The national outcry and months of Black Lives Matter protests against police brutality that followed the police killings of George Floyd and Breonna Taylor are a resounding demonstration of the public’s interest in combatting police violence, particularly excess force used on Black Americans. While media attention on police killings increased after Ferguson Officer Darren Wilson killed Michael Brown in 2014, one piece of the story is often missing: the story of the officers. In particular, the public rarely learns details about the involved officers’ personnel, disciplinary, or misconduct histories. Strong state confidentiality laws mean that the public cannot access these records. But civil rights suits against police officers should provide one way for the public to learn about officers’ misconduct.
This Note shows, however, that police combine protective orders and settlement to bind plaintiffs to silence and keep misconduct records from becoming public. Using a case study of 42 U.S.C. § 1983 suits filed against the New York City Police Department, I find that protective orders are common, and that officers and their city attorneys use them strategically. A textual analysis shows that every protective order explicitly protects police personnel and misconduct records. And cases with protective orders have statistically significantly higher settlement amounts than those without. These results are consistent with a conclusion that NYPD officers, and the City of New York, may pay more to keep misconduct records secret. The standards to modify a protective order are burdensome, and plaintiffs’ and judges’ incentives align against fighting or denying protective orders. Because of these roadblocks, the public has no opportunity to learn about problem officers’ past misconduct that comes to light during civil rights litigation. To facilitate the necessary changes to protect the public and reform or abolish police, the federal government should create a national database of police misconduct records.
Americans who watched the video of Minneapolis Police Officer Derek Chauvin fatally driving his knee into George Floyd’s neck or heard the news that Louisville Police Officers Brett Hankison, Jonathan Mattingly, and Myles Cosgrove shot and killed Breonna Taylor in her apartment shortly after midnight were undoubtedly enraged, but likely not surprised. Following the string of police killings of Michael Brown, Laquan McDonald, Eric Garner, and Tamir Rice in 2014, the Black Lives Matter Movement brought a renewed public consciousness to police brutality—though many Black communities and other communities of color needed no reminder.
While police brutality and police killings are not a new problem, the American media and general public seem to rediscover police violence in waves. President Herbert Hoover’s National Commission on Law Observance and Enforcement documented police abuses in its Report on Lawlessness in Law Enforcement in 1931. Modern policing had only begun a century earlier, but its first one hundred years were marked by serious police misconduct. The civil rights era brought graphic images of police brutality to the American public as police attacked peaceful protesters with dogs and fire hoses. And again, the video footage of four Los Angeles police officers brutalizing Rodney King in 1991 placed police violence and misconduct in the public mind. As Americans have begun to carry smartphones, bystanders—and even victims like Sandra Bland—have continued to capture concrete evidence of police violence.
The national response to the police killings of George Floyd and Breonna Taylor has demonstrated more than just public interest in police killings. Nationwide, Americans, led by community activists of color and the Black Lives Matter movement, have demanded police reform. From May 26, 2020, the day after Floyd’s killing, through June 9, 2020, hundreds of thousands protested police brutality against Black individuals. Protests occurred in two thousand cities and towns in the United States. Protesters chanted the names of George Floyd, Breonna Taylor, and a litany of other victims of police brutality.
But the names and histories of the officers responsible for civilian deaths tend to be less well-known than those of their victims. This is not because the public is uninterested in officers who use excessive force. Instead, police have successfully mobilized a variety of tactics to keep officers’ on-duty misconduct secret. Strict state confidentiality laws keep officer misconduct records concealed in most states. And prosecutors’ failure to indict officers who kill civilians keeps details about these officers from coming out at a criminal trial. Even when the state prosecutes officers, rules of evidence governing relevancy can prevent prosecutors from introducing evidence of prior police misconduct.
Federal civil rights suits against police are one way that the public could gain information about misconduct that local police officers and departments work hard to conceal. Section 1983 of Title 42 (section 1983) allows individuals to sue government officials for violations of their constitutionally protected rights. Using this private right of action, civilians who face police force can sue the perpetrating officers and the departments that employ them without the political baggage that local prosecutors face when deciding whether and how to prosecute police officers. Because FRCP 26(b) authorizes broad discovery, victims of police brutality who sue police can obtain police misconduct records and department training materials before federal trials.
Though federal civil rights suits are, in theory, a tool for increasing transparency regarding police misconduct, in practice, the public learns little from them. This Note finds that the evidence of past police misconduct or harmful departmental policies that civil rights plaintiffs obtain during pretrial discovery typically must remain confidential. This is because a large proportion of plaintiffs suing police agree to settlement terms that keep police personnel information, misconduct, and training records secret. Using a case study of section 1983 suits brought against the New York City Police Department, I find that stipulated protective orders covering police records are common in these suits. Strikingly, cases with protective orders have statistically significantly higher mean settlement amounts than those without. This finding correlates with the possibility that NYPD police officers, and their city attorneys, prefer to pay out higher settlement awards that bind parties to secrecy than risk the chance that damaging misconduct records will become public.
To explore the role that protective orders play in civil rights suits against police, this Note proceeds in five parts. Part One describes the current legal standards for issuing protective orders and the policy justifications for permitting them in federal litigation. Part Two explains that protective orders are an ill fit for section 1983 civil rights suits against police due to the public’s interest in obtaining officer misconduct records. Part Three explores the prominence of protective orders in federal civil suits against police through an empirical, case-study analysis of suits against New York City police officers between 2014 and 2019. Part Four argues that under current law, the incentives of all parties enable police to exploit protective orders to favor their confidentiality interests over public health and safety concerns. Part Five proposes a solution. Specifically, Congress should create a national database for police misconduct records. The ever-increasing death toll of civilians, particularly Black Americans, killed by police and the mounting public outcry over police brutality demand nothing short of a systemic overhaul. This proposed national misconduct database would be a step toward increasing transparency, putting public pressure on officers who have engaged in misconduct, and preventing future police violence.
Parties to civil litigation may “share what they learn in discovery with other persons, including news media, as they see fit.” Courts in a majority of federal circuits have affirmed a litigant’s ability to use pretrial discovery for any lawful purpose. Some authorities ground a litigant’s ability to disseminate pretrial discovery in the First Amendment while others derive this power from the absence of language in the Federal Rules of Civil Procedure, which “limit[s] a party’s use of information or documents it obtains through discovery.” While the source of the power is not settled, the power itself seems uncontroversial and can be broadly exercised. For example, in an Eastern District of Virginia case, the court found that the plaintiffs—former employees of the private security companies owned by Erik Prince—could publish on their counsel’s website discovery materials that a valid judicial order did not make confidential. Thus, even though “pretrial discovery . . . is usually conducted in private,” parties may disseminate and “use discovery information as they wish” when it is not protected.
But judicially issued protective orders sharply curtail litigants’ ability to share pretrial discovery with non-parties. The Supreme Court has made clear that a court may restrict a litigant’s power to disclose pretrial discovery by issuing a protective order that complies with Rule 26(c) of the Federal Rules of Civil Procedure (FRCP).
Under FRCP 26(c)’s permissive terms, a court may choose to keep discovery materials confidential upon a party’s showing of “good cause . . . to protect a party or person from annoyance, embarrassment, oppression, or undue burden or expense.” In the Third Circuit,“[g]ood cause means ‘that disclosure will work a clearly defined and serious injury to the party seeking closure.’” The Sixth and Ninth Circuits have adopted similar definitions. To demonstrate this injury, the moving party must “articulate specific facts to support its request.” The rule gives courts “substantial latitude to fashion protective orders.” Once entered, a protective order may keep discovery material confidential long past the end of the litigation. Protective orders stay in place unless a party, or an interested intervenor, moves to lift or modify the protective order.
Parties may also present a stipulated protective order to the court. Under these orders, parties agree to confidential terms. Theoretically, courts must find good cause exists before issuing such an order, even when parties agree to keep discovery confidential. In fact, in 1995, the Advisory Committee on Civil Rules reaffirmed that courts must engage in this inquiry. It rejected a proposal to modify Rule 26(c) to permit courts to grant a protective order for either good cause or “on stipulation of the parties.” A letter from the Committee on Rules of Practice and Procedure of the Judicial Conference of the United States explained that members “voted to delete the words ‘on stipulation of the parties’” because of concern that such a rule “would tie the hands of trial judges reluctant to accept agreed orders.” Therefore, the plain text and history of FRCP 26(c) show that courts must find good cause to grant a protective order, regardless of whether it is contested or stipulated.
Some federal courts have affirmed this rule. The Seventh Circuit has squarely held that district courts must “independently determine if ‘good cause’ exists” before granting a stipulated protective order. Decisions in the Third and Federal Circuits similarly suggest but do not clearly establish, that district courts must find good cause for stipulated protective orders.
However, other federal courts have ruled to the contrary. The Ninth Circuit has held, without citation to authority, that “[w]hile courts generally make a finding of good cause before issuing a protective order, a court need not do so where . . . the parties stipulate[d] to such an order.” The Eleventh Circuit has also shifted the requirement that parties establish good cause to a district court’s consideration of “a motion to modify . . . a stipulated protective order” instead of the moment the district court first grants it. The Second Circuit has intimated that “parties might enter into an agreement or stipulate to protect the confidentiality of discovery materials before presenting a proposed protective order to a court,” and later use this agreement to prevent the court from modifying it on reliance grounds. The Second Circuit’s approach acknowledges that a court must approve a stipulated protective order but simultaneously suggests that the parties can prevent a court from engaging in a thorough good cause analysis before approving it. And while the Sixth Circuit in an earlier case appeared to condemn courts issuing stipulated protective orders without finding good cause, it now acknowledges that “courts often issue blanket protective orders that empower the parties themselves to designate which documents contain confidential information.” Therefore, in these circuits, when all parties agree to a protective order, they “postpone, perhaps indefinitely, the obligation to make a particularized showing” explaining why a protective order is justified.
The trend toward permitting parties to stipulate to protective orders without a judicial finding of good cause leads to a “disturbing” pattern in which “some courts routinely sign orders which contain confidentiality clauses without considering the propriety of such orders, or the countervailing public interests . . . .” A review of one hundred proposed stipulated protective orders issued in federal courts in January 2018 found that courts and parties use “generic language to describe the need for the protective order” and fail to reach the particularized good cause standard.
But even when courts do engage in the required good cause analysis before granting a stipulated protective order, they typically consider only the moving parties’ interests and disregard broader public interest concerns. The Third Circuit’s Pansy factors stand alone in asking district courts to consider the “public importance” of a case, its impacts on general health and safety, and litigants’ status as public officials before granting a protective order. This singular focus on the litigants’ interests follows from Seattle Times Co. v. Rhinehart’s holding that the public has no First Amendment or common law right to access pretrial discovery. And because the amended FRCP 5(d) no longer requires parties to file discovery documents until “they are used in the proceeding,” the public cannot claim that pretrial discovery is a judicial document subject to public access. This precedent suggests that private parties can invoke judicial power to endorse their private agreements with little to no consideration of public interest.
Parties began using stipulated protective orders in complex, commercial cases with a large volume of discovery. The drafters of the FRCP recognized the expanding use of protective orders in commercial litigation when amending Rule 26(c) in 1970. While the advisory committee comment explains that they amended the rule “to give [protective orders] application to discovery generally,” they also added language explicitly acknowledging that courts may issue protective orders to protect trade secrets, confidential research and development, or commercial information. This new language “reflect[ed] existing law” applying protective orders to “confidential commercial information” exchanged in litigation.
But stipulated protective orders are now “commonplace in the federal courts” and associated with litigation of all types. And it’s not hard to see why. Stipulating to confidentiality accelerates discovery and satisfies both parties. The proponent gets to keep discovery materials confidential, and the opponent may access information it could not otherwise receive without extended litigation. In many ways, a stipulated protective order is a contract to not distribute discovery materials. But unlike a standard private contract, stipulated protective orders receive judicial approval and are subject to judicial monitoring and enforcement. Parties that violate a protective order are subject to contempt—a more immediate and potentially harsher sanction than breaching a contract.
