint that an officer had made prejudicial statements after the officer posted statements on social media that promoted a negative stereotype about Black parents and reposted an article that implied that all Syrian refugees were terrorists. In its investigation of the complaint, the department did not analyze the officer's use of stereotypes; instead, it concluded that it was "unable to find any racially derogatory remarks" and that the allegation of prejudice was "clearly false."

An additional problem with the local departments' investigations of biased conduct is that they relied heavily on the officers' denials that bias influenced their actions, without considering whether an officer's conduct created the reasonable appearance of bias. This approach is problematic both because officers are unlikely to admit that bias affected their actions and because it fails to consider the likely perspective of the community. We did not find this problem at San Bernardino Police because the records we reviewed included no evidence that the department had directly questioned the related officers about whether their conduct had been motivated by bias.

As a result of the deficiencies in the departments' investigative practices, they are at higher risk for failing to identify instances when their officers engage in biased conduct and failing to take action to prevent those officers from engaging in biased conduct in the future. None of the departments that we reviewed have adequate guidance for how to determine whether bias influenced officers' conduct. In the absence of such guidance, we observed that at San José Police—the only department we reviewed that consistently provided formal analysis in its investigations about whether conduct was biased—some investigative analyses and conclusions differed widely from one another for no discernible reasons. A statewide definition of biased conduct and guidelines for applying this definition during investigations would clarify expectations for law enforcement departments.

Greater statewide oversight could increase law enforcement departments' adoption of best practices for addressing bias. Reviews by other external agencies, including the California Department of Justice (DOJ), show that the problems that we identify in this report are not limited to the five departments we reviewed. That said, external reviews and oversight have proven effective in advancing law enforcement departments' adoption of practices for preventing bias. Requiring the Racial and Identity Profiling Advisory Board (RIPA Board)—an entity within DOJ charged with eliminating racial and identity profiling in law enforcement—to review and report on departments' implementation of best practices for preventing bias could provide greater visibility into local department practices and assist the public and policymakers in holding departments accountable.

In addition, DOJ currently performs occasional, in-depth reviews of law enforcement departments' operations, including efforts to mitigate bias. These reviews have resulted in positive change in the departments in question. However, from 1999 until March 2022, it had completed only six reviews of city or county departments that examined their broader policing practices, with another four ongoing. Requiring DOJ to perform reviews more frequently could provide the public with in-depth assessments of departments' implementation of best practices.

The work officers perform is both critically important to our State and often extremely challenging. We found no evidence that the majority of the officers we reviewed had engaged in biased conduct. Nonetheless, biased conduct by even a few officers can cause harm to members of the public and erode the community's trust in law enforcement. For these reasons, it is incumbent upon law enforcement departments and the State to take the steps necessary to guard against officer bias in all its forms.

Agency Comments

Each of the departments that submitted written responses to our report generally agreed with our recommendations, with some exceptions. Stockton Police did not clearly state whether it would implement our recommendations, but noted that it would analyze our audit and see how it can align its policies and procedures with best practices. San Bernardino Police did not submit a written response to our report; however, we incorporate its perspective on our findings and recommendations throughout the text of the report.

---

Introduction

Background

Law enforcement departments and the peace officers (officers) that they employ provide an essential function in the communities they serve: they are responsible for the safety of all residents in their communities. To fulfill this responsibility, officers must work in inherently challenging and dangerous environments and must sometimes make quick decisions about how to uphold public safety. Because their responsibility is essential to the safety of their communities, officers are entrusted with significant and unique authority. This authority includes the ability to make decisions to deprive individuals of personal freedoms. For example, officers may detain or arrest individuals. In some cases, they may also use physical, including deadly, force.

In light of this authority, the public expects that officers will perform their duties in a fair manner. Under any given set of circumstances, the way that officers treat individuals and the enforcement actions that they take should be the same or similar, regardless of the individual's identity characteristics. In fact, state law specifically prohibits officers from engaging in racial or identity profiling, which it generally defines as considering or relying on a person's actual or perceived identity characteristics when deciding to detain or search that person or when deciding on the scope or substance of law enforcement activities after that detention or search. ¹ State law allows an officer to rely on identity characteristics that are listed in a specific suspect description. For example, if a witness describes a suspect as a "white male," officers can consider that information in their search for the suspect. The text box lists the specific identity characteristics that are protected by state law.

Explicit and Implicit Bias

Bias represents a threat to officers' ability to treat all people fairly. In general, bias constitutes a lack of objectivity that can take the form of preconceived judgments, opinions, or attitudes about a person or group of people based on actual or perceived identity characteristics. As Figure 1 shows, bias can be explicit or implicit, and both types can affect a person's behavior and treatment of others. Individuals with explicit biases are aware of their prejudices and attitudes toward certain groups. These positive and negative attitudes and beliefs about groups of people are conscious and deliberate. In contrast, implicit bias is unconscious. It is an automatic association or attitude about a group of people, and it operates outside of a person's awareness. All people—including those who outwardly reject prejudice and stereotypes—possess and are affected by implicit biases. In fact, people can possess implicit biases about others with whom they share key identity characteristics.

Figure 1

Explicit and Implicit Bias Can Both Affect Behavior

Because the two forms of bias are distinct from one another, the effect they can have on behavior ranges widely. For example, a particularly extreme manifestation of explicit bias is affiliation with a hate group. The Federal Bureau of Investigation (FBI) defines a hate group as "an organization whose primary purpose is to promote animosity, hostility, and malice against persons of or with a race, religion, disability, sexual orientation, ethnicity, gender, or gender identity which differs from that of the members or the organization." Because explicit biases are rooted in consciously held preferences for or against certain groups of people, they are incompatible with being an officer.

Implicit bias can also cause individuals to engage in biased conduct, although the presentation of that bias can, at times, be subtle. For example, implicit bias may lead an individual to give certain groups of people the benefit of the doubt while treating others with suspicion under similar circumstances. Because implicit bias is unconscious, individuals are likely to be unaware of its effects on their behavior. Nonetheless, multiple studies have shown that implicit bias at times affects the decisions made by members of numerous professional groups, including doctors, educators, and judges.

When bias influences an officer's actions, it can have particularly detrimental effects because it may interfere with the fairness with which officers perform their duties. For example, studies have shown that implicit bias can increase the likelihood that officers will perceive individuals who are Black as threatening. Studies have also suggested that bias can play a role in officers' decisions to use force against specific individuals.

Because of the effects that bias can have on officer conduct and the potential damage it can cause, it is critical that law enforcement departments take action to address it. In general, law enforcement departments should strive to identify and exclude from their ranks any officers with biases that render them incapable of fairly exercising their duties. Because everyone has implicit biases, law enforcement departments generally must take steps, such as providing training and education, to mitigate these biases' effects. Studies have suggested that an officer's implicit biases can be overcome through targeted interventions that reduce the effect of bias. Throughout this report, we describe the best practices that law enforcement departments can implement to effectively address officer bias.

Recent Action by the State to Address Biased Policing

Recent events have heightened national awareness of and attention on the subject of bias in law enforcement departments. These events have included officer-involved shooting deaths of unarmed Black individuals, protests associated with the Black Lives Matter movement, and the surfacing of biased social media content associated with some officers. Further, reports in the media have included allegations that some officers are affiliated with groups that espouse bias against others, including the Proud Boys and the Ku Klux Klan. These events have led to calls for increased accountability and oversight.

One of California's key tools for making such changes is the Racial and Identity Profiling Advisory Board (RIPA Board). State law required the Office of the Attorney General to establish the RIPA Board by July 2016 to eliminate racial and identity profiling and to improve diversity and racial and identity sensitivity in law enforcement. The RIPA Board includes, among many others, a representative from the California Police Chiefs Association, a university professor who specializes in policing and racial and identity equity, and two representatives from community organizations. With the assistance of staff within the California Department of Justice (DOJ), the RIPA Board is responsible for annually analyzing data from law enforcement departments across the State and other responsibilities as the text box shows. It then issues annual public reports on a variety of subjects, including the total number of public complaints alleging racial or identity profiling. The RIPA Board's annual reports contain recommendations for local law enforcement departments, the Legislature, and the Commission on Peace Officer Standards and Training (POST) regarding best practices and proposed changes to state requirements to address and mitigate the effects of bias. POST is a governor-appointed commission within DOJ that sets minimum selection and training standards for California law enforcement departments. ² Law enforcement departments' compliance with POST standards is voluntary and incentive‑based. Participating departments agree to abide by the standards in exchange for services and benefits, such as reimbursement for officer training. Most departments in the State, including all of the departments we reviewed during this audit except for CDCR, participate in POST.

In addition, the State has recently enacted laws aimed at increasing public accountability and transparency in law enforcement. For example, effective January 2022, state law requires POST to develop a system for revoking or suspending an officer's certification—thereby revoking or suspending his or her ability to be employed as a peace officer—for serious misconduct, including excessive force, sexual assault, dishonesty, and demonstrations of bias. Further, a separate law—also effective in January 2022—requires records related to sustained findings of certain forms of officer misconduct, including discriminatory or prejudicial behavior, to be disclosed under the California Public Records Act.

In light of concerns about bias and hate group affiliation among officers, the Joint Legislative Audit Committee (Audit Committee) directed our office to conduct a review of a selection of law enforcement departments to determine whether their officers were members of hate groups or had engaged in biased conduct. It further asked us to evaluate the departments' practices for addressing bias within their ranks. We reviewed four local law enforcement departments: the Los Angeles County Sheriff's Department (Los Angeles Sheriff), the San Bernardino Police Department (San Bernardino Police), the San José Police Department (San José Police), and the Stockton Police Department (Stockton Police). At the direction of the Audit Committee, we also included the California Department of Corrections and Rehabilitation (CDCR) in portions of our review. The text box shows the number of sworn employees at each department.

---

Chapter 1

Some Officers at Five Law Enforcement Departments Displayed Bias

Key Points

• At each department we reviewed we identified some officers who had engaged in biased conduct. Some of this conduct occurred during their on-duty interactions with individuals. We also identified 17 officers who made biased statements or promoted biased content on social media.
• We did not identify sufficient evidence that any of the officers we reviewed were members of hate groups; however, six officers promoted content on their social media pages related to groups that foster animosity toward people with protected identity characteristics, including Muslims, individuals who identify as LGBTQ, and immigrants.
• None of the departments we reviewed have fully developed and implemented comprehensive efforts to address bias among their officers. Without a comprehensive approach to guard against the presence and effects of bias, the departments will be less able to identify, mitigate, and address bias.

At Each of the Five Law Enforcement Departments, We Identified Officers Who Have Promoted Negative Stereotypes and Engaged in Biased Conduct

Some officers at each department we reviewed—CDCR, Los Angeles Sheriff, San Bernardino Police, San José Police, and Stockton Police—engaged in biased conduct either in their on-duty interactions with individuals or online through their social media posts. We reviewed a total of 25 investigations the departments had conducted, such as in response to complaints they received from the public or other officers, and we also reviewed the publicly available Internet activity for about 750 officers across the five departments. As we describe in the next section, we did not identify any officers with affiliations with hate groups as a result of these reviews. However, we did identify conduct by officers that was either explicitly biased or reasonably appeared to have been influenced by bias. Throughout the rest of this section, we describe examples of that conduct. We caution that the text and images that follow contain content that some readers may find disturbing.

We identified some officers' biased conduct through complaints that the departments had received from members of the public and other officers. We reviewed documentation, including body-worn camera footage when it was available, for a selection of these complaints, as well as for other internal investigations. Not every investigation we reviewed contained clear and direct evidence of biased conduct. Nevertheless, we identified incidents in which officers had made blatantly biased statements. Figures 2, 3, and 4 provide examples of this conduct.

WARNING: This figure contains content that readers may find disturbing.

Figure 2

Examples of Biased Conduct by Peace Officers Identified in CDCR's Internal Investigations

---

WARNING: This figure contains content that readers may find disturbing.

Figure 3

An Example of Biased Conduct by Officers at a Law Enforcement Department We Reviewed

---

WARNING: This figure contains content that readers may find disturbing.
We have redacted some of the posts below.

Figure 4

Example of Biased Conduct That Was the Subject of a Complaint

In addition, we identified other incidents in which there were indications that bias may have influenced an officer's conduct. Because the departments documented these incidents in their investigation files, it is clear that they were aware of the conduct. However, as we explain in more detail in Chapter 4, the four local departments often did not take appropriate action to recognize and address possibly biased conduct after it had occurred.

We also identified some officers who had posted biased content on their public social media accounts, as the first text box shows. When we reviewed about 750 officers we selected from across the five departments, we found that about 40 percent did not have any identifiable social media accounts. Further, some of the remaining officers had privacy settings that limited the amount of content that was publicly viewable. However, of the 450 officers who had discoverable social media accounts with some level of public activity, 13 had posted or supported content that contained biased statements or images. Although the number and age of these posts varied among these officers, the records we reviewed indicate that all of the posts occurred while they were employed as officers.

The biased statements that we identified targeted various groups, including those listed in the second text box. We considered the content we identified to be biased because it generally either promoted negative stereotypes or contained hateful or derogatory speech about a group of people. Figure 5 contains examples of these statements.

In addition to the 13 officers who made biased statements on social media while they were employed by the law enforcement departments, as we show in the third text box, we identified four officers who posted biased content on their social media accounts before they were hired by the departments we reviewed. These posts were similar in nature to those we discuss above. These cases highlight the importance of departments effectively screening for bias and prejudice in their hiring processes, an area we describe in more detail in Chapter 2.

The departments we reviewed have recognized the harm that an officer's social media activity can cause. The departments have policies prohibiting certain conduct on social media by officers, including on an officer's personal social media accounts. Some of the policies specifically prohibit making derogatory or biased statements. For example, San José Police's policy states that department members must not post online any material that is disparaging, insulting, or disrespectful to a group or class of people. The policy indicates that such posts may adversely reflect upon the department, impair its working relationships, or inhibit its ability to operate efficiently and effectively.

WARNING: This figure contains content that readers may find disturbing.
We have redacted some of the posts below.

Figure 5

Examples of Biased or Prejudiced Statements Peace Officers Made on Social Media

When an officer engages in biased conduct or makes statements like those we describe in this section, it casts doubt on that officer's ability to treat individuals fairly. For example, one officer made comments about Muslims that raise questions about how that officer might perceive and react to Muslim individuals in the course of the officer's duties:

"There was a time where our country, we were letting in, we were a safe zone and I was totally against it, especially with Muslims. . . . In Iraq. . . . they do not like us and they never will. . . . they hate Americans…they strap on bombs and they come at us. . . . if they're really willing to do that, for their own people to kill us, what do you think is going to happen when we allow a lot of them [to] come over here? So at that point in time in my life I was completely against it."

The biased conduct we identified was generated by a small number of the officers at each department. Nevertheless, concluding on that basis alone that bias is not a significant problem at these departments would be incorrect for a number of reasons. By its nature, our review was not designed to catalogue every instance of biased conduct or statements by officers at these departments. Our work encompassed only a limited number of internal investigations and the publicly shared views of a selection of officers. Moreover, the behavior of even a few officers can erode a community's trust in law enforcement and damage the relationship between a department and the community it serves. According to the U.S. Department of Justice (US DOJ), community trust—which we discuss further in Chapter 3—reinforces police legitimacy and is critical to preventing and controlling crime.

Most importantly, the departments we reviewed have not fully adopted strong safeguards against bias and its effects, as we describe in the chapters that follow. As a result, these departments are at a higher risk of being unaware of and unable to effectively address the ways in which their officers exhibit bias. Without a comprehensive set of practices to address bias—such as those we describe later in this chapter—law enforcement departments cannot know the extent to which bias is a problem in their organizations or whether they are effectively combating that problem.