The proliferation of protective orders in civil litigation reflects an assumption that such litigation is primarily a mechanism for resolving private disputes. The fact that discovery and “information exchange . . . takes place out of the public eye and without involvement by the judge” supports this perspective. And despite the Federal Rules’ operation as a “broad discovery regime,” scholars like Arthur Miller argued that they “never intended that rights of privacy or confidentiality be destroyed in the process.” Under this understanding of the court system, allowing litigants to contract for confidentiality is more important than facilitating public access. This is because it allows litigants to resolve disputes without sacrificing their privacy, protects their property, and improves court efficiency while reducing costs. For advocates of protective orders, the paradigmatic litigant in need of protection is businesses with commercially valuable information.
Richard Marcus, a supporter of protective orders, noted that civil discovery does sometimes alert the public to concerns and provides information the public would not otherwise have access to. But he contended these are “collateral effects” that “should not be allowed to supplant [the] primary [dispute-resolution] purpose” of litigation. Marcus argued that regulators, rather than courts, should be the means by which the public is informed of issues that implicate public health and safety.
Arguments defending litigants’ ability to stipulate to protective orders assume that the purpose of civil litigation is to resolve private disputes. These policy rationales may resonate for commercial disputes between private parties and relationship-based litigation like divorce. But Part II questions whether these same arguments justify courts endorsing litigants’ confidentiality agreements in section 1983 suits, which fundamentally concern the behavior of public officials.
II. The Public’s Interest in Police Misconduct Records Make Stipulated Protective Orders an Ill Fit for Civil Rights Suits Against Police
Stipulated protective orders should not be “commonplace” in civil suits against police. Section 1983 suits against police officers are unlike the typical suits that the private dispute resolution model contemplates because a section 1983 suit involves a public officer by definition. Under 42 U.S.C. § 1983, local or state officials acting in their official capacity “under color” of law are liable for depriving the “rights, privileges, or immunities secured by the Constitution and laws” of any person in the United States. Common causes of action in section 1983 suits filed against police officers include: use of unreasonable force in violation of the Fourth or Fourteenth Amendments and unlawful arrests, stops, frisks, searches, or seizures in violation of the Fourth Amendment.
Because “[l]awfulness of police operations is a matter of great concern to citizens in a democracy,” the concept of civil litigation as resolving private disputes seems to be an ill fit for suits that involve public officers. These cases “represent a balancing feature in our governmental structure whereby individual citizens are encouraged to police those who are charged with policing us all.” Instead, section 1983 suits are better conceptualized under a model of public law litigation.
Public law litigation introduced an alternative approach to the private dispute model of litigation. As first described by Abram Chayes, the original concept of public law cases had these key attributes:
These . . . cases involved amorphous, sprawling party structures; allegations broadly implicating the operations of large public institutions such as school systems, prisons, mental health facilities, police departments, and public housing authorities; and remedies requiring long-term restructuring and monitoring of these institutions.
Civil rights cases and the structural remedies they demanded first exposed federal courts to public law litigation and set the characteristics of the model. These suits proliferated following the 1966 amendments to the Federal Rules of Civil Procedure. Modifications to nonparty joinder in Rule 19, class actions in Rule 23, and nonparty intervention in Rule 24 made “party structure more flexible” and enabled large groupings of plaintiffs to sue governmental bodies to effect policy change.
Section 1983 suits against police do not, and cannot, always satisfy each of the typical characteristics of public law litigation. A private litigant’s power to sue police departments for injunctive relief, and therefore create court-enforced policy changes, has been stymied by City of Los Angeles v. Lyons’s equitable standing doctrine. Following Lyons, private individuals can only seek injunctions against police departments when they can show that they are “realistically threatened by a repetition” of the same injury they already suffered. To receive injunctive relief, those who have suffered police abuse must now allege that “all police officers” in the locality “always [harm] any citizen with whom they happen to have an encounter” in the same way or “that the City ordered or authorized police officers to act in such manner.” These doctrinal developments have slowed “institutional reform litigation in policing” and prevented individuals from acting as “direct agent[s] in effecting meaningful social change through America’s courts.”
But key attributes of public law litigation remain in section 1983 suits against police. As Chayes described, these modern suits still “do not arise out of disputes between private parties about private rights. Instead, the object of litigation is the vindication of constitutional or statutory policies.” The defendants in these suits are police officers, police departments, and often the city, county, or state government that grants police their authority to operate. The causes of action in section 1983 suits often extend beyond discrete officers and implicate police department training and policy. Therefore, these suits often “seek to vindicate important social values that affect numerous individuals and entities.”
And section 1983 suits against police still have the power to spur police departments into changing their policies and practices. Suits for money damages can force police departments to enter into settlement agreements with provisions for reform. These cases include a suit that required “Wilmington Police [to] evaluate its deescalation tactics and training for officers”  and a suit that forced the “Los Angeles Police Commission [to] . . . agree to ban a controversial form of restraining” suspects known as “hogtying.” Certain judicial decisions in section 1983 cases have even required police departments to release documents detailing systemic misconduct. In a federal case brought against the Houston Police Department and City of Houston for the shooting of Kenny Releford, a federal judge refused to seal certain discovery materials despite the City of Houston’s attorneys’ and the police union’s opposition. Following the court’s ruling, “the Houston Police Department was compelled to release previously secret internal reviews of Releford’s shooting as well as of other unarmed Houstonians.”
But for these suits to have a chance at sparking change, there must be publicity. A successful suit against police or simply credible allegations of harm before a court “will focus public attention on the problems,” and “increased scrutiny will generate diffuse but sometimes powerful pressures for responsible behavior.” Further, as Joanna Schwartz has found, “most departments ignore lawsuits that do not inspire front-page newspaper stories, candlelight vigils, or angry meetings with the mayor.” Therefore, transparency in private citizen suits against police is critical. Making the public privy to filings and discovery in these suits can do necessary work to “help citizens police the government by forcing governmental entities to release information that would otherwise be kept secret.”
The public has an interest in police misconduct and disciplinary records produced in discovery because access to these records enables the public to identify abusive officers. This access is necessary to hold these abusive officers, and, more importantly, the departmental policies that produce them, to account. But due to the dearth of criminal trials brought against police officers, police union record destruction policies, and strict state confidentiality protections for officers, civil discovery may provide one of the few remaining avenues through which misconduct records could be disclosed.
Residents of a given community and the public at large have an interest in knowing about dangerous officers, as well as police departments’ systematic failures to train or reprimand officers for misconduct or crime. Knowing this information can give the public critical data that it can use to pressure police departments and governments into terminating abusive officers and reforming training and oversight policies. And the public has a particular interest in police misconduct records produced in section 1983 litigation discovery because these suits tend to identify officers who have engaged in or have histories of misconduct. Specifically, police officers who seek to protect their records as defendants in section 1983 suits may desire confidentiality because they have disciplinary histories that if publicly exposed, could reveal their past serious misconduct.
Many police officers involved in high-profile killings have previously been the subject of both civilian complaints and lawsuits. Derek Chauvin, the Minneapolis officer who killed George Floyd “received at least 17 complaints” in his nearly twenty years on the force. He was “involved in the fatal shooting” of another person and shot a suspect who survived. He was also “named in a brutality lawsuit.” While Minnesota was one of the few states that published civilian complaint data at the time of the incident, the public data itself could not predict Officer Chauvin’s future dangerousness because it both undercounted the number of complaints brought against Chauvin and lacked details about the facts of the incidents.
Tou Thao, one of the three other officers involved in Mr. Floyd’s killing, had six misconduct violations and was sued for brutalizing a man in 2017. In that 2017 suit, all parties entered into a protective order that made “Minneapolis Police Department Personnel files,” “Minneapolis Police Department Internal Affairs records,” and “Minneapolis Civilian Review Authority and Office of Police Conduct Review records” confidential. The order required counsel to “return all Confidential Materials” once the action terminated, and prohibited parties from using the materials “or information derived from them . . . for any other purpose other than for this Action.” Minneapolis settled the case for $25,000.
Similarly, Jason Van Dyke, the Chicago police officer who murdered Laquan McDonald, had a history of complaints about excessive force. “[I]n his 17 years on the police force, [he] accumulated 20 documented citizen complaints against him, mostly for excessive force.” A rare jury trial also found him liable for using excessive force during a traffic stop. This case bound the parties with a protective order during pretrial discovery. The order protected “personnel files, disciplinary actions, histories, [and] files generated by the investigation of complaints of misconduct by Chicago police officers.” It explicitly determined that the Illinois Personnel Records Review Act, and Section 7 of the Illinois Freedom of Information Act, protected the records. The protective order still covers those discovery materials that did not become part of a judicial record.
Empirical analysis substantiates the intuition that officers with more misconduct records are more likely to use excessive force against civilians in the future. Kyle Rozema and Max Schanzebach analyzed fifty thousand civilian allegations of misconduct made against Chicago Police Officers. They found “a strong nonlinear relationship” between officers who were the subjects of civilian allegations and officers who engaged in serious misconduct, as measured by the officers being named as defendants in section 1983 suits. A nonlinear relationship in this case means that each increase in civilian complaints against an officer did not correspond with an equal increase in the probability that the officer would be a defendant in a civil rights lawsuit. Instead, the “worst 5 percent of officers” in terms of the volume of complaints against them were subject to the most litigation, and the “worst 1 percent of officers . . . generate[d] almost 5 times the number of payouts and 4 times the total damage payouts in civil rights litigation.”
Rozema and Schanzenbach’s research links histories of misconduct to abusive officers who are sued under section 1983. Yet, it is precisely these histories of misconduct that protective orders obscure. Had the public been privy to the details of Officer Chauvin’s, Officer Thao’s, and Officer Van Dyke’s misconduct records prior to their killings of George Floyd and Laquan McDonald, respectively, Minneapolis or Chicago residents could have pressured the departments to fire these officers or establish robust mechanisms for evaluating and terminating problem officers. The section 1983 suits against them could have filled this informational hole but did not because of protective orders.
Misconduct and training records provide a key source of information that the public needs to hold police officers accountable for their actions. Without record transparency, police departments can “claim that a fully functional police accountability system exists––whether true or not––without any contradictory evidence publicly accessible.” And “[e]ven if it is a functional system, depriving the public of any ability to judge for itself is not justice.” The public and policymakers need access to records to judge whether internal police disciplinary procedures are effective.
Outsiders must apply external pressure because internal investigations rarely punish officers. Rachael Moran identified that “the DOJ has . . . found repeated instances in which civilians’ complaints reported through an internal intake process were never investigated.” When officers are investigated, “the officers investigating these reports have an inherent inability to conduct impartial investigations” and ultimately resist “disciplining their own officers . . . [for] even the most obvious misconduct.” Empirical evidence confirms that police departments reject civilian complaints at high rates. Law enforcement agencies in California upheld only 8.4 percent of civilian complaints between 2008 and 2017. In Chicago, the complaint process takes “about one year” to complete and may be followed by “a lengthy appeals process for any resulting discipline.” And in New York, between 1975 and 1996, the department only terminated 2 percent of NYPD officers following misconduct allegations.
It is possible that police departments discipline few officers following complaints because the complaints themselves are frivolous. But researchers like Phil Stinson have found that a small percentage of officers are the subjects of consistent complaints, suggesting through sheer repetition that the complaints against these officers have merit. Further, for the police department or associated complaint review board to treat a claim formally as a complaint, the complaining party must speak with an investigator, often at the investigating body’s office or at the police department. Because a complaint is not officially a complaint when the agency receives a letter, email form, or phone call, the effort required to make a complaint may serve as a fair proxy of how serious these claims are.
Having access to disciplinary, personnel, and training records would provide the evidence necessary for future litigants to successfully seek injunctions against harmful police department policy under Monell v. Department of Social Services. Section 1983 suits on a respondeat superior theory cannot hold local governments liable for the actions of individual officers. However, Monell allows section 1983 suits against a local government or municipality that “implements or executes” an unconstitutional “policy statement, ordinance, regulation, or decision.” To meet the requirements of Monell liability, plaintiffs need documents that can prove the department engaged in an unconstitutional custom or failed to hire adequately, train, and supervise. Giving the public access to police misconduct and training records can therefore remove an informational block to succeeding in these actions. With adequate access to proof of departmental policy or custom, Monell suits can encourage broader departmental change by enjoining unconstitutional departmental conduct.
Unfortunately, section 1983 suits provide one of the only remaining ways that members of the public may gain access to officers’ disciplinary and misconduct records, as well as police department training manuals.
Criminal cases have not been a strong mechanism for bringing police misconduct records to light because the criminal legal system was not designed to punish police officers. Local and federal prosecutors have been resistant to charging officers who kill or injure civilians. A 2010 study evaluated 8,300 misconduct accusations based on a dataset that collects credible incidents of misconduct from media reports. Of this subset of publicized, credible police misconduct accusations, only 39 percent resulted in legal action of any kind.
Prosecutors also may present half-hearted cases to grand juries that fail to indict police officers for assault or murder. For example, the District Attorney’s Office in Hennepin County where Officer Chauvin killed George Floyd did not return an indictment for a single officer for any of the forty-two officers involved in killings in the county between 2000 and 2016. Prosecutors similarly failed to return indictments for the officers who killed Eric Garner and Michael Brown. And while grand jury records, and the experiences of grand jurors themselves, are presumptively secret, a juror was permitted to speak about their experience on the grand jury in the case against the Louisville police officers who killed Breonna Taylor. The grand jury returned only three counts for wanton endangerment against Brett Hankison for shooting at Ms. Taylor’s apartment and hitting a neighboring apartment during the midnight. But the grand juror contended that the prosecutor did not present evidence to support a homicide offense or explain those laws. In fact, the juror stated that other grand jurors “asked about additional charges” but were told “there would be none because the prosecutors didn’t feel they could make them stick.” This juror was only allowed to speak after a court granted their motion to describe their experience over the objections of the Kentucky Attorney General Daniel Cameron, who led the prosecution. These anecdotes lend credence to Kate Levine’s argument that due to the close police-prosecutor relationship, conflict of interest laws should disqualify local prosecutors from prosecuting officers, but they rarely do.
This combination of grand jury secrecy and lack of prosecution creates another barrier to public disclosure. Thus, criminal suits against police officers rarely serve as fora for revealing details of police officer crime and misconduct.
Police unions also include a variety of clauses in their union contracts to keep misconduct records secret. Stephen Rushin’s analysis of 178 police union contracts from U.S. cities with over one hundred thousand residents shows that many police contracts provide for regular destruction of misconduct records. Specifically, “eighty-seven . . . collective bargaining agreements” of the 178 contracts he reviewed (48.9 percent) “require[d] the removal of personnel records” at set times. This means that certain officers “can have his or her personnel file wiped clean” every two years even when records contain sustained claims and show a pattern of misconduct. Further, “many police union contracts prevent even police chiefs from fully using officer disciplinary records.”
State confidentiality laws also prevent the public from accessing misconduct files and disciplinary information about the police officers that serve them. Currently, thirty-eight states make disciplinary records completely confidential or sharply curtail public access. Some states have adopted statutes that explicitly make officers’ records confidential, like a Delaware statute entitled “Law-Enforcement Officers’ Bill of Rights.” Other statutes shield personnel or disciplinary records of all state or municipal employees, which necessarily include police officers, from public view. Departments in Arkansas, Hawaii, and Indiana only make the records of terminated or suspended officers public.
Police departments often broadly construe language in their states’ public records acts or state constitutional rights to privacy to make disciplinary records confidential in practice. Departments in D.C. rely on broad privacy protections in public record laws to deny requests for records. Kentucky police departments will respond to requests with “heavily redacted records” that merely list “disciplinary actions.”
Some courts have occasionally pushed back on broad readings that exempt certain police records from disclosure. The First District Appellate Court of Illinois granted a plaintiff access to complaints against Chicago Police officers, rejecting the CPD’s argument that the Illinois FOIA exceptions should be read broadly to include complaints. But others, like the court in Maryland Department of State Police v. Dashiell, have shielded records even from the individual who filed the complaint against an officer.
Only twelve states make police disciplinary records subject to disclosure under a state’s public disclosure law. But there are still limits on what records may reach the public. For example, in Florida, Georgia, Maine, and Minnesota, disciplinary records are available only after an internal investigation is finalized while records in Arizona are protected until an appeals process is completed. And in Arizona, public records are subject to a balancing test that considers an officers’ interest in “confidentiality [and] privacy.” In Utah, only substantiated disciplinary records are subject to public disclosure.
Further, civilian defendants in criminal cases face a high standard for obtaining misconduct records to impeach police witnesses despite their constitutional rights to due process and compulsory process. Criminal defendants in state court have a Sixth Amendment compulsory process right and a Fourteenth Amendment due process right to “a meaningful opportunity to present a complete defense.” This includes “[t]he right to offer the testimony of witnesses, and to compel their attendance” and requires the prosecution to “deliver exculpatory evidence into the hands of the accused.” To vindicate these rights, defendants must have access to records about officers who will testify against them at trial to enable impeachment. But Rachael Moran has found that many states “make it extremely difficult for defense counsel to access these confidential [police personnel] records.” Specifically, criminal defendants in Colorado must allege a “‘specific factual basis’” showing that the records exist and that they provide material evidence. Criminal defendants in Arizona, Connecticut, Georgia, and North Carolina must show that the records have information that is relevant to the theory of defense.
Even in states like California that have amended more restrictive confidentiality laws protecting police records, police still challenge practices that enable criminal defendants to obtain impeachment evidence on testifying officers. For example, in Association for Los Angeles Deputy Sheriffs v. Superior Court, the Association for Los Angeles Deputy Sheriffs sued the Los Angeles County Sheriff’s Department to prevent it from disclosing a list of officers whom prosecutors have determined have impeachment material regarding their credibility as a witness. The list, devised as a way for prosecutors to comply with their disclosure requirements under Brady v. Maryland, disclosed very little: only “(a) the name and identifying number of the officer and (b) that the officer may have relevant exonerating or impeaching material in [that officer’s] confidential personnel file.”
The California Supreme Court held that the prosecutors could maintain such a list. But defendants who receive information about this list do not receive the impeachment material directly from the prosecutors. This is because the Pitchess statutes restrict even “a prosecutor’s ability to learn of and disclose certain information regarding law enforcement officers.” Therefore, armed with the information from the list, defendants must still obtain the “personnel records and records of citizens’ complaints” through California Public Record Law if the material is subject to public disclosure under Section 832.7 or comply with procedural requirements under California’s Pitchess statutes. If the defendant must comply with Pitchess, the “party seeking disclosure . . . must file a written motion . . . identify[ing] the officer or officers at issue . . . describ[ing] the ‘type of records or information desired.’” The party must also show good cause for discovery by making allegations of materiality and reasonable belief of the information’s existence. Despite reform, California defendants continue to face obstacles in accessing police personnel records.
Confidentiality laws, prosecutorial resistance to charging, union conditions mandating record destruction, and protective orders combine to keep the public in the dark about dangerous officers in their communities. This opacity can be deadly. Courts can mitigate the effects of widespread secrecy around police records by not granting protective orders in civil rights suits against police or granting motions to modify or lift orders protecting these records. This discretion is a power that courts can and should use to protect the public interest.
Because of the strong public interest in section 1983 suits against police, Eastern District of New York Judge Weinstein in King v. Conde found that “‘[r]outinely’ issuing protective orders will not necessarily promote justice or the proper balance of interests.” While King predates the 2000 amendment of FRCP 5(d), eliminating the discovery filing requirement, Judge Weinstein’s admonition should remain a guiding force for discovery in civil suits involving the police. The mere fact that discovery is no longer publicly filed does not eliminate the public nature of a section 1983 suit or the public’s interest in identifying dangerous officers and holding them accountable. But, as I detail below, the countervailing legal principles and practical realities of case management suggest that courts will grant stipulated protective orders in a civil rights case against police without considering the public’s interest.
To begin to understand the role stipulated protective orders play in federal civil rights suits against police, I analyzed cases brought against New York City Police officers in the Eastern and Southern Districts of New York. I hypothesize that the presence of a protective order results in a higher settlement amount than cases without such orders. I ground this hypothesis in two related assumptions. I assumed that police officers will move for stipulated protective orders when plaintiffs request an officer’s misconduct and disciplinary records. I also assumed that officers are willing to pay a higher settlement award to settle a case with a guarantee that their misconduct records will remain confidential. I tested this hypothesis using a randomly selected probability sample of section 1983 cases filed against the NYPD between 2014 and 2019.
I found that stipulated protective orders are common in suits against the NYPD, but that New York City, representing the officers, moves for them strategically. In reviewing stipulated protective orders’ texts, I found that all protective orders explicitly protect NYPD disciplinary and misconduct records. Ultimately, I found that cases with protective orders terminate in statistically significantly higher settlement values than cases without orders in place.
The following presents my methodology, descriptive statistics, findings from reviewing the text of stipulated protective orders, results from my statistical analysis, a discussion of potential mechanisms that may explain my findings, and the limitations of the results.
To determine whether settlement awards meaningfully differ between section 1983 suits with and without protective orders, I randomly selected 20 percent of cases from New York City’s publicly available “NYPD Alleged Misconduct Matters” database. Below, I describe why I chose this subsample, my data collection process and its limitations, and the qualitative and quantitative methods I used to analyze the data.
I selected federal suits against New York Police officers as my case study sample for two reasons. The first reflects a practical data collection concern. While many major cities maintain open data portals, New York City is among the few municipalities that provides a comprehensive dataset of lawsuits filed against its police officers. New York City maintains the “NYPD Alleged Misconduct Matters” dataset pursuant to Local Law 166, which requires the City to publish information regarding suits against the NYPD. At the time of analysis, this database included all “civil actions alleging misconduct commenced against the police department and individual officers” between 2014 and 2018 in both federal and state court. The second reason for selecting this sample is because of the different access that section 1983 plaintiffs in state and federal court have to New York City and New York state police officers’ misconduct records. During this period, Civil Rights Law section 50-a set a high bar for New York state litigants seeking to obtain NYPD misconduct records. But as described further in Part IV.A, federal litigants were not subject to the same confidentiality provisions and could obtain the records through FRCP 26(b). Therefore, to ensure that litigants in my sample had the ability to obtain police misconduct records, whether unrestricted in discovery or through protective orders, I dropped all state cases from my analysis sample and included only section 1983 suits filed against NYPD officers in the Eastern and Southern Districts of New York.
I randomly selected 20 percent of cases within my sample parameters as my analysis sample instead of using the full dataset due to time and resource constraints. The “NYPD Alleged Misconduct Matters” database contains key variables relevant to my analysis, including docket number, litigation start and end date, attorney representation, disposition, and settlement amount. But the dataset did not include any information on protective orders. Therefore, I manually built my analysis sample by searching the available dockets of each randomly selected case, as described more fully in Part III.A.2. I chose 20 percent of the sample to reduce the time-consuming data collection process while ensuring I had enough data to accurately mimic the distribution of the full dataset. I tested whether the 20 percent subset analysis sample is representative of the full sample by comparing the data on relevant descriptive statistics. I present these results in Part III.B.
The final sample is limited to suits with a final disposition filed in either the Eastern or Southern Districts of New York. In each case, plaintiffs civilly sued individual officers, and often New York City, for constitutional violations under 42 U.S.C. § 1983. After dropping state court cases, the full dataset represents 2,929 unique civil rights suits brought against NYPD officers in federal court. The original federal suit dataset contained 3,545 cases, but 618 cases had not yet reached a final disposition. I dropped these non-final cases from analysis because they may not have reached a litigation stage where a party would move for a protective order; including them could artificially dampen the rate at which parties moved for protective orders.
Before collecting my protective order data, I randomly assigned a number to each case using the Stata software program’s “uniform” and “rank” functions. To select 20 percent of the 2,929 cases, my initial analysis sample included cases assigned a number between 1 and 595.