We Identified Six Officers Who Either Publicly Defended or Promoted Content From Problematic Groups

As we describe in the Introduction, affiliation with a hate group reflects an extreme form of explicit bias. Across the nation, concern has been growing regarding officer involvement in such groups. Our statutory access to records for an audit is limited to those held by public entities and private entities subject to review or regulation by public entities, and does not extend to records held by hate groups. However, to identify whether officers at the five departments we reviewed had links to hate groups, we interviewed relevant staff at each department, reviewed records associated with complaints and internal investigations, and examined data and other records held by DOJ and California Governor's Office of Emergency Services (Cal OES). Further, as we describe in the previous section, we investigated and reviewed the online activity of about 750 officers across the five departments.

We found insufficient evidence to prove that the officers that we reviewed were members of hate groups. That said, six of the officers posted content suggesting that they support groups with problematic principles or activities. One officer posted a statement defending the Proud Boys, which is an organization that the Canadian government states has promoted hostility toward Muslim individuals and women, as well as encouraged acts of violence against groups that it perceives to be its opponents. In his post, the officer stated that the Proud Boys had denounced white supremacy and that those against the Proud Boys are "in reality just against masculinity." That officer was among the officers we describe earlier who posted biased statements on their social media pages. The biased statements he posted were demeaning toward transgender individuals and promoted stereotypes about Asian individuals.

The remaining five of the six officers promoted content on their social media sites related to other problematic organizations. For instance, one officer's profile picture on his mainstream social media account was an image representing the Three Percenters, a group that the Canadian government and the Anti-Defamation League have described as having an extremist antigovernment ideology. Some of the ideology's followers have promoted hate and engaged in acts of violence toward Muslims. Another officer shared content from the same group on his social media account. The fourth of the six officers liked on his social media account a group that has publicly expressed animosity toward immigrants. The group publishes content that promotes stereotypes about immigrants, including that they engage in criminal activity, are involved in terrorism, and negatively impact American workers and taxpayers. The fifth officer identified himself as affiliated with a group that opposed same-sex marriage and promoted claims that having same-sex parents is harmful to children. Finally, the sixth officer liked a social media page that lauds the Confederacy. Given the public statements and activities of these groups, these officers' promotion or support of these groups on social media calls into question their ability to treat individuals fairly in the exercise of their duties.

Given the significant threat that officer involvement in hate group activities poses, law enforcement departments must strive to identify and address that behavior. However, some inherent challenges exist to identifying that an officer is affiliated with a hate group. For example, the U.S. Department of Homeland Security reported that some violent extremist hate groups use "leaderless resistance" tactics, wherein an individual or small group operates independent of any movement, leaders, or network of support, which allows them to "fly under the radar" and be difficult for law enforcement to detect.

An additional challenge to identifying officers involved in hate groups is the general protections that exist around freedom of speech and association. The U.S. Constitution provides broad and important protections that generally allow individuals the freedom to share their opinions and beliefs, both individually and as part of a group, without government interference. Accordingly, after reviewing the activities and records at DOJ and the Homeland Security Division within Cal OES, we found no evidence that indicated that those entities track or monitor individuals based on their beliefs or associations. Similarly, the FBI has publicly stated that it does not collect or maintain information on individuals solely for the purpose of monitoring First Amendment protected activity.

However, some state and federal departments, including the FBI, DOJ, and the Homeland Security Division within Cal OES, may conduct investigations or threat assessments of individuals or groups when there is a threatened or actual crime—such as a hate crime or act of terrorism. Further, information about those threats or investigations may be available to other law enforcement departments. Additionally, peace officer candidates are subject to a criminal records search at the local, state, and national levels to determine legal eligibility for peace officer employment and eligibility to carry a firearm. Law enforcement departments must forward the candidate's fingerprints to DOJ, which will conduct a state search that may reveal if an applicant has been convicted of crime that would disqualify the applicant from being a peace officer, including a felony or certain misdemeanors. Collectively, the required criminal records checks would assist law enforcement departments in identifying individuals with strong manifestations of bias, such as the commission of a hate crime.

Constitutional protections prohibit law enforcement departments from excluding officers solely on the basis of past or present membership in hate groups. However, a department may exclude an officer from employment who has engaged in an activity related to such an organization if the activity is incompatible with the mission of the department—including the department's need to maintain an image of fairness and impartiality. In fact, pending legislation would require law enforcement departments to terminate officers who have participated in hate group activities.

None of the Law Enforcement Departments Have Taken a Comprehensive Approach to Guarding Against Bias and Its Effects

The cases of biased conduct that we identified occurred at law enforcement departments that have not adequately implemented policies and practices that would safeguard them against the presence and effects of bias. Although each of the strategies we discuss in the subsequent chapters of this report provides benefits, no single practice is sufficient on its own to guard against bias. Instead, departments must adopt a comprehensive approach that addresses each of these areas of best practice. In the absence of such a comprehensive approach, departments will be less able to identify, mitigate, and address bias.

Numerous authorities, including those listed in the text box, have published best practices related to addressing bias and ensuring the fair and equitable enforcement of laws. These best practices are generally directed at local law enforcement departments; however, many of them are relevant and critical for correctional facilities as well. A correctional officer's role differs in key ways from a local law enforcement officer's role. However, correctional officers possess a similar responsibility to perform their duties in a fair and impartial way. Their treatment of incarcerated individuals should not be improperly influenced by an individual's protected identity characteristics.

According to many of these sources, a critical first step to providing bias-free law enforcement is adopting a formal policy against biased conduct. This policy should serve as the foundation for the department's efforts to address bias. The International Association of Chiefs of Police (IACP) states that a department's policy on bias should declare that biased conduct is prohibited, describe in detail what constitutes biased conduct, and outline key compliance mechanisms. Establishing such a policy conveys that a department's leadership is committed to preventing bias from interfering with its responsibility to treat members of the public equally in the exercise of its duties.

Nonetheless, neither Stockton Police nor CDCR has such a policy. As a result, both departments lack an important management tool for communicating expectations for standards of conduct for officers, as well as clear commitments to fairness for their communities. Both departments asserted that although they do not have specific policies prohibiting biased conduct, other policies address bias. For example, Stockton Police pointed to several policies—such as its policies related to restraining orders, the activities of its equestrian unit, and parole and probation violations—that include an express commitment to policing in a respectful, neutral, and trustworthy manner. However, those policies do not define bias-free policing and lack other key elements that best practices recommend that a bias-free policy should include.

The remaining three departments—Los Angeles Sheriff, San Bernardino Police, and San José Police—have established policies formally committing to bias-free policing but have not fully integrated the policies' principles into their key processes. For example, San Bernardino Police's policy prohibits biased‑based policing and further states the department is committed to providing law enforcement services to the community with due regard for the racial, cultural, or other differences of those it serves. It further outlines the general responsibilities for both patrol officers and supervisors related to its implementation. However, as we show in the subsequent chapters, San Bernardino Police's processes and practices do not consistently or comprehensively reflect this policy. We found similar issues at Los Angeles Sheriff and San José Police.

A policy prohibiting biased conduct is not sufficient on its own for a department to establish and maintain a culture of fairness and impartiality. Rather, a department must ensure that its actions consistently reflect the principles upon which that policy is based. Drawing a conclusion from settlement agreements that the US DOJ entered into with law enforcement departments, the RIPA Board states that departments should implement bias-free policing principles into their key functions and processes. These include job descriptions, recruitment and training efforts, personnel evaluations, and accountability systems, among others. In particular, authorities on law enforcement highlight the importance of departments' implementing the best practices that Figure 6 shows to combat bias, prejudice, and hate in their organizations. Multiple studies have shown that specific interventions can help to mitigate bias and its effects. In particular, research supports that simple awareness and acceptance of one's own implicit biases can help to reduce the effects of those biases.

Figure 6

Law Enforcement Authorities State That Law Enforcement Departments Should Address Five Key Areas to Combat Implicit and Explicit Bias in Their Organizations

Please refer to the recommendations section to find the recommendations that we have made as a result of these audit findings.

---

Chapter 2

Key Best Practices Could Strengthen the Five Law Enforcement Departments' Hiring Processes

Key Points

• A diverse workforce provides many benefits, including improved cultural sensitivity and reduced stereotyping; however, like law enforcement departments nationwide, each of the five departments we reviewed has struggled to ensure that its officers fully reflect the diversity of its community.
• Despite the challenges the five departments face in recruiting a diverse workforce, none had fully implemented key best practices for reaching diverse applicants.
• Although each of the departments has generally followed requirements in state law and regulation for hiring peace officers, they have not fully implemented key best practices—such as robust reviews of applicant social media pages—that could better enable them to identify individuals with potentially disqualifying biases.

None of the Five Departments We Reviewed Have Fully Implemented Best Practices for Recruiting a Diverse Workforce

Authorities on law enforcement agree that law enforcement departments must strive to create a workforce that is composed of individuals who are qualified to perform their duties and who reflect the diversity of the communities they serve. Diversity helps departments to develop culturally sensitive responses to community problems. It also reduces stereotyping of groups in the community by officers, as well as stereotyping of officers by the community. The IACP has highlighted that diversity helps a department build trust with a community, and US DOJ has shared research demonstrating that departments with greater diversity are more open to reform, more willing to implement cultural and systemic changes, and more responsive to the residents they serve.

The departments we reviewed have had varying levels of success in ensuring that their officers reflect the diversity of their communities. For instance, the proportions of Black officers and white officers that Los Angeles Sheriff employs are generally consistent with the Los Angeles County population. However, as Table 1 shows, according to the personnel data that the departments provided, none of the five departments we reviewed have workforces that proportionally reflect all the demographic groups within their communities. For example, San Bernardino Police and Stockton Police both employ a significantly lower proportion of Black officers than the proportion of Black individuals living in their communities. The same is true for Asian officers at CDCR, Los Angeles Sheriff, San José Police, and Stockton Police. Although San Bernardino Police, San José Police, and Stockton Police have lower proportions of Hispanic or Latino officers than their communities, CDCR and Los Angeles Sheriff have higher proportions than their respective communities. Further, the percentage of white officers in San Bernardino Police and Stockton Police far outpaces the overall population of white individuals in the communities they serve.

Table 1

The Diversity of the Departments Does Not Align With the Diversity of Their Communities

These disparities are not unique to these five departments. Certain demographic groups, such as female, Black, and Asian individuals, have historically been underrepresented as officers across the country, and law enforcement departments have reported struggles in recruiting individuals in these groups. In its guidance for law enforcement departments, the IACP notes that multiple factors can discourage women and members of certain demographic groups from considering a career in law enforcement. According to the IACP, conventional wisdom among police leaders holds that nonwhite community members are more likely to associate policing with civil rights abuses and present-day unequal treatment. In the same report, the IACP states that law enforcement leaders believe that women feel unwelcome in law enforcement and that they fear sexual harassment, being stereotyped, or being ostracized, among other concerns.

In the face of these challenges, it is critical that law enforcement departments implement best practices for recruiting applicants from diverse backgrounds. The text box lists key practices that authorities on law enforcement have identified. Each of the five departments we reviewed has taken steps to attract diverse applicants, such as through social media advertising. CDCR has contracted with a vendor to develop a marketing campaign and to research and leverage partnerships with diversity partners such as cultural groups. CDCR meets periodically with its vendor to assess the success of recruitment efforts. Similarly, San José Police has produced recruiting videos that it has posted online featuring diverse San José Police officers, including female officers, officers who identify as LGBTQ, and officers of different races. All of the departments except San Bernardino Police also demonstrated that they have conducted recruitment activities at multiple events for, or hosted by, diverse community groups. The lieutenant who oversees San Bernardino Police's recruitment unit asserted that it also recruits at these events, but the department lacked documentation to demonstrate that it had.

Although these efforts have likely improved the departments' ability to recruit diverse applicants, the departments have not implemented other key practices that could further increase their success. Table 2 summarizes our review of the departments' practices. For example, only Los Angeles Sheriff and Stockton Police have completed formal recruitment plans that include strategies for recruiting applicants from various demographic groups. A formalized recruitment strategy guides a department to document specific, measurable goals and the actions it will take to accomplish those goals. Los Angeles Sheriff's recruitment plan, for instance, includes a goal that 20 percent of its officers will be women. The plan also details strategies for achieving that goal, including partnering with Los Angeles County to sponsor an initiative to attract more women into public safety positions and working with a consulting firm to identify best practices for attracting female applicants.

Table 2

The Five Departments Varied in Their Implementation of Key Practices for Recruiting a Diverse Applicant Pool

Key Practice	CDCR	Los Angeles Sheriff	San Bernardino Police	San JosÉ Police	Stockton Police
Has a formalized strategy, including goals, for recruiting a diverse applicant pool	Partial	Adequate	Poor	Poor	Adequate
Advertises on social media	Adequate	Adequate	Adequate	Adequate	Adequate
Recruits at community events hosted by or for underrepresented populations	Adequate	Adequate	Poor	Adequate	Adequate
Assesses the demographics of current personnel against the demographics of the community to identify areas for improvement	Adequate	Poor	Poor	Adequate	Adequate
Collects and analyzes applicant demographic data to assess the success of recruitment efforts	Adequate	Poor	Poor	Poor	Partial

Source: Review of recruitment practices at the five law enforcement departments.

In contrast, neither San Bernardino Police nor San José Police have formalized recruitment plans. The lieutenant who oversaw San Bernardino Police's recruitment efforts during the time of our review could not explain why the department did not have a formal recruitment plan, but agreed that a formalized strategy would improve San Bernardino Police's recruitment efforts. Similarly, the lieutenant who oversees recruitment at San José Police noted that the department employs many strategic measures to recruit diverse applicants, but agreed that formalizing recruitment strategies would be beneficial.

At the time of our review CDCR did not have a formalized recruitment strategy for its entire workforce, but provided evidence of some tactics for reaching diverse applicants, such as a campaign to reach female applicants as well as a plan for reaching more diverse applicants. The plan for reaching diverse applicants includes steps that CDCR will take to target its advertising and attend certain events to recruit applicants, but it lacks formal goals for measuring its success in these efforts. CDCR's recruitment manager stated that CDCR is in the process of developing a formal recruitment plan. She provided a draft plan that includes strategies for increasing female and diverse applicants and also includes methods for monitoring the effectiveness of those recruitment efforts.

Another key practice that departments should adopt is the regular review of their diversity-related data. Two key types of diversity data can assist departments in evaluating their recruitment efforts. First, regularly monitoring the demographics of its current personnel can help a department assess whether those demographics match the community it serves and highlight areas to improve diversity. Second, monitoring the diversity of its applicant pool can help a department understand the effectiveness of its recruitment efforts and identify barriers its application processes may inadvertently present. For example, if a department receives a high number of applications from female applicants, it could indicate that its efforts to recruit female applicants have been successful. However, if few of those female applicants successfully make it through the hiring process, it could indicate that the department should evaluate the steps in its hiring process to identify potential hurdles that may be disproportionately affecting female applicants.

The departments vary in the degree to which their recruitment staff evaluate the first set of diversity‑related data: the demographics of their current personnel. ³ Because best practices call for the assessment of the diversity of current personnel to inform departments' recruitment strategies, we considered only demographic assessments reviewed by the departments' recruitment units. In line with best practices, Stockton Police annually evaluates both the race and gender of its officers, and had documented in a 2021 recruitment plan update an evaluation of the diversity of its full-time staff as a means to measure its progress toward its diverse staffing goals. San José Police's recruitment unit has reported information on the diversity of its current personnel and police academy trainees to a subcommittee of the city council, which the department indicated it began doing in February 2021 at the direction of that subcommittee. CDCR also recently began reviewing these data. According to the manager who oversees CDCR's recruitment unit, CDCR began reviewing the demographic makeup of its officers in May 2021. Further, CDCR's draft recruitment plan states that CDCR will evaluate reports on this information quarterly to assess the effectiveness of its recruitment strategies and identify areas for improvement.

Conversely, the recruitment units at Los Angeles Sheriff and San Bernardino Police do not analyze the demographics of their current personnel. The sergeant who oversees recruitment at Los Angeles Sheriff indicated that the reason the recruitment unit has not analyzed the demographics of its personnel in recent years was that the unit was downsized. Although the lieutenant who oversees the recruitment unit at San Bernardino Police agreed that reviewing the demographic data for its personnel would be useful, he stated that collecting and maintaining those data would be difficult because of the constant change in personnel and challenges in getting individuals to respond to questions about their demographics. However, we obtained and reviewed the city of San Bernardino's personnel data for San Bernardino Police's officers and found that the city possesses information on the race, ethnicity, and gender of all officers.