To determine whether a litigant in the random sample moved for a protective order, I accessed public docket information for each case. I searched Bloomberg Law dockets, Lexis Court Link, CourtListener.com, and PacerMonitor.com using the sample case’s docket number. When I matched a docket to a case in my sample, I read through the docket for indications of whether either party moved for a protective order or a confidentiality order. I manually collected five variables: (1) the presence or absence of a protective order or confidentiality order; (2) the party proposing the order; (3) whether the proposed order was stipulated to or contested; (4) whether or not the court granted the protective order; and (5) the date of any protective order entered. Where available, I also collected the associated proposed and granted protective orders for textual analysis.
Not every case was available from these sources. When I could not find a docket from any of my source websites, I indicated in my master dataset that the docket was missing. To ensure that 20 percent of my final sub-sample included only cases with a docket available, when I could not find a case, I added a replacement case to my master dataset according to its randomly assigned number. In total, I could not identify twenty-four dockets. This approach runs the risk of skewing results if publicly available protective orders are meaningfully different from protective orders that are not made public. However, because only twenty-four of the 595 cases included in the final analysis sample were unidentifiable. The chance that these twenty-four cases are fundamentally different is low.
During this period, some cases in the Southern District of New York were subject to Local Civil Rule 83.10. The rule applies to civil plaintiffs suing New York City, the NYPD, and individual officers under section 1983. Rule 83.10(11) indicates that for qualifying cases, a specified protective order “shall be deemed to have been issued in all cases governed by this Rule.” While Rule 83.10(11)’s “protective order” protects NYPD officers’ misconduct data and plaintiffs’ medical and arrest records, I did not classify these cases as cases with protective orders issued because they did not meet my established criteria for what constitutes a protective order. Specifically, neither party requested the order, and no court entered the order. Instead, when the stock protective orders from Rule 83.10(11) applied, I treated these “protective orders” as private agreements between the parties instead of a judicially granted and endorsed protective order.
This means I may undercount cases for which the parties believed themselves bound to keep NYPD records confidential. But because my analysis sample includes S.D.N.Y. cases covered by Rule 83.10(11) for which the court did grant a specified protective order on the record, I believe the protective orders actually entered by the court following a party’s motion better reflect the actual strategy and goals of the parties.
The dataset did not include some variables that would enhance this analysis. There are no concrete indicators of the plaintiffs’ or the officers’ races or ethnicities. Therefore, I cannot evaluate whether Black individuals or other persons of color are more likely to sue NYPD officers than White individuals. Further, this dataset does not provide information about the basis for the section 1983 claim or the severity of the alleged injury. Accordingly, the results cannot analyze whether specific categories of misconduct correlate with the presence of protective orders at higher or lower rates. Finally, because the data are limited to suits involving the NYPD, the results cannot be generalized beyond Eastern District of New York (E.D.N.Y.), Southern District of New York (S.D.N.Y.), or NYPD defendants specifically. A nationally representative dataset would be difficult, if not impossible, to compile because comprehensive datasets of each section 1983 suit filed against police officers are not available for every police department in the United States, or even for similarly large departments patrolling major cities. But as discussed in Part IV, there is reason to believe that police and their municipal attorneys across jurisdictions share similar interests in keeping misconduct records confidential.
To test my hypothesis that cases with protective orders terminate in higher settlement awards than those without, I used qualitative and quantitative research methods to analyze my sample data.
On the qualitative side, I analyzed the text of forty-eight protective orders entered in my randomly selected sample cases to determine what kind of information parties sought to keep confidential through protective orders. While I could review many case dockets using publicly available sources, full-text protective orders were not commonly available on sites without a paywall. Therefore, the forty-eight protective orders I analyzed represent all the orders I could view. My textual analysis involved reading each order and assigning the text a code if it embodied a relevant analysis category. Codes included: (1) protects NYPD records; (2) protects plaintiff’s information; (3) restricts records use; and (4) describes the protective order legal standard and explains how the order fulfills it.
On the quantitative side, I used an independent samples t-test to compare settlement amounts in cases with and without protective orders. My dependent variable was the log of the settlement price. I used the log of the settlement price because the distribution of settlement values is not symmetrical and skews toward $0, and the settlement values contain extreme outliers. To avoid dropping settlements for $0 after transforming the settlement value to the log, I assigned every $0 settlement a settlement price of $0.01. My independent variables were cases with protective orders and cases without protective orders. The observations are independent because duplicate cases were dropped prior to analysis. I conduct my significance testing at α=0.05.
The full dataset contains 2,929 cases filed against NYPD officers, the Department itself, and often the City of New York in federal court. As a result of the final disposition requirement, most cases began between 2014 or 2016 (73 percent). Civil litigants filed 56 percent of the cases in E.D.N.Y. and filed the remaining 44 percent of cases in S.D.N.Y. Parties settled 85 percent of cases. Only 1.3 percent of cases went to verdict. The median settlement amount was $11,500. Median litigation lasted 301 days. Approximately 10 percent of plaintiffs were pro se. Table 1 provides descriptive statistics for the full dataset compared to the randomly selected analysis sample.
The final analysis sample contains 595 randomly selected cases of the 2,929 cases with final dispositions filed in federal court. While this sample is only 20 percent of all section 1983 suits brought against the NYPD between 2014 and 2019, as Table 2 shows, the cases in the analysis sample and full sample share similar percentages for case disposition, the presiding district court, and pro se plaintiffs. Median time to case disposition differed by two days and median settlement amount differed by $1,000 between the full dataset and the analysis sample.
It is possible that the sub-sample fails to represent the full sample on variables not indicated in the dataset. Factors specific to the parties, like sex and age of plaintiffs and defendants and factors specific to the case, like the nature of the claims and the presiding judge, are not observed in these data. Critically, there is no indicator of whether an officer has a prior misconduct record. But because the sample cases closely approximate the descriptive statistics from the full dataset, it appears to mirror key, known attributes of the full sample.
Table 1: Descriptive Statistics for Full Sample and Analysis Sample
|Full Dataset||Analysis Sample|
|Median Settlement Amount||$11,500||$12,500|
|–Eastern District of New York||1,626||(55.5%)||341||(57.3%)|
|–Southern District of New York||1,302||(44.5%)||254||(42.7%)|
|Pro Se Plaintiffs||304||(10.4%)||54||(9.1%)|
|Suits with Repeat Plaintiff’s Firms||184||(6.3%)||35||(5.9%)|
|Median Time to Disposition (in days)||301||299|
The next sections present my qualitative and quantitative findings.
To understand what information parties to section 1983 suits naming NYPD officers as defendants seek to protect, I reviewed the text of forty-eight protective orders entered in cases within my analysis sample. While I identified 139 cases with protective orders in my analysis sample, the text of only forty-eight (35 percent) were available through publicly available sources. Taken together, each of the protective orders share striking similarities. Most judges signed the stipulated order without alteration or comment. The orders appear to modify the standard form presumptively issued in S.D.N.Y. cases governed by Local Rule 83.10(11). Each explicitly protected personnel and disciplinary records. Nearly all the orders prevent use beyond the purpose of the litigation and maintain the confidentiality of documents past the termination of the litigation. I describe: (1) if and how the orders describe the good cause standard and which rationales are offered; (2) how courts treat protective orders; (3) what information protective orders make confidential; (4) how information is protected; and (5) how the order describes the court’s power to modify and enforce protective orders.
A textual analysis of the forty-eight proposed and granted protective orders shows that most orders recognized that good cause is the standard for entering protective orders under Rule 26(c). The majority, forty-one cases, followed the basic text of the Local Rule 83.10(11) and stated that “good cause” exists under “Rule 26(c)” for the court to enter a protective order. Two orders justified the protective order with references to past agreement and four simply provided justifications for the protective order without mentioning the standard. Two court orders modifying the parties’ proposed protective order explained that “given the nature of the claims . . . certain documents produced in discovery may contain confidential private information for which special protection from public disclosure . . . would be warranted.”
While most orders recognized good cause as the standard, most did not engage in a rigorous good cause analysis. FRCP 26(c) requires a judge to find good cause exists before entering a protective order. Rule 26(c) broadly permits protective orders to “protect a party or person from annoyance, embarrassment, oppression, or undue burden or expense.” But most of the cases did not justify the order on these grounds. Instead, thirty-five orders found the order should be entered because defendant-officers and the City of New York would not produce requested documents unless “appropriate protection for their confidentiality is assured.” Nine cases expanded this rationale to include a statement that both defendants and plaintiffs’ request confidentiality protections. These objections merely reflect a conclusory desire for protection and fail to provide any rationale for why good cause exists to grant the order.
Another common rationale, offered in sixteen cases, was that the defendant police officers and City of New York claimed that the information sought was privileged, including privileges for law enforcement and government. In two cases, both parties claimed privilege covered the documents. While privilege would certainly protect the information the privilege covers from disclosure, it is not the standard for a protective order. As discussed in Part IV.A, state law government and law enforcement privileges do not control federal courts.
Ultimately, only one case adopted the language regarding confidentiality from Rule 83.10’s presumptive protective order. Both Rule 83.10 and the order in Jackson v. City of New York justified the order as satisfying good cause because “the parties seek to ensure that the confidentiality of these documents and information remains protected.” This justification, resting only on party agreement, seems to violate the requirement that the court finds good cause even when parties agree, as previously discussed in Part I.A.
Only one case offered a justification for good cause that seemed to satisfy the articulated good cause standard. The defendant officers in Marrero et al., v. City of New York asserted that disclosure of information to the plaintiffs “would impair the law enforcement operations and objectives of the NYPD” and “would impair the pending Civilian Complaint Review Board (“CCRB”) proceeding related to the underlying incident.” Both stated rationales, particularized to the facts of the case, would likely satisfy Rule 26(c)’s good cause if such harm could be shown.
Table 1 summarizes the “good cause” rationales offered in these cases.
Table 1: Good Cause Rationales
|Rationale||Number of Orders||Percent of Total|
|Objection to Disclosure Absent Confidentiality|
|–by Both Parties||9||18.8%|
|Claims of Privilege|
|–by Both Parties||2||4.2%|
|Agreement to Maintain Confidentiality||1||2.1%|
|Harm to Police Operations and CCRB Proceedings||1||2.1%|
Despite this perfunctory or entirely absent good cause analysis, only one case featured a denial of a protective order for lack of specificity. In Moore v. City of New York, Magistrate Judge Orenstein directed the City to “seek a more specifically targeted protective order during the discovery process,” and later granted the more specific order. This trend is consistent across the full scope of cases in the analysis sample. In only three other cases did a judge partially grant a protective order (2.9 percent of cases). The court in the remaining 135 cases granted the requested protective order (97.1 percent of cases).
Instead of ensuring good cause exists for protective orders by analyzing the moving parties’ reasoning or finding good cause according to its discretion, many judges directed the parties to enter into protective orders at the outset of discovery. In Corbett v. City of New York, Judge Woods stated that “[t]he Court thanks the parties for working together to submit their proposed protective order for discovery in this case. The parties are directed to consult the Court’s Individual Rule 4.C and to submit a proposed protective order that complies with that rule.”  Judges in Avila v. City of New York, Freese v. Mattina, Ortiz v. City of New York, and Siemionko v. City of New York, et al. all directed the parties to draft a protective order as the cases proceeded to discovery. Cases subject to S.D.N.Y.’s Rule 83.10 are presumed to accept the proposed protective order drafted by the court. Some litigants subject to the rule did modify and ask for the court’s endorsements of other protective orders, which S.D.N.Y. courts granted. But Rule 83.10 reflects a clear judicial preference for parties to conduct discovery without the court’s participation, which the court presumptively will enforce. The adoption of Rule 83.10 and the frequency of these directions in E.D.N.Y. cases suggest that judges often encourage parties to enter into stipulated protective orders.
Each order specifically protected certain information, nearly all of it focused on NYPD records and information pertaining to officers. Forty-five orders specifically protected NYPD officer’s personnel records and disciplinary records. Forty-four orders protected investigations by the NYPD and the CCRB into officer conduct. These protections reflect a frequent recognition that courts will grant orders covering this information because Rule 83.10 presumptively protects all personnel records, disciplinary records, and CCRB documents.