CDCR analyzes the second set of diversity-related data: the demographics of its applicant pool. CDCR performs this analysis despite its lack of certain demographic data, which we describe later. However, the other departments have either not done so regularly or have not analyzed the data at all. The captain who oversees Stockton Police's recruitment unit acknowledged that the department does not have a standardized process for analyzing applicant data, but provided evidence that around January 2021 it analyzed the diversity of its applicants from three different rounds of recruitment. Its analysis contained not only the gender, race, and ethnicity of applicants, but also the stages of its hiring process at which those applicants either voluntarily withdrew from consideration or were screened out.

Los Angeles Sheriff's recruitment plan states that the recruitment unit will, on a weekly basis, receive information on the number of applications the department received, including a demographic breakdown of the applicants based on gender, race, and age. However, the recruitment unit for the department could not provide evidence of any review of that information. The sergeant who oversees Los Angeles Sheriff's recruitment unit indicated that it had not done so from 2019 through 2021, in part because the department did not accept applications in 2021, and in part because the recruitment unit was downsized. Similarly, neither San Bernardino Police nor San José Police had analyzed the diversity of their applicants. The lieutenant who oversees San Bernardino Police's personnel and training division agreed the practice would be beneficial, and the lieutenant who oversees recruiting at San José Police indicated that it could begin performing this analysis.

One barrier to evaluating applicant data is that three of the departments—CDCR, San José Police, and San Bernardino Police—possess insufficient data on either the race and ethnicity or the gender of their applicants to allow these departments to perform a meaningful analysis. State law prohibits departments from requiring applicants to report characteristics such as their race, ethnicity, and gender. However, departments are permitted to ask applicants to volunteer that information. Each of the five departments requests this information during their hiring processes, but they have had varying degrees of success in obtaining the data. As Table 3 shows, Stockton Police and Los Angeles Sheriff each possessed data on the gender, race, and ethnicity of their applicants. However, San José Police lacked data on the gender of the majority of its applicants, and both CDCR and San Bernardino Police similarly lacked data on the race and ethnicity of the majority of their applicants. We reviewed each department's application forms to compare the methods each used to request demographic information from applicants and found that the departments' approaches differed. Given the success that some departments have had in obtaining information from applicants about their race, ethnicity, and gender, it may benefit CDCR, San Bernardino Police, and San José Police to evaluate the manner in which they request that information from applicants and consider changes to improve the rate at which applicants provide it.

Table 3

Three of the Five Departments We Reviewed Lack Data About the Race and Ethnicity, or Gender of Their Applicants

	CDCR	Los Angeles Sheriff	San Bernardino Police	San JosÉ Police	Stockton Police
Percentage of applicants who reported race/ethnicity	16%	99%	6%	90%	97%
Percentage of applicants who reported gender	100	99	99	22	99

That said, the lack of certain demographic information does not fully explain why San Bernardino Police and San José Police have not analyzed the diversity of their applicant pools. As Table 3 shows, applicants at San Bernardino Police generally report their gender, and the majority of San José Police's applicants have reported their race or ethnicity. As a result, these departments can, at minimum, analyze these data.

By not establishing robust processes for monitoring the diversity of both their current personnel and their applicants, the local departments we reviewed have limited their ability to track the success of their recruitment efforts and to identify areas requiring improvement. Until the departments develop such processes and implement best practices for recruiting a diverse workforce, they will likely continue to struggle to ensure that their officers reflect the demographics of their communities and to reap the benefits that such diversity provides. In Chapter 3, we present examples of how departments can engage more with their communities as part of their recruitment efforts.

The Departments Have Inconsistently Implemented Hiring Practices That Can Identify Applicants Who Have Disqualifying Biases

Strong hiring processes are critical to ensuring that law enforcement departments identify applicants' detectable biases. State law and POST regulations require local law enforcement departments to perform multiple hiring steps, including several—listed in the text box—that may reveal that applicants hold biases that disqualify them from serving as officers (disqualifying biases). POST and IACP have also published additional beneficial steps that law enforcement departments can take to provide further assurance that applicants do not have disqualifying biases.

CDCR is not required to follow POST regulations, which govern only participating law enforcement departments. State law assigns authority for developing, approving, and monitoring correctional officer selection standards to the Commission on Correctional Peace Officer Standards and Training (CPOST). Despite having this responsibility since at least 2015, CPOST has not formalized any hiring or screening standards. CPOST leadership attributed the lack of standards to a number of factors, including the broad responsibility of CPOST in contrast to its modest staffing, and the COVID-19 pandemic. A captain in CDCR's Office of Peace Officer Selection stated that it has instead followed POST's guidance.

When we reviewed five hiring decisions at each department, we found that all of the departments generally followed key hiring requirements from state law and regulation, as Table 4 shows. ⁴ Although state law requires CDCR to use POST standards as guidelines when completing background investigations, it is not required to follow the POST regulations that we reviewed. Nonetheless, we assessed whether it completed these steps because they are best practices for hiring officers. Further, each took additional screening steps beyond those that are required that could help them identify disqualifying biases. For example, the five departments all solicited information about applicants' tattoos. Because members of certain hate groups identify themselves through tattoos of specific symbols or logos, examining an applicant's tattoos can reveal that an applicant is associated with a hate group. Additionally, both San José Police and Stockton Police used a polygraph or similar examination during which examiners asked the applicants at least one question that could identify potential biases, including whether they are involved with or sympathize with hate groups. Despite these additional steps and their general adherence to state law, the five departments varied in their implementation of three other key practices that would have assisted them in more rigorously screening their applicants for disqualifying biases.

Table 4

Law Enforcement Departments Inconsistently Followed Key Steps for Screening for Bias During Hiring

Key Practice	CDCR	Los Angeles Sheriff	San Bernardino Police	San JosÉ Police	Stockton Police
Generally follows key requirements in state law and regulations	Adequate	Adequate	Adequate	Adequate	Adequate
Uses a structured oral interview that includes a question about an applicant's ability to interact with a diverse population	Poor	Adequate	Partial	Adequate	Poor
Contacts secondary references who can provide more candid information about an applicant than the references supplied by the applicant	Poor	Not applicable*	Adequate	Partial	Partial
Conducts social media reviews in which the investigator looks for indications of bias, prejudice, or hate group affiliation	Partial	Partial	Partial	Partial	Partial

Not All Departments Used Interviews to Assess Applicants' Ability to Work With a Diverse Community

Only Los Angeles Sheriff and San José Police effectively leveraged hiring interviews to identify whether applicants had experience with or were capable of working with diverse groups of people. POST regulations require departments to hold a structured oral interview with applicants. These interviews involve asking applicants predetermined questions in a standardized fashion and evaluating their responses against predefined, job-relevant criteria. POST requires that the departments' structured interviews address, at minimum, the six factors listed in the text box. Its interview guidance details that the community involvement and awareness factor includes elements that can aid departments in determining whether an applicant possesses disqualifying biases. These elements include an applicant's experience with and interest in community issues, freedom from social and ethnic prejudices, and sensitivity to and acceptance of differences based on demographic background.

Despite the benefit that structured oral interviews can provide, neither Stockton Police nor CDCR conducted them. The captain who oversees Stockton Police's hiring process stated that he and senior staff believe that other components of its hiring process sufficiently address the community involvement and awareness factor. Nevertheless, POST's requirements direct departments to assess an applicant through a structured oral interview and do not make allowances for departments to meet this requirement through other means. A captain in CDCR's Office of Peace Officer Selection stated that CDCR does not currently hold interviews because of the staffing and resources such an interview would take, but agreed that there would be value in doing so.

San Bernardino Police, San José Police, and Los Angeles Sheriff all conducted structured interviews of their applicants. However, only Los Angeles Sheriff and San José Police's interviews regularly contained a question designed to determine whether the applicant could work effectively with a diverse community. San José Police asked such a question in four of the five hiring decisions we reviewed, with the only exception being the oldest hiring decision we reviewed, which was for an applicant it interviewed in August 2019. Los Angeles Sheriff also asked applicants to respond to a question that was designed to assess their ability to interact with a diverse community, and it also provided interviewers with guidance that assisted them in this assessment. However, for the most recent interview we reviewed, which occurred in 2020, Los Angeles Sheriff revised this question so that it no longer specifically addresses whether the applicant has the ability to work with diverse individuals. The manager in the unit that develops interview questions at Los Angeles Sheriff indicated that the department changed the interview questions because of concerns about how long the previous questions had been in use, but agreed that it would be beneficial to reincorporate an assessment about applicants' ability to work with diverse groups of people.

In contrast to Los Angeles Sheriff and San José Police, San Bernardino Police's interview questions in this area were broad. For example, one question was a general inquiry into the applicant's experience working with the public. Although applicants could answer these questions in a way that revealed their ability to work with diverse groups of people, the questions would not necessarily lead applicants to share their experiences in that area. Additionally, San Bernardino Police did not develop guidance to inform interviewers that the purpose of these questions was to obtain information about an applicant's ability to work with diverse groups of people.

Most of the Departments Did Not Consistently Contact Secondary References

Four of the five departments generally did not follow another key practice: contacting certain individuals who could provide valuable information about an applicant. POST regulations require departments to contact each applicant's relatives, neighbors, and other applicant-provided personal references. In the hiring decisions we reviewed, all of the departments contacted these individuals and—with the exception of CDCR, which we discuss later in this section—typically asked some or all of these references whether applicants had ever exhibited biased or discriminatory behavior. However, the departments did not consistently contact additional individuals known as secondary references. State regulations describe secondary references as contacts provided to a department by the applicant's initial references and requires that departments contact these individuals during the hiring process. POST's guidance also emphasizes the value of secondary references, stating that they are sometimes more candid than references the applicants provide.

The departments varied in the degree to which they obtained and contacted secondary references. ⁵ In response to the departments' requests for secondary references, references sometimes named individuals whom the applicant had already listed as references. POST's background investigation manual describes secondary references as individuals not listed by the candidate as a reference; therefore, we did not consider individuals whom the applicant had listed as a reference to be secondary references. San Bernardino Police obtained secondary references for all five hiring decisions we reviewed, and contacted them for four. Conversely, Los Angeles Sheriff did not request secondary references from the references that investigators contacted, and did not receive any. San José Police and Stockton Police each obtained secondary references but contacted them only half of the time, and although CDCR obtained secondary references for four of the five hiring decisions we reviewed, it did not contact any of them. Stockton Police's captain speculated that we could not verify that it contacted secondary references due to the lack of documentation of this process. Nonetheless, the captain acknowledged the value that these contacts can add and stated that the department was open to documenting this process in the future. By not contacting secondary references, the departments forgo an opportunity to gain more complete information about applicants, including about whether the applicants have a history of biased behavior.

Finally, unlike the four local law enforcement departments, CDCR did not ask the references whom it contacted whether applicants had ever demonstrated biased or discriminatory behavior. POST's background investigation manual states that asking references about racial or ethnic biases is useful, and includes "Has the applicant displayed any bias or prejudice toward others?" in its sample reference questionnaire. CDCR asked references other helpful questions related, for instance, to the applicant's temper and emotional stability, but it did not inquire about applicants' ability to work with diverse individuals or about indications of disqualifying biases. An applicant's family, friends, neighbors, and current and former employers could provide valuable information about that individual's attitudes, past behaviors, and association with hate groups, all of which could reveal disqualifying biases. CDCR's chief of peace officer selection indicated that incorporating inquiries about bias into the questionnaire it uses when interviewing references would be a simple step that would provide value.

Each Department Needs to Strengthen Its Review of Applicants' Social Media

A final practice for identifying disqualifying biases that the departments did not adequately implement is a thorough review of the applicants' social media. Social media reviews can provide several benefits to departments, one of which is that they may reveal applicants who are affiliated with hate groups or hold the type of biased views that we report about in Chapter 1. In partnership with the Defense Personnel and Security Research Center—a division of the U.S. Department of Defense that conducts research and publishes best practices for effective personnel screening—the IACP published guidance for law enforcement departments on how to use social media to screen officer applicants and incumbents for potentially problematic behavior, including the sharing of posts like those we describe in Chapter 1. The IACP guides departments to review applicants' social media accounts and other online activity to ensure that they have behaved in a manner appropriate for the officer position. The guidance also provides some strategies for conducting a review and advises using applicant email addresses as search terms for accounts.

Although each of the departments we reviewed incorporated some assessment of applicants' social media accounts into its hiring process, most reviews were limited and their rigor was unclear based on the records the departments kept. Each department asked its applicants to report their social media accounts or usernames. However, only San José Police documented a broader search for applicant social media accounts, such as by using applicant email addresses or usernames to identify social media accounts that the applicants may not have disclosed. CDCR generally appeared to limit its reviews of social media accounts to one social media site. The remaining three departments all either did not document the social media sites they reviewed, or appeared to limit their reviews to the applicants' self-reported accounts. The limited social media reviews that CDCR, Stockton Police, San Bernardino Police, and Los Angeles Sheriff performed narrowed the departments' ability to learn about disqualifying biases.

Additionally, none of the departments except for CDCR had formalized guidance for their staff on how to conduct effective social media reviews. CDCR provided evidence that it had developed a social media investigation course for its background investigators, which includes strategies for navigating social media accounts and for how to locate applicants on social media. However, the training does not include guidance for how to identify or recognize biased content. The absence of such guidance, combined with the fact that the records the departments retained of their social media reviews were often limited to a note saying that the review had occurred, rendered us unable to verify that the reviews had been done with a consistent level of depth or with consistent consideration of the content.

When we performed the social media investigation that we describe in Chapter 1, we included the officers hired through the five hiring decisions we reviewed at each department. Of those who had public social media accounts, one applicant's account contained biased content at the time we conducted our review. Because of the limitations in the departments' documentation of their social media screenings that we describe above, we could not determine whether the department observed or considered this information during its background investigation. Nonetheless, stronger guidance and direction regarding the process that background investigators should use when reviewing applicants' online activity would increase the likelihood of their detecting online behavior that is indicative of potentially harmful biases.

One potential barrier to law enforcement departments performing in-depth reviews of applicants' social media is that state law prohibits employers from requiring or requesting that employees or applicants disclose their usernames or passwords for the purpose of accessing their personal social media. The statute does not clearly state whether that prohibition applies to employers that are state or local public departments, including law enforcement. Given the value that reviewing an applicant's public social media accounts provides in screening the applicants for a law enforcement position, an amendment to that statute to clarify that law enforcement departments are permitted to ask applicants to disclose their social media accounts for the purpose of employment screening would benefit those departments' screening practices.

Changes to POST's Oversight Would Better Ensure That Departments Adhere to Key Steps in Hiring

Although it regulates law enforcement hiring, POST has not conducted reviews to ensure that law enforcement departments are adequately performing the steps we describe above. State law requires POST to perform any inquiries necessary to determine whether law enforcement departments are adhering to hiring standards, and POST conducts regular audits of local law enforcement departments' hiring practices. POST audited each of the four local departments at least once from January 2020 through January 2022. However, POST's audits do not include a review to ensure that departments have conducted structured oral interviews and that the interviews adequately address the six factors that regulations require. Further, although POST verifies during its audits that departments have contacted references, it does not determine whether departments have contacted secondary references. The executive director for POST agreed that information from secondary references is critical, and that those references generally provide more candid information. He stated that POST could easily make an adjustment that would allow POST's staff to quickly verify during its audits that departments interviewed secondary references. Including in its audits a review of departments' compliance with these steps would allow POST to increase department accountability for adhering to key hiring requirements.

Please refer to the recommendations section to find the recommendations that we have made as a result of these audit findings.