But many orders went beyond the stock categories presumptively included in Rule 83.10. Eight orders specified that the materials the Internal Affairs Bureau produced were designated as confidential. Four orders protected “performance evaluations” and eight orders protected “NYPD training materials,” including the Patrol Guidelines, Administration Guide, Operation Orders, training manuals, directories, legal bulletins, and directives. Three orders protected video that NYPD-issued cameras or a NYPD officer took. Two orders included a catch-all provision that specifically protected “[a]ny other NYPD documents produced as a part of Monell discovery.” Only six orders specifically protected private information like the officer’s telephone numbers, home addresses, social security numbers, birthdates, financial information, and tax records.
The dialogue surrounding confidentiality focused squarely on officer records. Strikingly, only ten protective orders explicitly made any of the plaintiff’s records confidential. Seven orders provided protection for plaintiff’s medical or psychotherapy records. Three orders protected both plaintiff and officer medical records. Four orders protected officer’s medical records but not the plaintiff’s. Only four orders protected files related to the plaintiffs’ arrests or criminal history. The relative infrequency of protections for plaintiff’s medical and arrest records is noteworthy because the presumptive protective order under Rule 83.10 explicitly protects these types of documents. This means that the moving party, typically the defendant-officer(s), affirmatively removed plaintiff-protective provisions from the standard order before seeking the court’s endorsement.
Further, nearly every order included a catch-all provision: by labeling a document confidential “in good faith,” parties could choose to protect any document not already protected. Seven orders followed the standard language from Rule 83.10, which allowed any party or the court to make such designations. But thirty-three orders deviated from the presumptive order and narrowed the parties that could label discovery confidential to only the defendants or the court.
Limiting the discretion to label discovery as confidential to only NYPD officers and City defendants gives these defendants discretion, not shared by the plaintiffs, to determine which types of documents must be kept secret. Only Magistrate Judge Go attempted to limit the broad power to make documents confidential in Linton v. City of New York by allowing parties to designate future items as confidential “only if the document is entitled to confidential treatment under applicable legal principles.” But even this limitation still permits parties to define the scope of confidentiality through their interpretation of “applicable legal principles.”
Table 2 summarizes each of the categories of information these orders made confidential.
Table 2: Documents and Information Protected
|Document Category||Number of Orders||Percent of Total|
|Officer personnel and disciplinary records||45||93.8%|
|Officer’s home numbers, address social security numbers, birthdates, financial information, tax records||6||12.5%|
|Officer’s medical records||7||14.6%|
|NYPD and Agency Protections|
|NYPD and CCRB investigative files||44||91.7%|
|NYPD training materials||8||16.7%|
|Internal Affairs files||8||16.7%|
|Video taken by NYPD||3||6.3%|
|Related to Monell discovery||2||4.2%|
|Plaintiff’s medical or psychotherapy records||10||20.8%|
|Plaintiff’s arrest documents||4||8.3%|
|Any Documents Designated Confidential|
|–By Defendants or Court||33||68.8%|
|–By Either Party or Court||7||14.6%|
Nearly every order specified how parties could use information and discovery covered by the protective order. Thirteen orders followed the text of Rule 83.10’s presumptive protective order, which limited either party’s use of the protected discovery to the evaluation, preparation, presentation or settlement related to the action. But many more orders narrowed the permissible uses or the parties limited by the order. Fourteen orders stated that only plaintiffs and plaintiffs’ attorneys were bound by the use restrictions, although the permitted uses were again for evaluation, preparation, presentation, or settlement activities. Further, fourteen orders limited plaintiffs’ and plaintiffs’ attorneys’ permissible use of covered discovery to only the preparation and presentation of the case  , and just one order limited both defendant’s and plaintiff’s attorneys to preparation and presentation uses. Two orders imposed a more general bar on a plaintiff’s or plaintiff’s attorney’s ability to disclose confidential material. The Linton and Fedd v. City of New York orders, entered by Magistrate Judge Go, were comparatively unique. Instead, these orders informed parties that it could not “be construed as conferring blanket protection on all disclosures or responses to discovery.”
Several orders also included procedures to ensure that once litigation had concluded, neither party could use the protected discovery in the future. Nineteen orders required the plaintiff’s attorney to “destroy” or “return” to the defendants’ attorney all confidential material, including copies and notes, within thirty and sixty days of the case ending. Rule 83.10 includes no such destruction provision. In Windley v. City of New York, Judge Scanlon modified this destruction requirement by hand to clarify that the court would retain a copy of any documents filed with the court regardless of whether the protective order covered them.
Table 3 describes the use limitations included in these protective orders.
Table 3: Use Restriction in Protective Orders
|Use Restrictions||Number of Orders|
|Use limited to evaluation preparation, presentation, or settlement|
|–for all parties||13||29.2%|
|–for plaintiffs only||14||29.2%|
|Use limited to preparation and presentation|
|–for all parties||1||0.02%|
|–for plaintiffs only||14||29.2%|
|General bar on disclosure||2||4.2%|
|Destruction or return of material after conclusion of litigation||19||39.6%|
Many of the orders included language that described the court’s powers to enforce and modify the orders. Twenty-two orders mimicked the language of Rule 83.10 on two points: the orders informed parties that the court both (1) retained jurisdiction to “enforce” obligations created under the order and “impose sanctions” for violating the order and (2) reserved its right to modify the orders in its discretion at any time. But other orders deviated from this baseline. One order retained only a reference to the court’s right to modify, and six only retained references to the court’s jurisdiction to enforce and issue sanctions. Sixteen orders did not discuss the court’s powers at all. In one stipulated protective order that failed to address the court’s powers, Judge Scanlon in Windley wrote that “the Court may modify this order at any time.” Judge Scanlon also hand modified the scope of the order, striking out language that indicated that the “terms” of the order would be “binding upon . . . future parties.” The orders in Linton and Fedd also reminded parties that the court would “revisit . . . this protective order . . . in light of the right of the public to inspect judicial documents.” Removing references to the court’s powers to enforce and modify protective orders does not eliminate them. But it could reveal how the parties wish to frame the agreement either as more or less susceptible to change or enforcement.
Table 4 summarizes how protective orders addressed the court’s powers.
Table 4: References to Court’s Powers
|Court’s Powers||Number of Orders||Percent of Total|
|Enforce and Modify||22||45.8%|
The similarity in content and structure between all the orders tends to reflect the presumptive protective order issued under Rule 83.10 in S.D.N.Y. cases. But the modifications to the presumptive order show how important the City attorneys are in the protective order process and may reveal officers’ and the City’s priorities. Overall, protective orders tend to be more protective of individual officer’s files and NYPD records, while simultaneously being less protective of plaintiffs’ records than the Rule 83.10 order. The orders in my sample also tend to restrict only plaintiffs’ use of protected documents and give more discretion to officer and City defendants to label documents not already covered as confidential.
This document analysis shows that nearly every protective order in the sample protects police and NYPD records from public disclosure and attempts to keep these materials confidential indefinitely. It appears that these defendants value keeping NYPD employment, disciplinary, training, and policy materials confidential. The next section evaluates how the demonstrated preference for confidentiality in cases with protective orders may result in differences in settlement amount compared with cases that lack such confidentiality protections.
I conducted an independent samples t-test to test whether the difference between settlement values in cases with and without protective orders is statistically significant. My two-tailed independent samples t-test compares the difference in means of the settlement price for cases with and without protective orders. The test determines if the difference in the mean values is significantly different than zero. A finding that the settlement prices are statistically significantly different at α=0.05 would mean that cases with protective orders can be distinguished from those without by considering the settlement amount.
Considering the results of my t-test, I found that the 139 cases with protective orders (logged mean=8.3, standard deviation=5.5) compared to the 456 cases without protective orders (logged mean=6.0, standard deviation=6.3) settled for significantly different, and higher, settlement amounts, t(593)=3.9, p=0.0001. The p-value of 0.0001 means that if there is actually no difference in the mean settlement amounts between cases with and without protective orders, my observed result would be obtained in only 0.01% of analyses. Therefore, cases with protective orders are likely more expensive for NYPD Officers and NYC and more lucrative for plaintiffs.
Figure 1 describes the test.
Figure 1: Two-Sample t-Test for Cases with and Without Protective Orders on Settlement Amount
|Group||Observations||Logged Mean Settlement Amount||Standard Error||Standard Deviation||95% Confidence Interval|
|Cases Without Protective Orders||456||6.016||.297||6.329||5.433||6.510|
|Cases with Protective Orders||139||8.325||.478||5.515||7.400||9.250|
|t-Statistic||| 3.675 ||
|Degrees of Freedom||593|
The results from both textual analysis of the protective orders and the t-test comparison of settlement amounts suggest that officers act strategically in moving for protective orders. There are several possible explanations for my findings. One inference the analyses may support is that when plaintiffs have a case against an officer with a potentially damaging personnel file, the City and its officers are willing to settle for a higher amount. Additionally, another explanation is that officers could request a protective order because they want to protect their privacy or prevent their past conduct from being linked to the current action even when they do not have frequent or dangerous misconduct. Perhaps City attorneys will move for a protective order whenever a plaintiff requests an officer’s employment records, suggesting that plaintiff’s attorneys influence the officers’ litigation behavior. Regardless of whether the misconduct records have damaging contents or not, the significant differences in mean settlement amount between the two types of cases suggest the NYPD strongly values protecting the confidentiality of records and potentially being willing to pay more to do so.
Alternative explanations that explain the higher settlement amounts in cases with protective orders without considering confidentiality rationales also have support in the data. Cases with stipulated protective orders may have egregious facts, extensive injuries, or more aggressive or experienced plaintiff’s attorneys that serve as the primary drivers of higher settlement amounts. The following considers each alternative explanation.
Merits of the Case: While the data does not provide details on the case facts, the comparative numbers of cases in each group that settled for $0 could serve as a proxy on the merits of the plaintiff’s case. Comparing the two, twenty of the 139 cases with protective orders (14.4 percent) settled for $0, while a more substantial 116 of the 456 cases without protective orders (25.4 percent) settled for $0. After assigning a value of one to cases with $0 dispositions and a value of zero to cases with higher settlement amounts, I compared the difference in means between cases with and without protective orders on this variable using a two-tailed independent-means test. The test shows that this difference is statistically different from zero with t(593)=2.7 and p=.007. Figure 2 describes the test.
Figure 2: Two-Sample t-Test for Cases with and Without Protective Orders on $0 Dispositions
|Group||Observations||Mean $0 Dispositions||Standard Error||Standard Deviation||95% Confidence Interval|
|Cases Without Protective Orders||456||.254||.020||.436||.214||.295|
|Cases with Protective Orders||139||.144||.030||.352||.085||.203|
|t-Statistic||| 2.729 ||
|Degrees of Freedom||593|
The fact that cases without protective orders are more likely to settle for $0 does not cut against the inference that valuing record confidentiality influences officers’ willingness to settle for higher amounts. These findings also support the theory that the City and NYPD act strategically to protect confidentiality. If a $0 disposition signals a weak case, this finding could suggest that NYPD officers may care more about keeping their records confidential when they are defendants in a case that credibly charges them with misconduct. It could also signal that officers with past misconduct histories who need protection are more likely to face future litigation that have a higher chance of success on the merits. Together, the statistical findings that cases with protective orders settle for higher amounts than those without and are less likely to be settled for $0 correlate with the possibility that NYPD officers seek to keep police misconduct records confidential when a plaintiff’s case is strong.
Attorney Experience: A case may result in a higher settlement amount because the party benefits from an aggressive or experienced plaintiff’s attorney. I measured attorney experience three ways. I considered repeat law firms representing police officers, repeat law firms representing plaintiffs, and cases for which the plaintiff represented themselves pro se. I considered police officers and plaintiffs who are represented by repeat firms to benefit from the firm’s presumptive institutional knowledge. Comparatively, I assumed that pro se litigants are less sophisticated litigants.
Overwhelmingly, the City of New York represented the police officers in nearly every case, regardless of whether a court entered a protective order. Non-City attorneys represented only one officer in a case without a protective order and only one officer in a case with a protective order. In both these cases, the City of New York was still named as a defendant and may have influenced the officer’s representation.
In contrast, there were few repeat plaintiffs’ firms in section 1983 suits against police regardless of the case’s protective order status. Only six plaintiffs in cases with a protective order were represented by a firm that had represented other plaintiffs in any suit across the full analysis sample, accounting for 4.3 percent of the 139 cases with protective orders. And only twenty-nine plaintiffs in the 456 cases without protective orders were represented by a repeat plaintiff’s firm that represented another plaintiff in the full analysis sample, accounting for 6.4 percent of cases without a protective order. To test whether this represented a significant difference, I assigned a one to all cases with protective orders headed by a repeat plaintiff’s firm and a zero to cases led by a new plaintiff’s firm. After conducting an independent samples t-test, I found no statistically significant difference in the frequency of repeat plaintiff’s firms in cases with and without protective orders, with t(593)=0.9 and p=.36.
Finally, only seven of the fifty-four pro se litigants in the sample entered into protective orders. This represents 5.0 percent of the total cases with protective orders. In contrast, pro se litigants were plaintiffs in forty-seven cases without protective orders, a rate of 10.3 percent, that corresponds closely to the full dataset rate of 10.4 percent for all 2,929 cases. After assigning a one to all cases with protective orders headed by a pro se plaintiff and a zero to pro se cases without protective orders, I conducted an independent samples t-test. But there was no statistically significant difference in the rates of pro se plaintiffs in cases with protective orders compared to those without, with t(593)=1.9 and p=.06.
The consistent City representation of police officers suggests that these attorneys have substantial institutional knowledge while the fairly limited numbers of repeat plaintiff’s attorneys cuts against the theory that differential experience drives the disparate settlement amounts in cases with and without protective orders. Instead, it suggests that City attorneys will move for protective orders based on their knowledge of whether police records, misconduct, or otherwise, may be discoverable. Significantly, across all cases with protective orders, the City of New York moved for the protective order in 116 cases, 95 percent of all cases for which information about who was the moving party was available. This inference is further supported by the comparatively low rate of pro se litigants in cases with protective orders compared to those without, although the difference is not statistically significant. Because pro se plaintiffs are typically less sophisticated, the NYPD defendants may see them as less likely to discover or disseminate information that they would typically shield with a protective order.
Table 3 summarizes the potential mechanisms discussed above and provides descriptive statistics comparing cases with and without protective orders on these key variables.
Table 3: Descriptive Statistics of Cases with Stipulated Protective Orders and Cases Without
|Protective Order||No Protective Order||p-Value|
|Mean Settlement Amount (Log)||$347,335.90 (8.3)||$63,909.02 (6.0)||.0001*|
|Frequency if Settlement Amount Is $0||20||(14.4%)||116||(25.4%)||.007*|
|Repeat Plaintiff’s Firms||6||(4.3%)||29||(6.4%)||.36|
|Frequency of Pro Se Litigants||7||(5.0%)||47||(10.3%)||.06|
The selected sample and analysis performed make the generalizability of these findings limited. This analysis cannot describe how litigants use protective orders in suits against police departments in other fora in separate time periods because the sample includes only federal cases filed against NYPD officers in 2014 and 2018. And because the available data cannot describe important cases and litigant attributes, as discussed in Part III.A, the 20 percent analysis sample may not adequately represent the full universe of the 2,929 cases filed against the NYPD in the sample period. Relying on these data necessarily flattens the actual litigation process to known and recorded variables.
Further, the statistical analyses do not permit causal inferences about the mechanisms driving the differential settlement amounts between cases with and without protective orders. At most, these findings suggest a correlation between materials that officers seek to protect, like misconduct records, NYPD training materials, and higher settlement amounts.
But this analysis does provide important descriptive findings that detail a previously unstudied phenomenon of the use of protective orders in section 1983 suits against police. These initial analyses provide fertile ground for further, more comprehensive studies of how secret discovery may inhibit the flow of information about police officers and police misconduct to the public.
The qualitative findings and descriptive statistics of the analysis sample show that despite Judge Weinstein’s warning that routine protective orders issued in section 1983 suits against police “will not necessarily promote justice or the proper balance of interests,” protective orders are common in suits against NYPD officers. Courts in the Eastern and Southern District of New York regularly grant them. The City, on behalf of the officers, is generally the moving party. But most protective orders are stipulated. These orders explicitly make information like police personnel and misconduct records confidential and bind the plaintiffs and their attorneys to use the material only for the purposes of the litigation or settlement. There is no indication that the courts consider the public’s interest. For those cases covered by protective orders, plaintiffs cannot use the discovered misconduct records or make them public, even if they implicate public health and safety.
These findings, combined with the result that cases with protective orders covering police disciplinary records settle for statistically significantly higher amounts than those without, correlate with a theory that the NYPD and the City of New York value keeping police officers’ records confidential and will settle for higher amounts in cases where plaintiffs discover these records.
The data show that New York City and NYPD officers are the movants for stipulated protective orders in nearly every case. NYPD officers can therefore exploit and benefit from perfunctory grants of stipulated protective orders, skew incentives between the police and civil rights litigants, and procedural mechanisms that make future disclosure of police records difficult.
A. Case Law Development Reflects Trend Toward De Facto Acceptance of Protective Orders in Suits Against Police
The shift seems stark between Judge Weinstein’s proposal for a strict good cause analysis in suits against the police and the present default to granting stipulated protective orders without much scrutiny. But tracing the history of confidentiality protections in suits against police shows that stipulated protective orders are the natural outgrowth of the King v. Conde framework. In King, Judge Weinstein concluded that “routinely” entering protective orders covering police records would not adequately serve the public interest. But in the same case, the court determined that state law privileges covering law enforcement records “may be a useful referent” in federal court but are not “binding.” This finding created a tension between the harms caused by an outright ban on both plaintiffs’ and the public’s access to law enforcement records due to privilege and the legitimate interests that the privilege sought to protect. The following describes how the case law has trended away from a case-specific weighing of the interests toward widespread acceptance of protective orders in the wake of these tensions.
For years, police have been able to protect personnel and misconduct records from public disclosure through statutes like New York Civil Rights Law § 50-a and California Penal Code section 832.7. The legislature has recently amended these statutes to provide greater transparency as discussed in Part V.A, but they helped shape the current protective order status quo in section 1983 suits against police. In the 1980s, when civil rights litigants sued police, police invoked these statutes and the concept of an official records privilege to keep litigants from gaining access to these materials in discovery.
This specific privilege-based argument lost traction after federal cases like King and Kelly v. City of San Jose rejected the idea that a state law privilege could control federal cases. Citing public interest in accessing information about police, both courts adopted a balancing test “pre-weighted in favor of disclosure.”
Although these rulings favor access to police misconduct records, they laid the groundwork for the current status quo of de facto protective orders for police records in discovery. Kelly noted that protective orders could solve the discovery problem, “emphasiz[ing] that in many situations what would pose the threat to law enforcement interests is disclosure to the public generally, not simply to an individual litigant.” King concurred, finding that while routine protective orders are not desirable, they may be “an effective way to permit discovery without undermining law enforcement.”
Following King and Kelly, courts adopted protective orders as a solution to the tension between civil rights litigant’s access needs and police’s privacy needs. Instead of requiring narrowly tailored protective orders, courts granted protective orders that labeled the entire category of personnel and disciplinary records as confidential. They cited officer privacy as the grounds for good cause. But courts have gone further by requiring both police and civil rights litigants to stipulate to protective orders. For instance, in Megargee v. Wittman, the court advised the parties to resolve their discovery dispute around officer personnel and training files with a stipulated protective order. It advised the parties to stipulate primarily because reviewing the contested files would be a drain on judicial resources. In other cases, courts granted protective orders not because of concerns with protecting police privacy but because of an unwillingness to give plaintiffs the power to “disseminate unfiled police officer disciplinary information . . . without restriction or limitation.” Ultimately, courts are deferential in finding good cause to protect police privacy, even though police are public employees.
B. All Parties to Section 1983 Suits Have Strong Incentives to Request and Assent to Protective Orders
Given that most courts treat general allusions to police privacy as enough to satisfy the good cause analysis, civil rights plaintiffs have little incentive to oppose defendant police officers’ proposed protective orders. Civil rights plaintiffs focus on settlement because they are “one-shot” players “who have only occasional recourse to the courts” rather than repeat players “who are engaged in many similar litigations over time.” Therefore, they do not expect to routinely challenge police misconduct, and they have “little interest in the state of the law.” Plaintiff’s firms that engage in repeat section 1983 litigation against police do have an interest in favorable precedents. But because they are bound by their clients’ goals, many plaintiff’s firms must behave as one-shot players. And the fact that in the S.D.N.Y., certain police misconduct records are presumptively protected from disclosure under Local Rule 83.10 may disincentivize experienced firms from contesting protective orders covering police records because they believe the court will almost certainly grant the orders. Experts who regularly consult with plaintiffs in civil rights cases against police must also behave as one-shot players despite their expertise. Once the client chooses to enter into a stipulated protective order, any analysis the expert does with confidential discovery materials must remain confidential. These cases present the very circumstance that Professor Arthur Miller claimed was rare: civil rights litigants “who could reveal the dangers, often ha[ve] litigation interests other than obviating a potential risk to the public—most commonly, securing an advantageous settlement—and [they] will use acquiescence to the entry of the protective order as a bargaining chip.”
Additionally, the presence of stipulated protective orders in section 1983 suits involving repeat plaintiffs’ firms enables police to prevent these firms from directly using the institutional knowledge about officer police misconduct in future cases. Nearly all the protective orders I reviewed prevent future “use” beyond the scope of the litigation, and some required plaintiffs’ attorneys to destroy or return protected materials. Thus, plaintiffs’ firms cannot build a database of problem officers or collect misconduct and training materials that show systemic departmental failures without risking sanction.
The police department and the associated city or municipality, as repeat players, have “low stakes in the outcome of any one case, and . . . ha[ve] the resources to pursue [their] long-run interests.” This Note’s statistical findings are consistent with a conclusion that a central interest for NYPD officers and their City attorneys defending civil rights suits is keeping police records confidential.
The NYPD’s preference for confidentiality is unsurprising. Police across the United States regularly advances three main arguments for why their misconduct records should remain secret: concerns about privacy, concerns about the public’s ability to interpret records, and concerns about harassment of officers.
First considering police officers’ personal privacy arguments, there is no doubt that police, like other individuals, are entitled to keep the personal details about their life, including addresses and social security numbers, secret to protect their safety. Indeed, “[p]olice use their considerable bargaining and lobbying power to ensure their personal privacy.”
But in practice, personal privacy appears to serve more as a rhetorical justification for record confidentiality than as the true motivator. My review of cases against the NYPD shows, protective orders rarely cover this type of information explicitly; instead, the orders focus on protecting misconduct and personnel files.
Moreover, police argue that privacy rights granted to civilians should be more protective of police officers and their misconduct records because of officers’ “responsibility to testify in court and [their] vulnerability to public smear campaigns by private citizens.” But scholars have argued that police do not have the same claims to privacy rights that private citizens do when it comes to employment-related misconduct records. Rachel Moran explained that privacy theory does not “support a constitutional right to privacy in most police misconduct records.” This is because “[m]ost police misconduct records do not fall within . . . categories of highly personal, intimate, or political information” like medical, financial, or voting records. As Moran found, the Tenth Circuit, a Missouri appellate court , and the Louisiana Court of Appeals have all distinguished between records of substantiated police misconduct and the type of private information protected by privacy laws. Further, police officers’ status as public officers weakens their claims to privacy protections that private citizens have. Their public role makes their on-duty misconduct “matters with which the public has a right to concern itself,” and these matters should not be considered “private, intimate, personal details of the officer’s life.”
Second, police officers worry that because discipline and training may vary between and within departments, the public is ill-equipped to interpret the significance of a police personnel file’s contents. Kate Levine argued that:
“It is not difficult to draw an inference . . . that an officer’s disciplinary file may have as much to do with his supervisor’s attitude and biases as it does with his adherence to the code—let alone his quality as an officer, as defined by those of us who are policed, rather than those who are doing the disciplining.”