---

Chapter 3

The Five Law Enforcement Departments Have Not Implemented Key Best Practices for Mitigating the Effects of Officer Biases

Key Points

• Training can help officers better understand and address their own biases, but each department's training about bias needs to include additional content to be most beneficial. Also, none of the five departments require their officers to attend training about bias as frequently as law enforcement authorities recommend.
• Building strong relationships with the community is a critical step to mitigating bias, but the four local departments have not taken key steps for building and strengthening relationships with their communities.
• None of the departments we reviewed have established adequate systems for proactively identifying and correcting problematic trends in officer performance. These systems can be a critical tool for identifying potentially biased conduct and providing support or training to correct that behavior.

The Five Departments Have Not Provided Frequent Enough Training to Mitigate Officer Bias

Multiple authorities on law enforcement practices have indicated that training officers about bias is a key strategy for mitigating its effects. An effective training program reinforces expectations regarding fair and impartial law enforcement, improves officers' understanding of the effects of bias and ways to mitigate those effects, and strengthens officers' ability to interact effectively and respectfully with the diverse individuals in their community. The US DOJ, IACP, and other law enforcement experts have published best practices for the frequency with which officers should receive training about bias, as well as the content and approaches departments should include to ensure that those trainings are effective.

Despite the importance of effective training, the departments we reviewed have not required that training as frequently as recommended. For example, multiple sources recommend that departments provide officers with frequent training related to bias. Frequent training on bias can reinforce key principles and ensure that officers are up to date on current circumstances, applications, and methods for bias-free policing. However, rather than following this guidance, the departments we reviewed have generally required officers to meet the lower threshold set in state law. ⁶ Los Angeles Sheriff requires officers assigned to custodial duties in the county jails to participate in more frequent trainings that address how the officers should interact with incarcerated individuals from certain vulnerable populations, such as those who are LGBTQ. This threshold requires officers to participate in training about racial and identity profiling and bias only once every five years, although it allows departments to provide that training more frequently if needed. In contrast, POST's regulations require certain officers to receive 24 hours of training every two years on other important topics, such as driving, making arrests, and communication.

Three of the departments we reviewed—Los Angeles Sheriff, San José Police, and Stockton Police—have made additional trainings available to their officers, but none of those departments require the additional trainings at regular intervals. For example, in addition to the training required by state law, Stockton Police began requiring in 2014 that its officers take a three-course series that discusses implicit bias, as well as topics such as treating citizens fairly and respectfully. However, Stockton Police shared that it requires its officers to take this series of courses only once during their employment and does not require other regular training on those topics. From 2014 through 2018, San José Police required each of its officers to attend three one-time trainings that included discussions about implicit bias, meaning that during that time period San José Police required frequent training for its officers. However, the department has not required training to continue occurring at this frequency. Los Angeles Sheriff offers a voluntary course about bias up to twice a year, but only 15 of its more than 10,000 sheriff's deputies attended this course from 2018 through 2021.

Most of the local departments agreed that more frequent trainings on bias would be beneficial. Each of the departments referred to its academy training as an example of where relevant content on bias is also taught to officers. Although academy training is important, these are generally one-time trainings required before an individual is sworn in as an officer. As we discuss earlier in this section, frequent training on specific content related to bias reinforces the expectation for bias-free policing and strengthens officers' ability to interact effectively and respectfully with the diverse individuals in their community.

Departments also indicated that additional trainings would be cost-prohibitive. Regulations that POST has issued provide for departments to receive reimbursement from the State for all POST‑certified courses that officers take while on-duty, including mileage, lodging, meals, and tuition. Generally, training expenses for a required course may be reimbursed only once for a given officer, unless the officer is authorized to repeat the course periodically. In other words, departments may receive reimbursement only once every five years for expenses related to training on racial and identity profiling and bias. San José Police claimed that, due to staffing shortages, additional training would need to be taught during overtime hours. According to the department's calculations, it costs an estimated $130,000 per hour to train every officer, sergeant, and lieutenant. For example, a two-hour training would cost the department about $260,000.

Local law enforcement departments are unlikely to offer training about bias more frequently without change at the state level. The Legislature could address this issue by aligning the requirements for training about bias with those that already exist for training on other vital subjects—driving, communicating, and making an arrest. This new requirement would likely represent an increased cost for the State. However, bias has the potential to negatively affect many of the activities in which officers engage on a daily basis. For this reason, we believe the additional cost is worthwhile.

Beyond the inadequate frequency of their training, the departments we reviewed vary in the quality of their training content. To adhere to best practices, law enforcement departments should incorporate into their trainings the key topics that Table 5 lists. As the table shows, the local departments we reviewed adequately explain bias: providing relevant definitions of implicit and explicit bias, describing scientific research supporting that all individuals possess bias, and discussing the effects that bias can have on officer behavior. CDCR did not include this content because—until 2022—it did not require its officers to attend a training that specifically addressed these issues.

However, none of the departments incorporated adequate training on cultural awareness and sensitivity. The US DOJ notes that law enforcement departments should educate their officers regarding the various cultures within their communities so that officers are better prepared to engage with and respond to residents in a manner that is situationally appropriate. However, the law enforcement departments do not consistently provide this type of training. San José Police's training includes content about people who are Muslim and on the importance of respectful treatment of people who are LGBTQ. In alignment with guidance from the RIPA Board, San José Police also requires an in-depth training on the impact that past negative experiences with law enforcement can have on interactions with some groups, including Black individuals. The training includes locally relevant incidents and history. However, we expected San José Police to include in its training information about additional cultures, religions, races, and ethnicities that represent the San José community. For example, we did not identify in its ongoing trainings education on various Asian cultures, despite the fact that Asians represent almost 40 percent of the city's population. As a result, its officers may not be as prepared to interact with the community they serve and in a manner that is situationally appropriate.

In our review of the departments' internal investigations, which we describe in Chapter 4, we identified times when officers might have benefited from additional training on cultural awareness and sensitivity. For example, one officer made several inappropriate comments to a Black man that indicated a problematic lack of awareness about the historical experiences of Black people with law enforcement, and about the impact that racial slurs can have on people of color. The officer accused the man of playing the "race card," stated that the officer wished "we lived in a world, back in the [19]60s and [19]70s, where we could feel comfortable," and listed aloud a series of derogatory slurs as examples of protected speech. Without sufficient training to improve officers' understanding of different cultures, they may lack critical knowledge, leading to misperceptions and miscommunications that negatively affect their interactions with individuals from different cultures.

Table 5

Law Enforcement Departments Did Not Consistently Implement Best Practices for the Content of Bias Training

Recommended Content	CDCR	Los Angeles Sheriff	San Bernardino Police	San JosÉ Police	Stockton Police
Explanations of bias, including scientific research on implicit and explicit bias and content on how bias can influence behavior	Poor	Adequate	Adequate	Adequate	Adequate
Community engagement strategies, including the benefits of effective community engagement and the means to achieve that engagement	Not applicable*	Poor	Partial	Partial	Partial
Cultural awareness and sensitivity, including the various cultures and identity characteristics within the communities they serve	Poor	Partial	Poor	Partial	Poor
Reporting obligations, including how officers should respond after observing biased behavior by peers	Partial	Adequate	Poor	Poor	Poor
Realistic or actual scenarios that allow officers to develop skills that can be applied in situations they may come across in their day-to-day work	Partial	Adequate	Adequate	Adequate	Adequate

The Four Local Law Enforcement Departments We Reviewed Have Not Implemented Best Practices for Building Trust With Their Communities

Best practices for safeguarding against bias emphasize the importance of local law enforcement departments' engaging with the communities they serve to build collaborative, trusting relationships. ⁷ We did not include CDCR in this section because the vast majority of its officers' interactions are with incarcerated individuals, rather than with the local communities in which its prisons are located. Both officers and their communities benefit when they interact frequently in positive settings. Research has shown that officers' day‑to‑day work may actually promote a belief in negative stereotypes. Policing inherently involves dealing with criminal and other problematic behavior. If officers are consistently exposed to negative contacts within a given community, they may increasingly believe negative stereotypes regarding the members of that community. Positive interactions between those officers and members of the community help to dispel such biases. Further, these interactions may also dispel any negative stereotypes that the community may believe regarding officers.

Each of the local departments we reviewed has taken some encouraging steps toward establishing positive relationships with its community. The departments all host and participate in events aimed to increase their presence within the community, including various cultural celebrations at which members of the community can interact with officers. The departments also operate programs for youth. For example, San José Police participates in the Police Activities League, a goal of which is providing athletic and educational programs for children. In addition, Stockton Police has hosted reconciliation sessions, at which the chief of police and other leadership staff have met with members of minority and other marginalized communities to discuss the detrimental effects of those communities' experiences with law enforcement departments. However, none of the local departments have fully implemented key best practices that are critical for effectively building trust with the community, as Table 6 shows and as we discuss in the sections that follow. ⁸ In February 2022, the city of San José published the results of an assessment it commissioned of San José Police's adherence to best practices for policing. A significant portion of that assessment addressed the way that San José Police interacts with its community, and the assessment's findings generally mirror our conclusions.

Stockton Police Is the Only Local Department We Reviewed With a Comprehensive Community Engagement Plan

Other than Stockton Police, none of the local departments we reviewed have a comprehensive community engagement plan that institutionalizes their strategies for engaging with the community. According to the US DOJ, a community engagement plan includes strategies for partnering with community members, nonprofits, and other nongovernment organizations. These plans help departments to establish and communicate expectations regarding their community engagement, focus their efforts on areas where they would like to improve, and establish measurable goals to help them to monitor their progress.

Table 6

Local Law Enforcement Departments Did Not Fully Implement Best Practices for Effective Community Engagement

	Los Angeles Sheriff	San Bernardino Police	San José Police	Stockton Police
Formalize Expectations and Strategies for Community Engagement
Establish formal expectations regarding participation by officers in community engagement activities	Adequate	Poor	Partial	Poor
Create a comprehensive community engagement plan that includes strategies for engaging community members and setting clear, measurable goals	Poor	Poor	Poor	Adequate
Require Positive Interactions With the Community
Host or participate in community events, including those for diverse community groups	Adequate	Adequate	Adequate	Adequate
Track officer participation to ensure that all officers have positive interactions with diverse groups within the community	Poor	Poor	Poor	Poor
Solicit Feedback From the Community
Form community advisory committees to provide perspectives on operations	Adequate	Poor	Adequate	Adequate
Involve the community in key functions, including policy development, hiring and recruitment, training, and community engagement strategies	Partial	Poor	Partial	Poor
Regularly survey the community on perceptions of fear, safety, trust, and fairness in policing	Partial	Poor	Partial	Partial

Source: Review of best practices from sources such as US DOJ and IACP, and analysis of the steps taken by the departments in the table.

Stockton Police's strategic plan establishes a goal to increase trust between the department and its community. One strategy it lists to achieve this goal is conducting trust-building workshops, where members of the community and members of the department convene to promote healing through conversations, such as the reconciliation sessions we mention earlier. Another strategy involves deploying a neighborhood impact team—consisting of a public information officer and other members of the department—into the community after a traumatic event, such as a homicide. The plan states that the impact team can offer support to affected community members, in part by providing referrals to available resources. According to the officer responsible for the team, Stockton Police provides information about the San Joaquin County District Attorney's Office Victim-Witness Program, which that office advertises as providing a variety of services for crime victims, including crisis intervention.

In contrast, Los Angeles Sheriff, San Bernardino Police, and San José Police all generally lack formalized plans for building trust with their communities. Los Angeles Sheriff requires its service areas to develop location-specific community policing plans. We reviewed the plans that two service areas have developed and found that they lack key elements. For example, neither plan includes measurable goals to monitor progress in community engagement. Los Angeles Sheriff's deficiency in this area is particularly notable in light of its history. In 2011 US DOJ began an investigation of Los Angeles Sheriff's operations in its Antelope Valley service area, because of allegations of discriminatory policing by Los Angeles Sheriff. To resolve the investigation, Los Angeles Sheriff entered a settlement with US DOJ in 2015 in which it agreed to develop a plan for its officers' attendance at community meetings and events in the Antelope Valley service area, among other things. The settlement requires that the plan account for the need to improve the department's community relationships, including with communities of color, and that it indicate the number and types of community engagement events that department representatives will attend on a regular basis. In January 2020, the Antelope Valley service area revised a crime prevention and community engagement plan that it indicated was developed to respond to the settlement agreement. Although the US DOJ highlighted the benefits that this plan would provide to the department and the community, based on our review of other service area plans, Los Angeles Sheriff has not ensured that this beneficial practice is adopted throughout the county.

In the absence of documented plans for engaging with their communities, Los Angeles Sheriff, San Bernardino Police, and San José Police have no formalized expectations, targeted goals, or accountability mechanisms for building trust with the individuals they serve. Therefore, those departments' various community engagement events are not structured around any dedicated strategy, and they lack actionable information on the success of those efforts.

Three of the Four Departments Have Not Established Adequate Expectations That Their Officers Participate in Community Engagement Activities, and None Comprehensively Track This Participation

If a department does not ensure that all of its officers participate, the benefits of community engagement efforts are limited. The US DOJ suggests that one way departments can institutionalize community engagement is by tracking and rewarding positive interactions between officers and community groups. These positive interactions can occur through organized community events or informally through an officer's daily patrol activities, such as playing basketball with community members or patronizing local businesses. However, only Los Angeles Sheriff has a formal policy for officer participation in community engagement. The other three departments have either inadequate policies or no formal expectation for participation in community engagement efforts. Although the departments could demonstrate some participation by various officers in these events, none comprehensively tracked that participation or could otherwise demonstrate how many officers had participated in a given period or how frequently they had done so.

One of Los Angeles Sheriff's service areas demonstrates that tracking officer participation in community engagement is feasible. In the 2015 US DOJ settlement with Los Angeles Sheriff, the Antelope Valley service area agreed to determine the number and types of events that all officers must regularly attend. According to reports by the service area and the team monitoring the implementation of the agreement, since the settlement the service area has set a minimum requirement for officer participation in community engagement activities and tracks this participation. However, this tracking does not apply to all of Los Angeles Sheriff's service areas. According to the captains of two other service areas, the captains and those responsible for community engagement primarily attend the community engagement events and patrol deputies are asked to attend the events when available, limiting the effectiveness of these events, as not all deputies are in attendance. Further, neither service area tracks officer participation in community engagement activities.

The four local departments agreed that community engagement is an important aspect of effective law enforcement, but not all agreed that they should track officers' participation in such activities. For example, the captain that oversees the special operations division at Stockton Police stated that he is skeptical about setting an expectation for a specific number of community engagement activities and tracking these activities because he worries that officers would perceive this requirement as another box to check rather than understanding and being committed to the value of community engagement. However, by setting clear expectations and tracking officer participation, a department can clearly demonstrate and communicate its commitment to community engagement.

The Departments Have Not Adequately Involved or Sought Feedback From Their Communities

Guidance from the US DOJ states that, because a community is the primary recipient of law enforcement services, it has a vested interest in and a unique perspective on what constitutes effective law enforcement. Thus, best practice suggests that law enforcement departments should involve their communities in key departmental processes as part of their efforts to build trust in their communities. For example, law enforcement departments should involve their communities in recruitment and hiring processes by soliciting their communities' input on the qualities and skills officers should possess. The RAND Corporation also suggests that law enforcement departments use recommended strategies for increasing diversity in the U.S. military, such as partnering with organizations that work directly with minority groups to help increase the pool of eligible recruits. Similarly, the RIPA Board recommends that departments include in their trainings representatives of the community, particularly individuals knowledgeable about issues of race, religion, sexual orientation, and disability.

One way that best practices recommend that departments involve their communities is by forming community advisory committees—groups composed of representative members of the community who can provide advice regarding the department's development of policies and of recruitment, hiring, and other strategies. Although San Bernardino Police does not use this approach, Los Angeles Sheriff, San José Police, and Stockton Police have done so. For example, San José Police has established two different committees: a general community advisory board and an LGBTQ advisory board. San José Police shared agendas that showed how they partnered with the LGBTQ advisory board to discuss the department's attendance at related community events, implementation of LGBTQ-specific trainings, and development of both a training bulletin and policy changes related to officer interactions with LGBTQ members of the community. Finally, a public report states that Stockton Police has included a community advisory board in the development of policies, but the department could not provide evidence that the board's involvement had occurred.

Nonetheless, none of the four local departments could demonstrate that they had fully involved their communities in the key areas recommended by best practices. For example, although San José Police and Los Angeles Sheriff sometimes involved community representatives in their trainings, they did not do so consistently. The other two departments did not involve their communities at all. San Bernardino Police agreed that including community members in the development of training on bias would be beneficial and would ensure that the trainings cover cultural awareness and sensitivity that is relevant to its community. Stockton Police asserted that although it does not involve community members in its training, it has a diverse team of instructors from various backgrounds. Nonetheless, by not fully involving their communities in the development and operation of key law enforcement functions—in particular those related to addressing bias—the departments have not taken a critical step toward building strong relationships of trust with their communities.

Another key best practice that law enforcement departments should use to engage their communities is regular surveys. US DOJ guidance states that surveys are a valuable tool through which a department can better understand community fears, concerns, and perceptions of its performance. However, none of the local departments we reviewed regularly survey their communities. Los Angeles Sheriff has surveyed select segments of its community in the past; however, it stated that it has not conducted any community surveys since 2018, when it surveyed the community served by the East Los Angeles substation. That survey provided valuable information, including the community's degree of satisfaction with and primary concerns about law enforcement. Because of concerns expressed in that survey, the East Los Angeles substation indicated it would conduct more patrols, increase its presence in the community, and work on community outreach programs.

Similarly, Stockton Police has benefited from community surveys in the past but has not committed to conducting them regularly. As part of a 2014 national project funded by US DOJ, researchers conducted two community surveys—one before and one after Stockton Police implemented new initiatives to increase community trust. The results of the second survey, which researchers administered in 2017, showed a marked improvement in community perceptions of Stockton Police's policing. However, the survey also showed no significant improvement in community perceptions of police bias. According to Stockton Police, it attempted to conduct two additional community surveys between 2019 and 2020 through a contract with a third party. However, the department did not receive enough responses in 2019 to generate a usable report and was not able to conduct the 2020 survey because of the pandemic. As a result, Stockton Police has not surveyed its community since its last attempt in 2019 nor has it considered plans to resume these surveys, limiting its ability to determine whether its continued efforts to build relationships with its community have been effective.

Some of the departments stated that they face resource challenges in conducting regular community surveys. However, without such surveys, the departments lack key insights into community perceptions and the areas in which they need to build trust. Available guidance suggests methods for minimizing the number of people needed for conducting these surveys, such as partnering with local research institutions or universities to conduct the survey and analyze the results.

The Five Law Enforcement Departments We Reviewed Lack Adequate Systems for the Early Identification and Correction of Problematic Behavior

Alongside their other efforts to mitigate the effects of bias, law enforcement departments can proactively monitor key information about officer conduct to identify trends or signals that indicate an officer may need support. One recommended approach for this type of monitoring is the use of an early intervention system. In general, these systems entail collecting information about officers' behavior; analyzing that information to identify patterns, both positive and negative; assessing whether those patterns indicate a need to intervene; and providing supports or corrective actions when necessary. These systems address a wide range of conduct, including behaviors that may be the product of officers' biases. For example, the system should identify if an officer receives multiple complaints regarding biased conduct or disproportionately uses force against certain demographic groups.

Early intervention systems are an important component of departments' efforts to monitor and improve officer performance. Departments that do not proactively identify and correct officers' behavior can miss opportunities to provide important supports, coaching, and training to improve officers' performance before it harms members of the public or the officers themselves. In fact, some research has shown that these systems can reduce negative outcomes, such as complaints about an officer as well as officers' use of force against members of the public.

Despite the benefits, none of the departments we reviewed have established early intervention systems that are fully aligned with best practices, as Figure 7 shows. For example, San Bernardino Police does not have a formalized approach to early intervention. The department flags certain trends in officers' behaviors—such as how often officers use force—and, according to the former manager of its professional standards bureau, once an officer's conduct triggers an alert, either the professional standards bureau or the officer's supervisor will review the events that led to the alert and determine whether someone needs to meet with or provide support to the officer informally. However, the sergeant in charge of San Bernardino Police's professional standards bureau indicated that because the department does not have a formal policy on early intervention, it does not document any informal follow-up efforts that may result from these alerts. Without a formal approach, San Bernardino Police cannot ensure that it consistently identifies officers who could benefit from intervention.

Figure 7

The Law Enforcement Departments We Reviewed Lack Adequate Systems for Intervening With Officers Who Demonstrate Indicators of Bias

During our review, San José Police and CDCR were both in the process of significantly modifying their early intervention systems, the current versions of which are limited and not well documented. The new systems appear promising; for example, San José Police is conducting a pilot project in partnership with an entity that offers a predictive model to determine officer risk levels, followed by actionable alerts and interventions. However, because San José Police and CDCR had not implemented these modified systems at the time of our review, the remainder of this section focuses on our assessment of the early intervention systems at Los Angeles Sheriff and Stockton Police.

Los Angeles Sheriff has had an early intervention system, which it refers to as its performance mentoring program, since the 1990s as a result of external oversight measures. In response to an increase in shootings by Los Angeles Sheriff officers, the county board of supervisors in December 1991 appointed a special counsel to conduct a review of Los Angeles Sheriff's policies, practices, and procedures. The special counsel's resulting report described a number of problems, such as excessive uses of force, and made several recommendations, including that Los Angeles Sheriff develop an early intervention system. Los Angeles Sheriff has since developed and implemented that system, which has been the subject of subsequent reviews by the special counsel. Likely because of that oversight, its system was the most robust we reviewed, although it had limitations that we describe below.

To implement effective early intervention systems, departments must first collect quality information about officers' behavior. One type of data is especially critical for identifying potentially biased behavior: data on the demographics of individuals whom officers stop, detain, search, arrest, or take other actions against (stops data). Stops data are useful because they encompass common interactions officers have with the public. ⁹ Because CDCR officers work in custody settings, stops data are more relevant to the local departments. However, in addition to uses of force and complaints, other data on interactions between officers and incarcerated individuals—such as disciplinary actions that officers initiate against incarcerated individuals—may also benefit CDCR's early intervention system. Other data, such as complaints data and use‑of‑force data, represent far less frequent encounters. According to the RIPA Board, the 15 largest law enforcement departments in California reported almost 4 million total vehicle and pedestrian stops in 2019, compared to just 41,000 uses of force and 7,200 public complaints. In other words, these data suggest that at least 98 percent of reported stops did not include a reported use of force or public complaint. Because of the quantity of stops data, they can reveal evidence of possible bias that would be impossible to identify in the review of specific complaints or isolated incidents. ¹⁰ In isolation, stops data likely will not prove that officers have engaged in biased conduct, as other factors can also contribute to demographic disparities. However, early intervention systems are not designed to prove that officers are biased. Rather, they are meant to identify officers who might benefit from corrective action. In fact, the RIPA Board found that in 2019, officers from the 15 largest law enforcement departments in California were more likely to search Black and Latino individuals than white individuals but that these searches were less likely to reveal contraband than searches of white individuals.

The overall breadth of data that the departments use in their systems is mixed. None of the four local departments we reviewed uses stops data in their current early intervention systems. Stockton Police is particularly limited in its approach: the only indicator it uses to identify officers who may need intervention is the number of complaints it has received about them. Although complaints are a recommended component of early intervention systems, relying on them alone is a reactive and limited approach. By their nature, complaints occur after possible misconduct has already occurred, while the goal of an early intervention system is to intervene to support officers before they act or react poorly to the situations in which they find themselves. The lieutenant in Stockton Police's professional standards section—which oversees its early intervention system—stated that the system's use of complaints as its only indicator is archaic and that adding other data, such as data on uses of force, would be helpful. He generally attributed the current system's limitations to disjointed data systems and the absence of policies or procedures to include various data in the early intervention system.

Los Angeles Sheriff has also not incorporated stops data into its early intervention system, even though its special counsel has recommended that it do so since at least 2003. A recent effort to review these data at two of Los Angeles Sheriff's 23 patrol stations shows their potential value. Because of a settlement agreement with US DOJ, researchers analyzed Los Angeles Sheriff's stops data from the two service areas for the first half of 2019. The published results from September 2020 show that a few specific officers were outliers with respect to certain racial disparities: these officers, for instance, stopped Black individuals at higher rates than their peers did. In their report, the researchers stated that although legitimate explanations may exist for such findings, using stops data to identify officers who exhibit problematic trends and may require targeted intervention "is clearly an important managerial tool." Nevertheless, Los Angeles Sheriff has still not incorporated similar analyses into its departmentwide early intervention system. The former captain of its risk management bureau indicated that doing so could be beneficial but will take time, resources, and negotiation to implement.

The purpose of an early intervention system is not punitive, but rather to educate and guide officers to improve their performance. Accordingly, the nature of the interventions can vary widely and may include mental health services, training, or reassignment. However, neither Los Angeles Sheriff nor Stockton Police has outlined a range of specific intervention options or provided guidance to its staff about which interventions may work best to address particular behaviors. Stockton Police's policy lacks any requirement for supervisors to analyze or articulate the officers' specific behavioral issues when deciding on interventions. As a result, Stockton Police's documentation for its interventions mostly consists of one-page memorandums noting that a supervisor met with the identified officer to explain that the officer had been flagged by the early intervention system. Stockton Police's professional standards section lieutenant confirmed that these memorandums are general in nature and indicated that documenting an adverse comment about an officer outside of the formal disciplinary process could be problematic. Although we agree that early intervention systems should not be disciplinary in nature, the systems should still document specific trends in officers' conduct and include supports and education targeted to address those trends. In fact, Los Angeles Sheriff documented intervention plans, assigned mentors, and sometimes implemented targeted trainings for flagged officers. Particularly when taken as a whole, these steps make Los Angeles Sheriff's interventions more likely to achieve the desired behavioral changes.

Finally, if departments want to identify the effectiveness of their interventions, they must monitor officers' conduct following the interventions. Los Angeles Sheriff's process for doing so generally aligns with best practices. Once it begins providing interventions for an officer, it requires quarterly updates about that officer's progress that include narrative summaries and performance metrics. At the conclusion of the intervention and monitoring period, which can last up to two years, Los Angeles Sheriff requires a memorandum documenting that the officer has met requirements for removal from the program, such as achieving specific performance objectives from the intervention plan. Our review determined that Los Angeles Sheriff has generally implemented this framework in an adequate fashion. By contrast, Stockton Police has no formal process for monitoring the effectiveness of its interventions. Consequently, its professional standards section lieutenant indicated that unless Stockton Police placed an officer on a performance improvement plan, which is uncommon, it would not have any documentation regarding the outcomes of its interventions.

The Legislature could take action to ensure that local departments statewide are making use of stops data, given that doing so is an effective means of proactively identifying indications of bias. Departments must already collect stops data and report them to DOJ. The RIPA Board, which performs analyses of the stops data and issues public reports detailing the results, has recommended that departments annually review information about officers' individualized stops data and that departments identify officers with outlier trends regarding stops and searches, in conjunction with other performance metrics. However, it has not yet published technical guidance about how departments should incorporate stops data into an early intervention system. Such guidance could be useful given the nuances and complexities of using stops data to identify trends indicative of potential bias. The Legislature could require the RIPA Board to develop and publish that guidance and, as we describe in greater detail in Chapter 5, could require the RIPA Board to provide oversight of departments' implementation of early intervention systems that effectively incorporate analysis of stops data.

Please refer to the recommendations section to find the recommendations that we have made as a result of these audit findings.

---

Chapter 4

The Local Law Enforcement Departments Did Not Adequately Investigate and Address Possibly Biased Conduct

Key Points

• The four local law enforcement departments we reviewed sometimes failed to adequately investigate their officers' possibly biased conduct.
• Many of the investigations we reviewed were narrowly focused on blatant signs of bias, relied heavily on officers' denials of bias, or did not account for how officers' conduct reasonably appeared.
• Without a sufficient framework for consistently identifying, investigating, and tracking incidents of bias-related misconduct, law enforcement departments are less able to properly address such cases and the public cannot know their true prevalence.
• CDCR has historically used a decentralized complaint investigation process and has not fully implemented body-worn cameras, both of which have left it unable to adequately investigate and address biased conduct in its prisons.

The Four Local Departments Have Not Adequately Investigated Possibly Biased Conduct

Proper identification of and response to biased conduct demonstrates that, in practice as well as in principle, departments are committed to mitigating the effects of bias and building relationships of trust with the communities they serve. Each of the departments we reviewed has complaint and investigation processes that serve as its primary means for responding to instances of biased conduct. In accordance with state law, which requires departments to have processes for handling and reporting complaints by members of the public, each department has established ways that its employees or members of the public can submit complaints regarding misconduct, including biased conduct.

Each department's policies direct how it should respond to allegations of misconduct, which is often by having its internal affairs unit investigate the allegations and generally reach one of the four conclusions we list in the text box. State law requires departments to use the preponderance of evidence standard when determining whether to sustain an allegation, which means that departments must sustain an allegation if the evidence supports that it is more likely to be true than not true. ¹¹ This standard is commonly used in court proceedings about civil matters. It is a lower burden of proof when compared to other standards such as the clear and convincing evidence or beyond a reasonable doubt standards. Each department's policies indicate that sustained allegations typically lead to formal discipline or corrective action.

Further, each department can also investigate possible misconduct without receiving a complaint. For example, since January 2021 state law has required law enforcement departments to have a policy that provides a minimum standard on the use of force, including the factors that the department will use to evaluate and review all use‑of‑force incidents. Each of the four local departments we reviewed has procedures that direct it to perform investigations into officers' conduct after they use force against members of the public.

We reviewed a selection of misconduct investigations and found that in the investigations we selected, each of the four local departments failed to consistently conduct thorough and rigorous investigations of possible biased conduct by their officers. ¹² We discuss our review of CDCR's bias-related investigations later in this chapter. Among the 20 investigations we selected for review, there were only two—both at San José Police—in which a department reached a sustained determination specific to biased conduct, meaning that it determined that biased conduct had occurred. The investigations often exhibited multiple, interrelated problems. Figure 8 shows the key deficiencies we identified in the departments' investigations, each of which we discuss further in following text.

To determine how well the four local departments have responded to possible biased officer conduct, we assessed five investigations that each department conducted. Our selection process involved reviewing the allegations and basic facts of hundreds of investigations. Some did not include a fact pattern for which further investigation was likely to reveal biased conduct by the officers involved. For example, we came across investigations in which a complainant alleged that officers had engaged in biased conduct during a traffic stop or other detention. However, the fact pattern did not contain additional details to support the complainant's belief and there appeared to have been a violation of law that reasonably explained that specific stop or detention. Because we did not randomly select the 20 investigations we ultimately assessed, the frequency with which we found possible biased conduct or investigative deficiencies among these 20 investigations is not projectable to all of the investigations these departments conducted.

Figure 8

Because of Deficiencies in Their Investigations, the Local Law Enforcement Departments We Reviewed Failed to Recognize and Address Potentially Biased Conduct

In addition, one reason we did not select some of Los Angeles Sheriff's investigations for review was that it did not have body‑worn cameras in place at the time of the incidents. The lack of body-worn camera footage likely limited our ability to identify biased conduct at this department. For instance, in one case we did select for review, an officer mocked a Spanish-speaking individual using Spanish phrases and a fake accent, yet the video evidence showing that this conduct had occurred was available only because the officer chose to record the incident on a personal cell phone. Los Angeles Sheriff's inspector general reported that in October 2020 Los Angeles Sheriff began deploying body-worn cameras for its patrol officers, and a lieutenant in its body-worn camera unit stated in March 2022 that the department had deployed cameras for its patrol officers and was finishing implementation within other specialized units. The lieutenant indicated that Los Angeles Sheriff does not currently have a plan or policy for using body-worn cameras in its custody settings but may develop one based on a current pilot project of 50 body-worn cameras at its Men's Central Jail.

We concentrated our selection on investigations in which the fact pattern of the related incident presented reasonable questions—questions that an investigation of the incident would be able to resolve—about whether bias had influenced an officer's conduct. These include some of the incidents we describe in Chapter 1, in which officers exhibited bias toward members of the public. Many, but not all, of our selected investigations included complaints from members of the public.