These distortions, according to Levine, make police records such a reflection of discretion, by both civilians and police personnel, that they become virtually useless to the public in interpreting office competency.
While discretionary discipline may misrepresent an officer’s history to some degree, Levine’s preference for confidentiality goes too far. It may be difficult to test the degree to which office discipline reflects supervisor bias rather than actionable misconduct. But recent empirical work by Kyle Rozema and Max Schanzenbach undermines claims that civilian complaints are poor indicators of misconduct, finding that high volumes of citizen misconduct complaints correlate with higher settlement values in section 1983 suits. Further, the possibility that supervisor bias taints disciplinary records provides further reason for the public to examine them. The way to address and excise discriminatory treatment by supervisors within departments is to make that information public, not shield it. Ultimately, record transparency is a necessary first step for the public to learn about and understand how police disciplinary procedures operate and begin to build capacity to differentiate between discipline and training procedures. Certainly, context is critical to interpreting police disciplinary records, but, as Erik Luna concluded, the public cannot begin to “assess the performance of its [public officers] without knowing what decisions were made, what actions were taken, and the factual basis for both.”
Third, police express concern that if misconduct records are subject to public dissemination, disgruntled citizens will target police for physical and emotional harm. However, recent research suggests that these concerns are largely unsubstantiated. Rachel Moran and Jessica Hodge’s work surveying officers in states with public disclosure of police records found that only 16 percent of police department administrators “believed officers in their department had been harmed by public access to records of officer misconduct.” They found only a single administrator who recounted that an officer was subject to “physical and verbal harassment” due to public access to records. But administrators also recounted incidents where “media misinformation” and “very biased news reporting” did result in firings of officers that the administrators believed was unjustified, and officers reporting to 6 percent of administrators surveyed did receive threats, often on social media.
Regardless of their merits, these demonstrated police interests in halting the spread of information about their personnel records in litigation incentivize police to exploit the power of stipulated protective orders. Seeking a plaintiff’s consent and stipulation to a protective order, instead of filing a motion for a protective order without the plaintiff’s agreement, allows police to lessen the possibility that courts will find good cause lacking. A one-sided motion for a protective order could signal to the court that a good cause analysis is appropriate. Therefore, an officer is induced to obtain a plaintiff’s consent to protect police records because a court has little incentive to consider the need for disclosure limitations when both parties agree. As soon as parties begin to file discovery materials, the court may treat those items as “judicial documents.” Judicial documents are presumptively accessible to the public under the common law right of access. The common law right of access is rebuttable. But allowing a case to travel too far down the litigation path is risky because it makes police records vulnerable to public access. Thus, as the cases in this sample show, police do tend to settle without filing police records on the public docket.
Moreover, the City attorneys who represent NYPD officers also have an interest in keeping police records confidential. As Joanna Schwartz found, officers rarely contribute money to settlements arising from their misconduct. In New York specifically, NYPD officers “were required to contribute to just .49% of the civil rights cases in which plaintiffs received payment” between 2006 and 2011. Instead, cities and towns are responsible for paying settlement awards. As the party responsible for paying the settlement, the municipal governments are interested in keeping settlement payouts low. As a result, they are interested in preventing plaintiffs’ lawyers from publicizing information about officers with histories of misconduct to discourage future suits targeting those officers as known bad actors. While the municipalities also theoretically have an interest in reducing future lawsuits by reforming police department practices and terminating misconduct-prone officers, “pressure by police unions . . . [and] an interest in appearing tough on crime” may all combine to incentivize municipal governments to absorb the cost of police misconduct. This phenomenon likely extends beyond New York City. Many states have statutory requirements that municipalities represent officers in civil suits. Some police union contracts include conditions that government attorneys represent their officers.
In summary, the interests of plaintiffs, NYPD officers, and City attorneys align to enable police to protect their privacy interests and keep misconduct records confidential.
C. High Bars to Modify Protective Orders Prevent Litigants from Disclosing Police Misconduct After Receiving Discovery
Because civil rights litigants do not know what the police records will contain before stipulating to a protective order, they remain bound even if they find that the records have information relevant to public health and safety. For example, in the case giving rise to Mullally v. City of Los Angeles, the plaintiffs received seventy-nine Los Angeles Police Department (LAPD) records that they believed would show that the LAPD willfully ignored its officers’ patterns of domestic violence. The parties entered into a stipulated protective order to protect these files, and, after receiving them, the case soon settled. Mullally, an expert retained to review the files, disclosed their summaries to a television reporter following the settlement. After the broadcast, the court held Mullally in contempt for violating the protective order. The Ninth Circuit in Mullally found that the concerned expert had other options; he did not need to violate the protective order outright because it was subject to the court’s “modification or review.”
While there is little doubt that the public has no access rights to pretrial discovery materials, the court has the unilateral power to consider the public interest when modifying a protective order. As the Supreme Court noted in Seattle Times, “[w]hether or not the Rule itself authorizes [a particular protective order] . . . we have no question as to the court’s jurisdiction . . . under the inherent equitable powers of courts of law over their own process, to prevent abuses, oppression, and injustices.” But despite a court’s inherent power to lift protective orders, the current case law impedes parties from successfully modifying protective orders. The showing a movant must make to convince the court to modify a protective order is steep. A concerned litigant, expert, or intervenor is unlikely to meet the standard under current precedent in most federal circuits.
For example, the Second Circuit will modify a protective order only with a “showing of improvidence in the grant . . . or some extraordinary circumstances or compelling need.” While other courts will modify on a lesser showing, litigants still need to show that the opposing party did not rely on the confidentiality order to agree to settle the case. In the Fourth, Fifth, Sixth, Ninth, and Eleventh Circuits, litigants must overcome reliance-based modification criteria. Reliance even remains a factor in the Third Circuit’s more expansive balancing test. Among the other factors articulated in the Third Circuit’s influential test, reliance “depend[s] on the extent to which the order induced the party to allow discovery or settle the case.” In the Eighth Circuit, litigants must prove that there has been an “intervening circumstance” that changes the good cause calculus, which led the original court to enter a protective order. Under any of these standards, police can credibly argue that the confidentiality of police records induced discovery and settlement and that circumstances have not meaningfully changed since the protective order was entered. The threat of contempt and the difficult modification standards prevent plaintiffs from exposing discovery; therefore, the records stay confidential indefinitely after parties to a protective order settle or the case terminates.
A recent Eastern District of California case shows the difficulty plaintiffs have in proving the good cause standard needed to lift a protective order after obtaining police misconduct records in discovery. In Perkins v. City of Modesto, Modesto officers shot Plaintiff Perkins in his car unprovoked. Perkins sued the officers, the department, and the City and obtained the officers’ misconduct, personnel, and Internal Affairs records pursuant to a stipulated protective order. After Perkins moved to modify the protective order, the court lifted confidentiality for the fifteen records that met one of the conditions for public disclosure set forth in California Penal Code section 832.7. But the court read section 832.7 narrowly and refused to permit public disclosure of records that seemingly described use-of-force incidents because they did not result “in death or ‘great bodily injury.’” The court applied the balancing test set forth in In re Roman Catholic Archbishop of Portland in Oregon, and denied the motion. It specifically found that disclosing the “files will violate privacy interests and lead to potential embarrassment of the officers and reporting parties.” And in weighing public and private interests, the court determined that while the officers are public officials, and the files “might be important to public health and safety,” the officers’ privacy interests weighed more heavily in favor of confidentiality.
Under current legal standards and incentives for civil rights litigants, police can respond to civil rights litigants’ discovery requests without fear that their personnel, misconduct, or training records will ever become public. By granting stipulated protective orders in cases that settle at high rates and applying strict standards for modification, courts defer to police interests. This gives police officers virtually unchecked ability to control what information the public knows about them.
Litigants and experts in section 1983 suits against the NYPD are often bound by protective orders. They cannot expose information about dangerous officers, harmful department policies, or practices that they learn in discovery to the public. And the public has no access rights to discovery. Therefore, the public’s best hope for preventing officers and their departments from willfully concealing misconduct records is legislation that makes this material public. But transparency alone will not address the problems that the American public was forced to confront following 2020’s worldwide Black Lives Matter protests. This Part describes current state-level transparency efforts, the need for a national police misconduct database, and the ways in which a national database may provide a springboard for necessary systemic reform.
State legislatures have begun to realize the threats to public health and safety that strict statutes protecting police officer misconduct records pose. Specifically, California and New York state legislatures have lifted restrictions on access to police records. This trend should continue. These states’ recent legislation to make police records more transparent are discussed in turn.
Before 2019, California so limited both litigant and public access to police records that state legislators characterized California as “one of the most secretive states in the nation in terms of openness when it comes to officer misconduct and uses of force.” Under the prior version of California Penal Code section 832.7 and associated statutes, law enforcement records were unavailable to the public under California’s Public Records Act, and the penal code even “restrict[ed] a prosecutor’s ability to learn of and disclose” certain information about officers.
The California legislature enacted Senate Bill (SB) 1421 in 2018 to address the opaqueness of police misconduct records, explaining, “[t]he public has a right to know all about serious police misconduct, as well as about officer-involved shootings and other serious uses of force.” The legislature determined that concealing this information about law enforcement “undercut the public’s faith in the legitimacy of law enforcement” and “endanger[ed] public safety,” among other concerns. Movement leaders like Black Lives Matter and the Anti Police-Terror Project, both Black-led organizations that fight against police violence in communities of color, sponsored the bill along with the American Civil Liberties Union (ACLU) of California, California Faculty Association, California News Publishers Association, and Youth Justice Coalition. But official publicity for the bill and reporting focused on police violence generally and the need to update outdated transparency laws without directly acknowledging the disparate impact of police violence on people of color, and African Americans in particular.
Under the amended statute, “personnel records of peace officers” are still generally “confidential.” But the state provides an exception for incidents “involving the discharge of a firearm . . . by a peace officer”; involving “the use of force by a peace officer . . . result[ing] in death, or in great bodily injury”; records confirming a peace officer committed sexual assault; and sustained findings of dishonesty, including perjury. In a case reviewing the newly enacted statute, California’s Court of Appeal for the First District explained that the law “acknowledg[ed] . . . the extraordinary authority vested in peace officers and the serious harms occasioned by misuse of that authority.”
This law has enabled news organizations and advocacy groups to obtain California law enforcement records over the objections of the California Department of Justice. Becerra v. Superior Court describes the records requests submitted under the new law. In that case, the First Amendment Coalition requested “all records” held by the California Department of Justice that met section 832.7’s public disclosure criteria, and KQED Radio requested records from 2014 to 2018. The Department refused to provide records that it obtained from “other state and local law enforcement agencies,” even if they fell into one of the new law’s categories explicitly permitting disclosure. The Department instructed both requesters to seek the records from the state and local departments directly. The court found that the Department must produce the requested records after it interpreted section 832.7 in light of the California Public Records Act (CPRA) and found the Department’s claims of undue burden unavailing.
But the limits of California’s amended law have already begun to show. The court in Perkins v. Modesto cited the limited and defined scope of section 832.7 as justification to deny a litigant’s request to lift a protective order keeping personnel records and Internal Affairs records confidential even though the California law did not control the federal court’s authority to modify the order. Particularly, the court acknowledged that this would mean officer files involving use of force incidents that “might be important to public health and safety,” remain secret in California.
Because police departments are the responsive parties to these CPRA requests, they maintain control of the files and have the discretion to narrowly construe terms like “great bodily injury.” Consequently, police departments can continue to deprive the public of materials the legislature intended them to have. The San Francisco Police Department, for example, selectively “released only partial documentations from four shootings by officers” but refused to release disciplinary records in the first six months that the new law was in effect. The San Jose Police Department produced only six full files and ten partial files out of the eighty-five that the Mercury News alone had requested over the first 1.5 years of the new regime. And as of June 2020, the City of Richmond had simply refused to “release [police] documents pertaining to sexual assault, dishonesty and use of force.”