Each Department Focused Heavily on Blatant Signs of Bias but Overlooked Other Indicators

The four departments' investigations often overlooked or dismissed more subtle indicators of possible bias, suggesting that the departments were focused on only the most direct signs of bias—such as the use of racial slurs or other overt demonstrations of bias. However, as we explain in the Introduction, the effects that bias can have on behavior vary widely and its presentation can, at times, be subtle. This is especially true for implicit biases. Officers who genuinely believe they are unbiased may still make statements or take actions that are influenced by their implicit biases. Therefore, it is important to look for more nuanced signs of the influence of bias in officers' statements and actions.

An investigation from one department that we reviewed typifies our concerns in this area. In this investigation, a member of the public submitted a complaint that an officer's social media posts had exhibited prejudices. We reviewed the posts that the department's investigation documented and found one that promoted a negative stereotype about Black parents and another that reposted an article that implied all Syrian refugees were terrorists. Although the posts did not contain racial slurs, they demonstrated that, at a minimum, the officer likely lacked awareness about common racial or ethnic stereotypes. They also raised questions about whether those stereotypes could affect how the officer might treat members of the demographic groups in question.

Nonetheless, the department's investigation did not analyze the stereotypes that the officer's posts promoted. Instead, the department concluded that it was "unable to find any racially derogatory remarks" in the posts and, as a result, concluded the allegation of prejudice was "clearly false." The department closed this investigation without providing the officer with any formal corrective action. When we raised these issues with the department, a lieutenant involved in monitoring complaint investigations did not provide a comment about the specific case and instead stated that it is difficult to determine whether an employee displayed implicit bias short of self-admission. In the next section we discuss the limitations of relying on officer admissions of bias, and later in this chapter we discuss a framework that could help departments' investigators more consistently recognize possible indications of bias.

The Departments Relied Heavily on Officers' Denials of Bias

Another issue with some of the investigations we reviewed was investigators' reliance on officers' after-the-fact statements about the intent behind their conduct. We identified this problem at Los Angeles Sheriff, San José Police, and Stockton Police. For example, in two investigations we reviewed, Stockton Police asked officers whether their behavior was motivated by bias but did not document any additional analysis about whether the officers had acted in a biased manner. We did not find this problem at San Bernardino Police because the records we reviewed included no evidence that the department had directly questioned the related officers about whether their conduct had been motivated by bias.

Relying heavily upon officers' denials of bias is a poor investigative practice. Officers who have acted out of explicit bias are unlikely to admit it if they believe such an admission would result in discipline. On the other hand, officers whose actions were influenced by implicit bias are likely unaware of this bias and thus also unlikely to admit to it. The text box provides an example of the limited usefulness of an officer's after‑the‑fact denial that bias played a role in his actions.

An investigation that Stockton Police conducted of the encounter that we depict in Figure 9 is another example of the deficiencies of this approach. Much of the incident involved two officers arguing with a Black man about whether they had detained and searched him because he was Black. After the incident, the man filed a complaint alleging that the officers had engaged in racial profiling. Nonetheless, Stockton Police did not perform an analysis of whether bias influenced the officers' actions. Instead, it simply asked the officers whether race had been a factor in their actions, and then assessed whether their conduct had violated other policies, such as its policy requiring professionalism.

WARNING: This figure contains content that readers may find disturbing.

Figure 9

An Example of a Complaint Investigation That Failed to Thoroughly and Objectively Evaluate Whether Bias Was Present

When we shared our concerns with Stockton Police, the lieutenant in charge of its professional standards section did not disagree with our conclusions about the lack of analysis in this investigation. However, when we described the indications of bias or deficient cultural awareness, some of which we illustrate in Figure 9, the lieutenant again cited the officers' denials of bias. He further stated that their conversation with the man had not been rude or discourteous, and that the officers had the right to defend themselves in a professional and appropriate manner. We disagree with the notion that the officers behaved professionally and appropriately during that conversation or at other points in the encounter.

The Departments Did Not Consider the Reasonable Appearance of Officers' Conduct

Because biases are not always demonstrated in clear and simple ways and are not always consciously held, one approach to determining whether they have influenced an officer's behavior is considering how a reasonable third party would perceive that behavior. This approach allows a department to evaluate the presence of bias in a way that incorporates the likely perspective of the community a department serves. It is already a best practice for identifying and mitigating bias related to hate crimes, workplace discrimination, and legal proceedings. In fact, the departments we reviewed have even standardized this approach for assessing other possible policy violations, such as unbecoming conduct or workplace sexual harassment. Nevertheless, in several of the cases we reviewed, the local departments failed to reach conclusions that incorporated the reasonable appearance of officers' conduct. Instead, their conclusions reflected much narrower thresholds for determining whether the officers' behavior was problematic.

Figure 10 illustrates an example of how narrow considerations affect the quality of an investigation. In this investigation of an encounter between two officers and a member of the public, the department in question concluded that the officers did not say or do anything "a reasonable person could believe" would indicate that the complainant's race or ethnicity played a role in the encounter. However, the department's analysis was flawed. As the figure shows, one officer in particular made statements to the complainant that could reasonably be understood to be connected to his ethnicity, such as asking for the man's "papers" and mocking his complaints about the encounter as a "Telemundo special." However, during its investigation, the department took the officers' explanations for these statements at face value and used flawed reasoning to conclude that the conduct had not been influenced by bias. For example, the department acknowledged that the phrase "papers" has a context specific to immigration, but because this encounter was not related to immigration, it dismissed the possibility that the officer—even unconsciously—might have used that specific term because of the complainant's ethnicity.

The department similarly concluded that the officer could not have been influenced by bias when he made the "Telemundo novela" and related comments because he was of Mexican descent, had grown up with his family watching novelas on television, and stated that he did not make the comments "with malice or in a demeaning manner." However, that logic focuses on intent and does not consider how the comments reasonably appeared in the context in which they were actually made—a context in which the complainant did not know the officer's personal habits and in which the officer was mocking the complainant. Further, as we note in the Introduction, people can possess implicit biases about others with whom they share key identity characteristics. When we asked the department for its perspective on this case, the lieutenant in charge of the unit responsible for investigating complaints did not wish to comment because he was not in that unit at the time and did not think it appropriate to comment on other investigators' analyses.

Finally, in two of the investigations we reviewed, the departments failed to hold officers accountable to basic expectations of professionalism and courtesy toward the public, casting doubt on the departments' objectivity in those investigations. In one of these investigations, an officer encountered a man experiencing homelessness and immediately drew his gun, began speaking in a raised voice and called him a "ding‑dong." Very soon thereafter, the officer proceeded to escalate the situation by using force. As part of its review of the use of force during the incident, the department acknowledged that the officer had used "street terms" that "could be construed as condescending." However, the investigator did not believe the officer had intended to be discourteous or disrespectful and stated that the officer "was merely in a very stressful situation." The investigator wrote that "although [the officer's] verbiage was certainly not textbook," it did not rise to the level of a policy violation.

A lieutenant who later reviewed the investigation's conclusions noted that the officer had used demeaning language that escalated the incident. However, the department declined to counsel the officer or apply corrective action because it did not complete its secondary review until more than a year after the initial incident—in part because the original investigator did not respond in a timely manner to reviewer notes directing the investigator to incorporate an assessment of the officer's use of proper tactics and communication. According to a sergeant in the unit responsible for investigating complaints, the officer's behavior in this incident appeared aggressive and demeaning, and his use of force was questionable. The sergeant stated that the review should have led to a formal internal investigation into why the officer escalated force and whether he had exhibited any possible biases.

WARNING: This figure contains content that readers may find disturbing.

Figure 10

Deficiencies in a Department's Complaint Investigation Caused It to Overlook Potentially Biased Conduct

Three Departments Prematurely Dismissed Complaints Alleging Biased Conduct

Los Angeles Sheriff, San Bernardino Police, and Stockton Police sometimes conduct only cursory reviews of complaints from the public. For example, San Bernardino Police allows investigators to classify some reports about officer conduct as service complaints if it determines the reports, even if true, would not represent a violation of law or department policy. The sergeant in charge of San Bernardino Police's professional standards bureau explained that service complaints generally focus more on the department's overall service than on possible misconduct by a particular officer. He further explained that because service complaints are not likely to involve misconduct on the part of officers, their reviews tend to require less documentation. The department classified two complaints we reviewed—both of which contained possible indications of racial bias—as service complaints.

Los Angeles Sheriff, San Bernardino Police, and Stockton Police each prematurely dismissed at least one complaint alleging biased conduct that we reviewed without addressing concerning elements of the cases. In one of the cases, an officer made several questionable statements to a Black man, including accusing the man of playing the "race card." The officer further stated that he wished "we lived in a world, back in the [19]60s and [19]70s, where we could feel comfortable"—a statement that overlooks the negative experiences of many Black Americans during that era. After the man suggested that a neighbor may have used a racial slur toward him, the officer indicated that doing so was protected as free speech. The officer then proceeded to list aloud other derogatory slurs to illustrate his point that distasteful speech was protected.

Although the incident contained multiple indications of possible biased conduct or deficient cultural awareness on the part of the officer, the investigative report consisted of a one-page summary indicating that the officer had behaved appropriately. This summary included a brief note that the investigator had told the officer he "did an excellent job" but could have used "a different remark other than 'race card,'" because the complainant "could only focus on [that] remark." Although the man had submitted a formal complaint alleging that the officer had engaged in biased conduct, the department marked this case as resolved without conducting any formal analysis or drawing any formal conclusions related to bias. A sergeant within the department's unit that investigates complaints agreed that it would have been beneficial to more formally investigate this complaint. He indicated that one reason such an investigation did not occur was because of a lack of guidelines about what constitutes biased conduct—an issue we address in the following section.

Each local department we reviewed has a policy for conducting lower‑level reviews of some allegations of misconduct, such as reviews by an officer's chain of command rather than by internal affairs—and we recognize that this may be an efficient way to handle some complaints. However, if a department uses lower‑level reviews to address complaints related to bias, those reviews should thoroughly analyze whether biased conduct has occurred and adequately address the concerns of the complaint. Los Angeles Sheriff, San Bernardino Police, and Stockton Police did not always do so in the cases we reviewed. Stockton Police disagreed with our assessment that it should have more fully investigated the complaint that we reviewed of which it had performed only a lower‑level review. However, San Bernardino Police generally acknowledged that the investigations we reviewed were deficient, and Los Angeles Sheriff provided us a revised complaint routing form that would require a full internal affairs bureau investigation for any allegations of bias or profiling.

The Four Local Departments Have Not Adopted a Framework for Identifying Bias During Investigations

When examined as a whole, the deficiencies we describe in the four local departments' investigations of bias demonstrate that these departments lack a sufficient framework for investigating such issues. Sufficient guidance for conducting investigations would necessarily include a uniform definition for what constitutes biased conduct by officers and factors that investigators should weigh in making their assessments. Nonetheless, the departments that we reviewed lacked such guidance. In fact, San José Police was the only department to consistently reference a bias-related policy in its investigations of misconduct. Its policy defines bias-based policing in a way that is similar to state law's definition of prohibited racial or identity profiling, which the text box shows.

Los Angeles Sheriff and San Bernardino Police also cite this definition in their bias-related policies; however, unlike San José Police, they conducted investigations of possible biased conduct without analyzing whether the conduct met this definition. One reason for this difference could be that San José Police's investigation policies explicitly list "bias-based policing" as one of a handful of categories that internal affairs investigations could consider when analyzing officers' conduct. As we explain in Chapter 1, Stockton Police does not have a policy specifically regarding biased conduct.

A statewide definition of biased conduct and a requirement that law enforcement departments apply it during their investigations would clarify expectations regarding what constitutes biased conduct. The definition of racial or identity profiling in state law provides a useful starting point for such a definition. However, the definition is limited to specific police activity and does not address other conduct, such as officers expressing biased or prejudiced viewpoints on social media. In addition, neither that statute nor the statute mandating that departments have complaint investigation procedures specifies how departments should approach their investigations of possible biased conduct. For example, these statutes do not require departments to initiate investigations and provide bias-specific analysis whenever they notice possible indications of bias.

Figure 11 shows an example of how having such established expectations might have altered one of San Bernardino Police's routine use-of-force reviews. In this incident, multiple statements and actions indicated that the department should have considered whether bias affected the officers' responses to the situation. However, the department did not do so. A requirement to investigate for bias wherever indications appear—and a uniform standard for what constitutes biased behavior—might have prompted the department to undertake a more thorough review. A sergeant in the department's professional standards bureau indicated that other explanations, aside from possible biases, existed for the officers' behavior; however, the sergeant acknowledged that given the circumstances of the incident, it would have been reasonable to formally assess whether the officers' conduct was indicative of possible biases—and that guidelines for when to perform these assessments would help.

Each department also lacks adequate criteria for how to determine whether bias influenced officers' conduct, directly affecting its ability to assess implicit bias and the reasonable appearance of that behavior. The text box outlines criteria that investigators could use to more fully and consistently consider the possible influence of bias on officer conduct. In the absence of such guidance, we observed that at San José Police—the only department we reviewed that consistently provided formal analysis in its investigations about whether conduct was biased—some investigative analyses and conclusions differed widely from one another for no discernible reasons. In three of the investigations we selected for review, San José Police recognized how officers' conduct reasonably appeared rather than deferring completely to the officers' explanations of their intent, and it mentioned implicit bias as a possibly relevant factor in one of those cases. In contrast, for the remaining two investigations we reviewed, San José Police's analyses were flawed, largely because it did not adequately consider the possible influence of implicit bias or how the officers' behavior reasonably appeared.

WARNING: This figure contains content that readers may find disturbing.

Figure 11

San Bernardino Police Failed to Consider the Possible Influence of Bias Despite Indicators It May Have Been Present

In the absence of strong guidance about how to investigate possible biased conduct, San José Police and the other departments we reviewed risk treating both complainants and officers differently from one investigation to the next. The lieutenant in charge of San José Police's internal affairs unit during our audit indicated that facts and evidence differ from case to case but that guidelines in this area could be helpful for investigators.

Finally, the four departments would likely benefit from training their investigators on how to identify biased conduct. The approaches the departments sometimes took in the investigations we reviewed suggest that investigators did not fully understand bias or its nuances. In addition, a few of the investigations considered criteria that were irrelevant to the question of whether conduct was biased, such as the fact that officers maintained a calm demeanor. For instance, in the case we describe in Figure 10 the department noted as a relevant factor in its investigation that an officer had made comments "in a casual manner."

These shortcomings demonstrate that training about effective investigations of biased conduct would be beneficial. At the time of our audit, none of the departments indicated that they had implemented training for their investigators specifically related to identifying and analyzing biased conduct. Further, POST was not aware of any statewide training specific to performing investigations of possible biased conduct. Training on investigations of possible biased conduct could supplement the broader training requirements related to bias that we describe in Chapter 3.

Three of the Local Law Enforcement Departments Have Poorly Tracked Their Bias Investigations, Obscuring the Extent of the Problem and Reducing Public Transparency

Poor investigation practices of Los Angeles Sheriff, San Bernardino Police, and Stockton Police not only have impaired their identification of individual instances of biased conduct but have also hindered their ability to monitor the prevalence of biased conduct by their officers. Accurate data on the frequency with which members of the public allege that officers have exhibited bias, as well as how often investigations sustain such allegations, can assist a department in evaluating the extent to which bias is influencing its officers' conduct and damaging its relationships with the public. However, these three departments' failures to formally determine whether biased conduct has occurred and to properly label their investigations has left them unable to effectively track such incidents.

In the investigations we reviewed, Los Angeles Sheriff, San Bernardino Police, and Stockton Police often did not reach a formal conclusion about bias even when the facts in a case indicated that such a conclusion would be reasonable. When a department performs a misconduct investigation, it generally identifies one or more allegations of misconduct and relates those allegations to specific policies that the officer may have violated. As we previously discuss, San José Police's investigators explicitly assess whether biased conduct has occurred when performing investigations, likely in part because its policies list "bias-based policing" as one of the categories that internal affairs investigations should consider. In contrast, the remaining three departments often limited their evaluations to whether officers violated policies related to other issues, such as professionalism or courtesy toward members of the public. For example, Stockton Police mentioned in one investigation that an officer's social media post had "promoted negative racial/ethnic stereotypes." However, its final determinations and discipline were related to its social media policy and its policy against "unbecoming conduct," obscuring the fact that the conduct appeared biased.

Each department has a system for electronically tracking information related to its misconduct investigations, including the specific allegations and their conclusions. However, because Los Angeles Sheriff, San Bernardino Police, and Stockton Police do not consistently make formal determinations of whether bias occurred, they cannot effectively track or quantify the extent of biased conduct. In contrast, San José Police can more effectively track the frequency with which it has received complaints about bias and how often it has substantiated these allegations.

Beyond the internal problems that poor tracking creates for a department, it also has external consequences. State law requires departments to report information that DOJ requests, and to fulfill its own responsibilities under state law DOJ has requested departments to report complaints they receive from the public, including those specifically alleging racial or identity profiling. DOJ collects and publishes—and the RIPA Board analyzes—this information, which provides a view of the extent of possible biased conduct by officers throughout the State. However, during our review it came to our attention that two departments we reviewed reported inaccurate information to DOJ.

Specifically, lieutenants in Stockton Police's professional standards section stated that the department does not report all complaints to DOJ. They indicated that Stockton Police does not report complaints it designates for lower levels of review—such as complaints for which a supervisor has determined that an officer acted reasonably and within department policy—because these complaints do not result in a formal finding. This practice is problematic because it results in Stockton Police underreporting its complaints. A log that Stockton Police provided us indicates that in 2020 it recorded more than 35 lower-level complaints, including more than 10 in which it noted that profiling was alleged. However, Stockton Police reported only three complaints alleging profiling to DOJ for that year.

Los Angeles Sheriff also appears to have reported inaccurate information. DOJ data show that Los Angeles Sheriff reported it had substantiated three bias-related allegations from 2016 through 2020. However, when we reviewed the cases that Los Angeles Sheriff advised us it had reported, we found that none of the investigations clearly reached sustained findings for biased conduct. In fact, two stated explicitly that biased conduct had not occurred. A sergeant in Los Angeles Sheriff's professional standards division told us that to report information to DOJ, an employee reassesses the details of each case and makes judgments about which bias complaints should be considered to have been substantiated. The sergeant indicated that Los Angeles Sheriff is in the process of making changes to better track its categorization of public complaints.

Because Los Angeles Sheriff, San Bernardino Police, and Stockton Police do not reliably make formal findings specific to bias or track the investigations that include allegations of bias, they are at increased risk of failing to disclose records made public under recent changes to state law. Recently, the Legislature and the Governor adopted statutory changes effective January 2022 that require departments to disclose certain records related to investigations of biased conduct that result in sustained findings. However, if departments do not properly make sustained findings for or track investigations of biased conduct, they may not be able to adhere to the intent of this law.

If implemented, our recommendations for establishing a uniform definition of biased conduct, adopting guidelines for conducting investigations of possible biased conduct, and providing training for investigators on how to assess possible biased conduct will likely curb problems with tracking and public transparency. In Chapter 5 we discuss additional recommendations related to independent oversight of departments' investigations that could also improve the consistency of departments' investigations and tracking efforts.

CDCR Is Beginning to Address the Significant Weaknesses in Its Misconduct Investigation Process

CDCR has taken a decentralized approach to investigating misconduct in its prison system, hindering its ability to assess the extent to which its officers have engaged in biased conduct. For instance, CDCR has allowed authorities at each prison the discretion to investigate incarcerated individuals' complaints of staff misconduct or to elevate those complaints for review by CDCR's office of internal affairs. As a result, authorities at each prison have been able to conduct inquiries into complaints from incarcerated individuals, and unless they expected the misconduct to result in adverse action, they could resolve the complaints themselves without elevating them. CDCR also has not centrally tracked the details of the complaints that prison authorities handled, instead allowing each prison to do so on a standardized complaint log. Because the handling and tracking of complaints has occurred in part at the prison level, CDCR has had limited oversight of how its prisons have dealt with allegations of biased conduct.

Because CDCR's office of internal affairs has tracked only the complaints that prison authorities chose to elevate to it, we were limited in our selection of biased conduct investigations to review. A number of the investigations we considered—and four of the five we ultimately selected—involved allegations of bias or discrimination that CDCR employees made about other CDCR employees, sometimes involving alleged misconduct toward incarcerated individuals. In each case we selected, CDCR concluded that the officers had engaged in conduct related to bias—such as making insults or engaging in harassment based on identity characteristics—and imposed discipline. Our review of these investigations did not identify any instances in which CDCR overlooked or failed to investigate possible biased conduct.

However, the decentralized nature of CDCR's complaint process means there is a significant risk that complaints were mishandled elsewhere in its system of 34 adult prisons. Indeed, external oversight of CDCR has revealed key weaknesses in prison authorities' handling of complaints made by incarcerated individuals. For instance, the Office of the Inspector General—which state law establishes as an independent entity responsible for oversight of CDCR's misconduct investigations—reported in 2019 that over half of the 188 investigations it reviewed from the Salinas Valley State Prison were inadequate and that 92 percent contained at least one significant investigative deficiency, such as poor interviewing techniques or a lack of objectivity. The Office of the Inspector General noted that the complaint investigation process at Salinas Valley State Prison was also in place at other prisons statewide, and it recommended that CDCR consider "a complete overhaul" of this process. Similarly, plaintiffs in an ongoing lawsuit—which began in 1994 and which alleges discrimination against incarcerated individuals with disabilities—have raised concerns about CDCR's process for investigating complaints of misconduct.

In response to these concerns, CDCR began reforming its misconduct investigation process. It filed emergency regulations effective January 2022 outlining a new process that, when fully implemented, will require a single centralized screening team to review every complaint alleging staff misconduct toward an incarcerated individual or parolee and to make determinations about the level of review the complaints require. When this screening team refers investigations to prison authorities to conduct, CDCR's internal affairs unit will review the investigations for completeness and independence. These changes—which appear to align with past recommendations from the Office of the Inspector General—should resolve our primary concerns regarding CDCR's lack of oversight of complaints of misconduct toward incarcerated individuals.

Under CDCR's new process, prisons must refer to the office of internal affairs allegations of officer misconduct toward incarcerated individuals that include identity-based insults, identity-based harassment, or discrimination. However, that referral process will depend on CDCR's screening team and prison-level investigators to recognize conduct that falls into these categories. To consistently identify and respond to instances of possible biased conduct by officers, CDCR's new investigation process would benefit from including the elements we discuss earlier in this chapter as helpful to local departments: using a clear and comprehensive definition of biased conduct; employing specific criteria for determining when conduct meets that definition, and documenting formal analysis based on these criteria; and training its investigators on how to analyze situations for possible bias. CDCR currently lacks formalized criteria for how to investigate potential indications of bias. Further, according to the chief of administrative operations in CDCR's office of internal affairs, all CDCR staff receive relevant training, such as implicit bias training; however, CDCR does not provide training that is specific to investigators.

Finally, a significant hindrance to CDCR's investigations of biased officer conduct is that CDCR has not yet fully implemented body‑worn cameras. That limitation has significantly impeded its ability to substantiate indications of bias in certain cases involving complaints from incarcerated individuals. For example, because of conflicting witness interviews, in one investigation, CDCR was unable to substantiate complaints alleging that officers used racial slurs or other offensive language. It likely could have substantiated or refuted these complaints with video evidence. According to one CDCR investigation, multiple incarcerated individuals alleged that an officer had used racial slurs, but the officer denied doing so and other CDCR staff did not observe that he had. CDCR found no additional evidence to substantiate the allegations, which it did not sustain.

CDCR is in the process of implementing additional video capabilities. In response to court orders resulting from the lawsuit we mention earlier, CDCR has required officers at six of its 34 adult prisons to use body-worn cameras, with four more prisons set to receive body-worn cameras if the Legislature fully funds CDCR's budget change proposal for fiscal year 2022–23. In addition, CDCR has already installed or has plans to install an audio-video surveillance system (fixed cameras) at 22 adult prisons, some of which also have body-worn cameras. Therefore, even if the Legislature fully funds CDCR's request for additional body‑worn cameras and fixed cameras—which includes $80.3 million in one‑time funding and $7.6 million in ongoing funding—24 adult prisons would still lack body-worn cameras and 12 adult prisons would lack both body-worn cameras and fixed cameras. These institutions will be less able to substantiate or refute that their officers have engaged in biased conduct.

According to an associate warden in CDCR's division of adult institutions, CDCR has not determined exact cost estimates for expanding the use of body-worn cameras and fixed cameras, but the 12 adult prisons that lack fixed cameras are in the beginning stages of completing the necessary designs. The associate warden indicated that body-worn cameras generally cost less to deploy than fixed cameras. According to the chief of administrative operations in CDCR's office of internal affairs, fixed cameras and body-worn cameras offer different advantages, and body-worn cameras are generally more useful for recording individuals' statements and for documenting officers' perspectives. These benefits of body‑worn cameras make them particularly useful for substantiating or refuting allegations that officers have made biased comments toward incarcerated individuals. CDCR's associate warden estimated that expanding the use of body-worn cameras to the remaining 24 adult prisons could cost about $1 million to $2 million per prison initially, along with ongoing annual operating expenses.

The Five Departments Do Not Have Processes to Consistently Impose Discipline Targeted at Correcting Their Officers' Biased Conduct

In some circumstances involving biased conduct, strong punitive discipline—such as termination—is warranted. For example, a department may determine that an officer's conduct demonstrates such strong bias that it renders that officer incapable of performing his or her law enforcement duties fairly. But in other circumstances, such as when an officer's conduct is influenced by implicit biases, a more beneficial approach may be corrective action—such as training—that is meant to change the officer's behavior. According to US DOJ and state civil service guidelines, the goal of discipline should be not to punish officers, but to correct their behavior—and creative alternatives to traditional punitive discipline may be most effective in doing so.

However, in 12 cases we reviewed in which departments sustained allegations of misconduct but did not terminate the officer's employment, four of the five departments relied upon punitive discipline alone to address the misconduct rather than using more targeted corrective actions. For example, after CDCR found that an officer had used a racial slur toward an incarcerated individual, it imposed a yearlong salary reduction without any required training or education. The officer indicated he had used the slur in response to an incarcerated individual using it first, which had bothered the officer. Corrective action may have helped the officer understand why repeating the slur was problematic. Similarly, after Stockton Police found an officer had made a social media post that "promoted negative racial/ethnic stereotypes," it only issued a written letter of reprimand to the officer that did not mention bias or stereotypes or explain why the officer's post was problematic. Although these types of disciplinary actions communicate to officers that their behavior was unacceptable, punitive discipline will not, on its own, educate officers about why their conduct was problematic or how to avoid that conduct in the future.

The departments have also sometimes implemented targeted corrective actions. San Bernardino Police, San José Police, Stockton Police, and CDCR occasionally took these corrective actions in some of the cases we reviewed. Although Los Angeles Sheriff did not in the cases we reviewed, it has an education‑based discipline option that generally allows officers to choose to receive less punitive discipline in exchange for completing other corrective actions. These corrective actions can include attending specific trainings, completing community service activities, or writing research papers on topics related to their misconduct. The department's guidelines include options and instructions for responding to different types of misconduct, such as alcohol-related offenses or inappropriate communication, and its lists of suggested courses include a course on cultural awareness for supervisors. However, the guidelines do not specify how or when to apply this framework to instances of biased conduct toward members of the public. A lieutenant in Los Angeles Sheriff's advocacy unit indicated that the guidelines have not been updated since the department adopted its bias-related policy in 2021.

None of the departments' policies require the consistent use of such corrective actions to address possible biased behavior—nor do the policies specify which corrective actions may be most effective for this purpose, such as specific training courses or specific approaches to take when counseling officers. In general, individuals involved in investigations at most of the departments shared with us that disciplinary decisions depend on the specific circumstances of each case, or are made by other staff or leadership outside of internal affairs. Nevertheless, requiring consistent use of targeted corrective actions would help departments address biased conduct.

To provide consistent and appropriate discipline, US DOJ suggests using a discipline matrix, which establishes in advance the range of allowable discipline outcomes based on the type of policy violation. CDCR has a formal discipline matrix and San José Police has a table of discipline standards that similarly lists policy violations along with ranges of discipline to be imposed, and both departments have portions dedicated to biased conduct. However, these portions rely entirely or heavily on punitive discipline, meaning that pairing punitive discipline with more targeted corrective action, like training or counseling, is optional. Los Angeles Sheriff also has a discipline matrix, but it does not have a section specifying discipline options for when officers engage in biased conduct toward the public. Neither San Bernardino Police nor Stockton Police uses a formal discipline matrix, although they indicated that they have standardized disciplinary outcomes for certain types of policy violations. To address our concerns, all five departments should ensure that their discipline matrices or broader discipline policies require consideration of nonpunitive corrective actions whenever such actions might effectively help address possible biased behavior, even in conjunction with punitive discipline.

Please refer to the recommendations section to find the recommendations that we have made as a result of these audit findings.

---

Chapter 5

Ensuring Broader Use of Practices That Address Bias Will Require State-Level Intervention

Key Points

• Local law enforcement departments—including some outside the scope of this audit—have not proactively adopted best practices to address the threat that bias poses to fair and impartial law enforcement.
• External reviews and oversight are effective measures for increasing departments' adoption of these sorts of best practices.
• We recommend three state-level interventions to increase law enforcement departments' transparency and accountability for addressing bias: requiring departments to report additional information to the RIPA Board about their policies and practices, formalizing DOJ's existing reviews of departments, and expanding local and statewide oversight of departments' misconduct investigations.

External Pressure Has Prompted Departments to Implement Bias-Related Reforms

As our report makes clear, the law enforcement departments we reviewed have not addressed bias in their organizations in a comprehensive manner. Despite taking beneficial steps in certain areas, each of these departments has deficiencies in its policies and practices that limit their abilities to mitigate the effects of bias on the manner with which officers perform their duties. Unfortunately, these deficiencies are not limited to the departments we reviewed. DOJ has identified the need for several other local law enforcement departments to improve in some of the same areas as we did during our audit: hiring and recruitment, training, and complaint investigation. For example, amid mounting public pressure for reform and after an officer-involved shooting death of an unarmed Black man, the Sacramento Police Department (Sacramento Police) asked DOJ to perform a review of certain aspects of its operations. DOJ found that Sacramento Police needed to develop and implement a community outreach plan, review and update its bias training, and improve its hiring and recruitment processes.

As the Sacramento Police example shows, local law enforcement departments are more likely to embrace changes to their approach to combating bias when significant external forces highlight the need for these reforms. In fact, some of the departments we reviewed implemented some of their positive practices because of lawsuits or external reviews. For example, Los Angeles Sheriff instituted its early intervention system following an external review by a special counsel that the county of Los Angeles Board of Supervisors appointed to investigate an increase in officer‑involved shootings. Similarly, the city of San José's Office of the Independent Police Auditor (San José Police Auditor)—an entity we describe in more detail later in this chapter—has made several recommendations for improvement to San José Police's practices, which we saw the department later adopting as reforms.

The level of harm that bias within law enforcement departments can cause to officers and to communities indicates that more should be done to address this issue. As we describe in the Introduction, officers are expected to provide essential services and protection, and to do so, they are empowered in unique ways that other public servants are not. Given the important position that officers hold, the repeated documented failures of law enforcement departments to sufficiently address bias, and the influence of external oversight on prompting change, we propose three measures California can take at the statewide level to address bias in policing. Figure 12 depicts these recommendations.

Figure 12

Three State-Level Interventions Would Increase Law Enforcement Departments' Accountability for Addressing Implicit and Explicit Bias

The RIPA Board Could Increase Departments' Public Accountability for Addressing Bias

To promote greater transparency and accountability around local law enforcement efforts to address bias, the Legislature could require departments to regularly report whether they are adhering to a standardized set of best practices. In performing our audit, we found a large volume of information about the best approaches that law enforcement departments can take to address bias. However, we found little information about how widespread the use of these practices is throughout the State. Increased transparency could incentivize local departments to adopt policies or practices that they would not have otherwise. Additionally, more visibility into local department practices would provide the public and policymakers with information that would assist them in holding local departments accountable.

The RIPA Board is positioned well for maintaining and verifying the information that the local departments would report. The Legislature required DOJ to create the RIPA Board for the purpose of eliminating racial and identity profiling, and for improving diversity and racial and identity sensitivity in law enforcement. State law already requires the RIPA Board to engage in activities such as analyzing policies and practices to make policy recommendations for eliminating racial and identity profiling. The RIPA Board has also recently reviewed and analyzed indicators of bias in stops data and civilian complaint data. Therefore, a statewide effort to catalogue and verify the steps departments have taken to address bias would be consistent with the RIPA Board's existing mission and activities.

For a statewide reporting effort to work optimally, the RIPA Board would need to establish a uniform expectation of the policies and practices that departments should use to combat bias. This guidance would help departments ensure that they are being comprehensive in their efforts. For example, in Chapter 3 we discuss how law enforcement departments can design effective early intervention systems to detect potentially problematic trends in officer conduct. However, even Los Angeles Sheriff—which has the most robust early intervention system among the local departments we reviewed—has not adopted key elements that make such a system work optimally. In addition, to ensure that the departments' responses to the RIPA Board are accurate, the Legislature could require that departments provide the board with copies of any of the policies, procedures, or plans that they attest align with the best practices. By establishing the expected components of a department's approach to combating bias, verifying a selection of department practices, and publicly reporting on each department's adherence to those best practices, the RIPA Board would add a layer of accountability and transparency that is missing from the State's current approach to addressing bias in law enforcement.

Expanding DOJ's Reviews Could Improve Departments' Implementation of Best Practices

DOJ could more regularly perform in-depth reviews of individual law enforcement departments and recommend improvements. Currently, DOJ's Civil Rights Enforcement Section is responsible for addressing a broad array of civil rights violations including hate crimes, failure to provide disability access, and police misconduct. This section conducts reviews of many areas of law enforcement departments' operations, including hiring, training, community engagement, early intervention systems, complaint investigations, and discipline. DOJ uses pattern or practice investigations to review agencies for civil pattern or practice violations where there is evidence to suggest systemic police misconduct. Similarly, through a mutual agreement, DOJ conducts reviews of local departments' adoption of best practices. DOJ has also conducted more narrow reviews related to law enforcement, such as a review it began in 2020 of Los Angeles Police Department's use of the CalGang system.

DOJ's reviews generally result in stipulated agreements or public reports that require departments to improve their operations. For example, DOJ entered a 2020 agreement with Kern County Sheriff's Office that required that department to develop a written plan that included strategies for recruiting a quality workforce and for providing its officers annual training on community engagement techniques. The agreement stipulated that the training must cover establishing formal partnerships and actively engaging community organizations such as those representing its youth, immigrant, and LGBTQ communities. After entering the settlement, DOJ and Kern County selected an independent monitor to oversee the department's implementation of the reforms in the agreement.

Despite the benefits that these reviews offer, DOJ has performed them infrequently. From 1999 until March 2022, it had completed only six—listed in Table 7—of city or county departments that examined their broader policing practices, with another four ongoing. DOJ's reviews have primarily been pattern or practice investigations, which depend on DOJ identifying reasonable cause to believe that a local department has a pattern or practice of violating civil rights. To conduct its other reviews, DOJ relies on mutual agreement with the local department, making those reviews less frequent than its pattern or practice reviews. DOJ will soon have additional authority under state law to review certain police practices, but that authority will still depend on it being invited to perform those reviews. Specifically, beginning in July 2023—subject to appropriated funding—state law will require DOJ to operate a police practices division that reviews law enforcement deadly force policies when a law enforcement agency requests such a review.

Table 7

DOJ Has Reviewed Police Practices Since 1999

DATE	Department	Review Type
2001	Riverside Police Department	Pattern or Practice Investigation
2009	Maywood Police Department	Pattern or Practice Investigation
2019	Sacramento Police Department (Phase 1)	Mutual Agreement
2020	Sacramento Police Department (Phase 2)	Mutual Agreement
2020	Kern County Sheriff's Office	Pattern or Practice Investigation
2021	Bakersfield Police Department	Pattern or Practice Investigation

Source: DOJ reviews of law enforcement departments and resulting judgments and reports.

Given the value that these reviews provide to both law enforcement departments and the public, we believe that DOJ should complete them on a more regular basis. Additionally, to ensure that DOJ performs these reviews for the departments that will benefit most, it should be required to develop selection criteria. For example, when selecting which departments to review, DOJ could consider the information departments report to the RIPA Board, the frequency and type of their officer misconduct allegations, and other factors.

An Independent Review Process Could Ensure That Law Enforcement Departments Adequately Investigate Their Officers' Misconduct

Among the four local law enforcement departments we reviewed, only San José Police and Los Angeles Sheriff have processes for independent review of the quality of their misconduct investigations. According to the city of San José's website, its city council established the San José Police Auditor in 1993. ¹³ Voters then added the San José Police Auditor to the city charter in 1996, and in 2020 voted to expand its authority and access to records, such as by allowing it to review misconduct investigations initiated by the department without a public complaint. The text box shows its key authorities and responsibilities. Similarly, Los Angeles Sheriff has an Office of the Inspector General and constitutional policing advisors, both of which can access records and review investigations. However, because the San José Police Auditor was the only entity whose review was clearly noted in and had impact upon the specific case files we selected, we focus much of this section on that entity.

The San José Police Auditor reports to the city council. It reviews a significant percentage of San José Police's misconduct investigations to determine whether they are complete, thorough, objective, and fair. Its feedback resulted in meaningful improvement to a San José Police investigation of a complaint of biased conduct. In that case the San José Police Auditor raised concerns that San José Police had not adequately examined whether statements that an officer made during an interaction with a member of the public had shown bias. In response to this feedback, San José Police reopened the investigation and reexamined the evidence in the case.

Following this reexamination, San José Police changed its conclusion about the bias allegation against that officer from unfounded—which indicates the allegation was clearly false—to not sustained, meaning that San José Police could not clearly prove or disprove the truth of the allegation. In that same case, the San José Police Auditor's feedback also caused San José Police to add and analyze a new allegation against the officer related to a possible improper search and seizure. For that new allegation, San José Police reached a sustained finding and indicated it took corrective action in the form of training and counseling. These significant changes were a direct result of the San José Police Auditor's input.

We identified some weaknesses in the San José Police Auditor's authority and approach. Most importantly, San José Police sometimes disagreed with and declined to implement key feedback from the San José Police Auditor. Specifically, for two cases we reviewed, San José Police determined that allegations of biased conduct were unfounded despite indications of bias, and despite the San José Police Auditor raising concerns with the internal affairs unit about its conclusions. In such situations, the San José Police Auditor has the authority to bring its concerns to the chief of police and the city manager. However, it did not do so in these cases.

The San José Police Auditor stated that the primary reason it has not elevated its concerns more frequently is that it sometimes has had minimal time to review investigations. State law generally requires law enforcement departments to resolve investigations of officer misconduct within one year of the date they discover the misconduct if they wish to take disciplinary action. It does not include any exceptions that explicitly grant specified time for independent oversight entities to review those cases. ¹⁴ Exceptions to the one-year time period in state law include, among others, a reasonable extension for investigations that involve more than one employee and suspending the time period when investigations involve a matter in civil or criminal litigation. In one case in which the San José Police Auditor did not elevate its concerns, San José Police sent the investigation to the auditor with about two and a half months remaining in the one-year time frame prescribed by state law.

In April 2021, the city of San José and the San José Police Officers' Association agreed that, until at least June 2022, the police department must send its investigations to the San José Police Auditor for review within nine months. This time frame guarantees the San José Police Auditor at least three months to elevate its concerns related to each case. According to the San José Police Auditor, this change has increased its ability to review investigations and make necessary appeals. However, it stated that it would be able to pursue more appeals if it had more time to review investigations.

Despite these limitations, the San José Police Auditor has increased the transparency and accountability of San José Police's misconduct investigation process, including in its investigations of bias. With the proper guidelines in place, independent oversight like that provided by the San José Police Auditor could help improve accountability in investigations of bias-related misconduct throughout the State. As we describe in Chapter 4, departments have frequently struggled to identify indications of bias, and their investigations at times caused us to question whether they had viewed officers' behavior objectively. Implementing an independent review process for bias-related misconduct investigations across the State would likely improve how well departments handle such allegations and promote increased public trust in departments' complaint resolution processes.

In 2019 US DOJ recommended that law enforcement departments consider instituting routine, third-party audits of internal affairs investigations to increase the public's trust in the departments' decisions. Because our review focused on investigations of possible biased conduct, we do not know whether the departments we reviewed regularly fail to adequately investigate other types of officer misconduct. However, it is reasonable to assume that external oversight would incentivize better performance of those investigations as well. The goals and principles of effective independent oversight, as described by US DOJ and the National Association for Civilian Oversight of Law Enforcement (NACOLE), relate to departments' investigations more broadly, rather than just bias-related investigations. ¹⁵ NACOLE is a nonprofit organization that is dedicated to promoting greater police accountability through the establishment or improvement of citizen oversight agencies.

One approach to a statewide oversight framework for misconduct investigations would be to rely in part on existing local entities. Many local governments across the nation have established processes for independent oversight that appear similar to the functions of the San José Police Auditor. These include California localities smaller than those we reviewed, such as the city of La Mesa. However, local governments have taken a wide variety of approaches in instituting these processes—and, according to a report by NACOLE, it is likely that no single model of independent oversight will work for all jurisdictions.

For these reasons, we recommend the State adopt a framework similar to that in Figure 13. Under this approach, DOJ would be responsible for establishing criteria for successful local oversight of law enforcement investigations, such as the criteria we detail in Figure 13. For local governments that do not provide oversight that meets these criteria, state law would require DOJ—which already conducts broader reviews of some departments and collects departments' data on bias-related complaints—to audit the associated law enforcement departments' investigations of possible biased conduct.

Figure 13

The State Could Implement Greater Oversight of Investigations of Officer Misconduct

Please refer to the recommendations section to find the recommendations that we have made as a result of these audit findings.

---

We conducted this performance audit in accordance with generally accepted government auditing standards and under the authority vested in the California State Auditor by Government Code section 8543 et seq. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on the audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives.

Respectfully submitted,

MICHAEL S. TILDEN, CPA
Acting California State Auditor

April 26, 2022

---

Appendix

Scope and Methodology

The Audit Committee directed the California State Auditor to evaluate the processes that CDCR and a selection of local law enforcement departments use to vet law enforcement officers for potential affiliation with hate groups and to address bias. As part of this audit, we reviewed Los Angeles Sheriff, San Bernardino Police, San José Police, and Stockton Police. The table below lists the objectives that the Audit Committee approved and the methods we used to address them.

Audit Objectives and the Methods Used to Address Them

AUDIT OBJECTIVE		METHOD
1	Review and evaluate the laws, rules, and regulations significant to the audit objectives.	Reviewed federal and state statutes, rules, regulations, and court cases related to local law enforcement departments' vetting of officers for potential affiliation with hate groups and biased conduct.
2	Determine the actions state entities, such as DOJ and Cal OES, take to collect and maintain information about hate groups and their membership. Identify and assess any processes these agencies employ in making information about membership in such hate groups available to local law enforcement agencies and CDCR for employment screening purposes.	Interviewed staff and reviewed records at both DOJ and Cal OES. Verified that neither department tracks or maintains information about individuals based on their beliefs or associations. Reviewed information the local departments received from DOJ related to criminal background checks.
3	Assess the sufficiency of guidance—including that from POST—provided to local law enforcement agencies and CDCR related to diversity and screening individuals for associations with hate groups during peace officer recruitment, hiring, and ongoing employment.	Collected and analyzed guidance and best practices regarding effective practices for identifying and addressing bias, including guidance related to recruiting diverse applicants and screening for hate group affiliation during peace officer recruitment, hiring, and ongoing employment.
4	Determine whether any legal impediments or employment-related issues exist that would prevent local law enforcement agencies and CDCR from identifying and making employment decisions about peace officers and peace officer applicants who are members or associates of hate groups, including but not limited to any provisions in state or federal law.	Reviewed relevant federal and state statutes, regulations, and court cases.
5	For a selection of four local law enforcement agencies, evaluate the agencies' processes for the following: Screening peace officer applicants for past or current membership in or association with hate groups, collecting information about any such associations, and reaching hiring decisions. Monitoring to determine whether peace officers develop associations with hate groups while employed. Recruiting and hiring peace officers who reflect the agencies' respective communities.	For the four local departments and CDCR, obtained and analyzed data on the diversity of the departments' current personnel and, to the extent possible, the departments' applicant pools. Reviewed documentation of the departments' policies and practices for recruiting diverse individuals and evaluated diversity-focused recruitment efforts. Reviewed the departments' policies and procedures for screening peace officer applicants during the hiring process. For a selection of five applicants at each department, assessed the degree to which the departments complied with state law and regulation and implemented best practices related to screening for biased conduct, including affiliation with a hate group. Collected and analyzed documentation regarding the departments' efforts to monitor for hate group affiliation or other potential signs of bias, including the departments' implementation of early intervention systems.
6	Assess whether the selected local law enforcement agencies have implemented effective practices in the following areas: Training peace officers about bias and prejudice and the threats that bias and prejudice can pose to effectively protecting and serving communities. Implementing processes to identify biased or prejudicial statements or actions by peace officers—including those alleged in complaints against officers—and to intervene through education, disciplinary action, or both. Promoting a diverse and inclusive workforce and professional culture.	Reviewed best practices for establishing cultures of fair and impartial policing, including best practices related to recruiting, hiring, training, community engagement, early intervention systems, and internal investigation. Reviewed best practices and requirements for providing training regarding bias, including recommended frequency, content, and method of delivery. Evaluated the effectiveness of each law enforcement department's training on bias against identified best practices. Reviewed and evaluated each local department's efforts to implement best practices regarding engaging and building strong relationships with its community. For a selection of misconduct investigations at each local department and CDCR—many of which originated from complaints against officers—assessed whether the departments had taken adequate measures to investigate, identify, and respond to indications of biased conduct. This review was designed to assess departments' responses to potential indications of biased conduct, not to holistically review departments' broader complaint intake and investigation structures or processes.
7	For a selection of four local law enforcement agencies and CDCR, do the following to the extent possible: Identify biased or prejudicial statements, postings, or actions by peace officers, including but not limited to those alleged in official complaints. Identify past or current memberships, associations, or participation in any known hate groups.	Reviewed records associated with complaint investigations and other internal investigations of misconduct by peace officers to identify whether officers had engaged in biased conduct or were affiliated with a hate group. For a selection of about 750 officers across the five departments, used a contracted social media investigation company to identify public social media profiles belonging to those officers, and reviewed the officers' public postings for biased statements and evidence of affiliation with hate groups.
8	Review and assess any other issues that are significant to the audit.	None identified.

Assessment of Data Reliability

The U.S. Government Accountability Office, whose standards we are statutorily obligated to follow, requires us to assess the sufficiency and appropriateness of the computer‑processed information we use to support our findings, conclusions, and recommendations.

In performing this audit, we relied on electronic data obtained from the U.S. Census Bureau, CDCR, Los Angeles Sheriff, San Bernardino Police, San José Police, and Stockton Police to compare demographic information of the corresponding cities, county, and State with the demographic information of applicants and current sworn employees for each department. We reviewed existing information about the data. As a result, we identified limitations with some of the data. Specifically, demographic information for applicants and officers at the departments is self-reported and incomplete. Therefore, we found that this data was of undetermined reliability for our purposes. Although this determination may affect the precision of the numbers we present, there is sufficient evidence in total to support our findings, conclusions, and recommendations.