The limited nature of these disclosures is not rare. According to a report released six months after SB 1421 amended section 832.7, “some of the state’s largest law enforcement agencies [had not] provided a single record.” Before the law went into effect, some police departments, including those in Downey, Inglewood, Freemont, and Morgan Hill, and the Yuba County Sheriff destroyed years of records before the public could request them.
California legislators have begun to recognize the significant transparency limits of amended section 832.7 and now explicitly acknowledge the disparate violence police have perpetrated on communities of color. On July 1, 2020, State Senator Nancy Skinner introduced SB 776 to “further increase[e] public access to long-hidden police records,” a decision explicitly driven by the “worldwide protests over racism and brutality in policing.” The bill would have closed police department’s discretion to resist disclosure by expanding “access to all records involving police use of force”; records of “police dishonesty . . . and on-the-job sexual assault” not just sustained complaints; and “all disciplinary records involving officers who have engaged in racist, homophobic or anti-Semitic behavior.” But the bill failed to pass after it was ordered inactive. For now, California law enforcement will retain the discretion to delay or deny access to police records.
Since 1976, New York’s highly restrictive Civil Rights Law section 50-a completely blocked the public’s access and even sharply limited criminal defendants’ ability to obtain police misconduct records to impeach officer testimony or prove their case. The New York Civil Liberties Union, an organization that has a history of being denied requests for police records under 50-a, described it as “arguably the worst law in the nation when it comes to the public’s ability” to access misconduct records. The law’s terms were so stringent that the New York Court of Appeals interpreted it as barring public disclosure of the outcomes and recommendations from disciplinary proceedings referred to the NYPD from the City’s Civilian Complaint Review Board. This is especially shocking, as the Board was created to establish better transparency over police actions. New York’s State Committee on Open Government found in 2014 that 50-a “has been expanded in the courts to allow police departments to withhold from the public virtually any record that contains any information that could conceivably be used to evaluate the performance of a police officer.”
On June 12, 2020, New York repealed 50-a. The effort to repeal 50-a was an eight-year battle led by Communities United for Police Reform, an organization dedicated to ending “[d]iscriminatory and abusive policing in New York City,” and Justice Committee, “a grassroots organization dedicated to building a movement against police violence and systemic racism in New York City.” These organizations’ arguments that 50-a’s confidentiality shields perpetuated racist policing did not take hold until George Floyd’s killing and the Black Lives Matter protests demanded action that the legislature amended section 50-a.
Pursuant to SB 8496 and Assembly Bill (AB) A10611, “law enforcement disciplinary records” including “complaints, allegations, and charges against an employee” along with officer names and the transcripts and outcomes of disciplinary proceedings are available to the public under New York’s Freedom of Information Law. Its only limit provides that law enforcement may redact medical history, home address and personal contact information, social security numbers, and mental health and substance abuse services used. In the justification for the repeal, the New York Senate found that “[r]epeal of § 50-a will help the public regain trust that law enforcement officers and agencies may be held accountable for misconduct.” While these were the official justifications, the influence of the Black Lives Matter movement, Communities United for Police Reform, and Justice Committee on the bill makes clear that organizing against racially discriminatory policing played a major role in getting 50-a amended. State Senator Zellnor Myrie made this link explicit when he stated in his vote to repeal the law that “[m]y life matters. Black lives matter.”
The new disclosure scheme in New York has not yet been tested, but law enforcement discretion may continue to limit disclosure despite New York’s broad disclosure mandate. Florida, Georgia, and Arizona, which enacted similar public disclosure laws, have demonstrated the continuing challenges citizens face in obtaining these records; officers in these jurisdictions continue to create delays in disclosing records until investigations are finalized and then may argue privacy grounds justify further secrecy. Even if New York is able to escape the fate of these other jurisdictions, police union contracts could allow New York police departments to destroy these misconduct records if they are not requested expediently.
The state-level trend toward transparency is a positive development and more states should join California and New York in expanding public access to police misconduct records. But there have been local trends toward transparency as well. Some cities and departments, notably in major metropolitan areas with large Black American populations, have created anonymized datasets on complaints, use of force, and officer-involved shootings. Currently, there are twenty-six departments that provide data on officer-involved shootings, eighteen departments that provide data on use of force, and six departments that provide data on complaints. None of these departments provide public access to the records themselves, the names of the officers involved, or the officers’ misconduct histories. This makes these anonymized datasets a poor substitute for public access to detailed misconduct records.
Efforts to create misconduct databases that identify specific officers have faltered. In June 2020, New York City Mayor Bill de Blasio announced an intention to create a public database of disciplinary case documents shortly after the legislature repealed New York Civil Rights Law section 50-a. The database was to include “an officer’s name, the internal charges they faced, the hearing dates, transcripts from the hearings, and the final ruling of the department.” But police and firefighters unions sued the City to prevent these disclosures. The Second Circuit has stayed the publication of the records while the case is litigated.
Apart from New York City’s planned, but not yet realized, publication of disciplinary records, these reforms do not necessarily make police misconduct information generally accessible. These laws continue to place the burden on the public to seek this information and only allow individuals or organizations with knowledge of public records acts actual access. The patchwork of national public disclosure laws means that the public, and often police departments, are unable to identify “wandering officers” who “are fired or . . . resign under threat of termination and later find work in law enforcement elsewhere.” A comprehensive, national solution is needed.
To ensure the public has the access to records that provide them with concrete information about the law enforcement officers who patrol their communities, Congress should create a public, national database of police misconduct information.
There is a demonstrated public desire for such a database. News organizations have begun to develop public databases to fill the void created by state and federal policies. Both The Washington Post and The Guardian developed projects tracking national police killings of civilians based on news reports and public sources. The Guardian’s project covered only police use of deadly force incidents between 2015 and 2016 while The Washington Post’s project extends from 2015 to the present. Neither project includes the names of officers involved or describes an officer’s prior misconduct history.
But some news and advocacy organizations have begun local efforts to publish state-wide police misconduct records. KQED Radio, partnering with other California newsrooms, has begun a systematic effort to obtain available records under California Penal Code section 832.7’s public disclosure law. While the organization is “still fighting for records from agencies that have yet to provide them,” KQED “is developing a database of all the records” and plans to make them public. ProPublica published a database based on 12,056 records from investigations by the CCRB into complaints against NYPD officers. The searchable database lists the officers name, the conclusion of the CCRB, officer and complainant demographics, and high-level descriptions of the allegations. The New York Civil Liberties Union published a similar, searchable database of CCRB complaints containing “279,644 unique complaint records involving 48,757 active or former NYPD officers.” In addition to officer name, a description of the allegation, and the CCRB’s finding, it also lists the incident date and the penalty, if the NYPD imposed any.
But only one current project provides access to the misconduct records themselves. In 2019, USA Today and the Invisible Institute published an online database of more than 30,000 officers who have been decertified in forty-four states. Both the names of the officers and the decertification record itself are available. This significant and singular effort in terms of its scope and documents provides the first national database of detailed police records. While limited to forty-four states and only publishing records on decertification, the database provides an opportunity for the public to learn about the officers who police them and identify patterns of officer misconduct that can lead to targeted reform. But the narrow scope and incomplete nature of the database further highlights the difficulties even sophisticated organizations still face in obtaining police records.
Moreover, the federal government is aware that the “lack of accurate data” and a “lack of transparency about policies and practices in place governing use of force” have driven the public to believe that “police use of force in communities of color . . . is unchecked, unlawful, and unsafe.” Former Federal Bureau of Investigations (FBI) director James Comey at a Department of Justice (DOJ) Summit on Violent Crime Reduction on October 7, 2015 stated: “It is unacceptable that The Washington Post and the Guardian newspaper from the U.K. are becoming the lead source of information about violent encounters between police and civilians.” Underscoring Comey’s point, the U.S. Commission on Civil Rights in 2018 found that neither the public nor police departments have “accurate and comprehensive data regarding police uses of force” and that no “national database exists” to capture the frequency of incidents. The Commission explained that “[s]everal representatives from public advocacy groups, government agencies, local police departments, and experts on the topic” agreed “that the lack of national data on police use of force incidents serves as one of the most significant impediments to identifying problems and implementing solutions” around police misconduct. It recommended that Congress require states to report data on use of force to the DOJ and that the DOJ should “create and maintain a public, national database of police use of force incidents.”
Following these findings, there are more comprehensive national data collection efforts on police misconduct currently, but none of the data is public. Specifically, under the Death in Custody Reporting Act of 2013, states must submit data to the DOJ when “any person who is detained, under arrest, or is in the process of being arrested, is en route to be incarcerated, or is incarcerated” dies. These data must include the demographic information and name of the deceased, date, time and location information, the law enforcement agency involved, and a description of the circumstances of the death. Notably, states do not have to report the names of the officers involved. Additionally, the FBI also collects data on police use-of-force incidents that result in the death or serious bodily injury of a person as of January 1, 2019, under its Use of Force Program. But the FBI has not made any of these data available to the public. In response, President Trump signed Executive Order on Safe Policing for Safer Communities on June 16, 2020, directing the attorney general to “create a database . . . to track . . . terminations or de-certifications of law enforcement officers, criminal convictions of law enforcement officers for on-duty conduct, and civil judgments against law enforcement officers for improper use of force.” While these data will be made public, the data are “anonymized” and only provide information for officers who face the most severe outcomes.
Making a publicly available police misconduct database has remote chances of successfully getting past Republican senators, but the proposal has Congressional backers. As part of the George Floyd Justice in Policing Act of 2020, House Representatives have proposed that “the Attorney General . . . establish a National Police Misconduct Registry” that should contain: credible complaints, complaints pending review, unsubstantiated complaints, disciplinary records, termination records, certification records, and records of lawsuits against police and settlement amounts if applicable. The law would make “the Registry available to the public on the Attorney General’s website in a manner that allows members of the public to search for an individual law enforcement officer’s records of misconduct.” The House passed the Act on June 25, 2020. It is currently pending in the Senate, where Republican senators are expected to stall and prevent a vote on the bill.
Without concrete information about dangerous officers, the public is unable to effectively scrutinize the officers who patrol their streets and the departments that train and discipline them. Access through public records laws is improving, but they provide a woefully incomplete picture of police misconduct in the United States. Criminal prosecutions are too rare to provide the public access to records. Furthermore, the prevalence of protective orders in civil suits keeps this information secret. Without police misconduct records, the government, advocacy organizations, and communities themselves cannot hold officers and departments accountable. A public, national registry of police misconduct records would fill this dangerous information gap.
But transparency alone cannot address calls to reform, defund, or abolish police and cannot prevent police brutality. It is merely a tool, not a policy solution. Yet this tool could prove useful in pushing forward reforms Black Lives Matter protestors have demanded.
The Black Lives Matter protests have resulted in two competing campaigns for change: a “research-based” list of reforms “with the strongest evidence of effectiveness at reducing police violence,” exemplified by Campaign Zero’s 8 Can’t Wait; and an expansive list of measures to “build toward a society without police or prisons,” as exemplified by 8 to Abolish. 8 Can’t Wait encourages policymakers to implement the following eight proposals: (1) “ban chokeholds & strangleholds”; (2) “require de-escalation”; (3) “require warning before shooting”; (4) “require [police to] exhaust all alternatives before shooting”; (5) impose a “duty to intervene”; (6) “ban shooting at moving vehicles”; (7) “require use of force continuum”; and (8) “require comprehensive reporting.” 8 to Abolition, on the other hand, rejects “reforms that do not reduce the power of police.” The campaign calls for the following eight reforms: (1) “defund the police”; (2) “demilitarize communities”; (3) “remove police from schools”; (4) “free people from prisons and jails”; (5) “repeal laws that criminalize survival”; (6) “invest in community self-governance”; (7) “provide safe, accessible housing for everyone”; and (8) “fully invest in care, not cops.”
A national police misconduct registry could assist in achieving some of the goals of either approach. Under 8 Can’t Wait’s approach of implementing reforms to reduce police brutality but not eliminate police, a misconduct database would extend and enhance the proposal to “require comprehensive reporting.” Campaign Zero’s “Model Use of Force Policy” provides one public publication requirement: it encourages departments to release information about an officer’s disciplinary record following a use of deadly force incident. A national police misconduct database, made reliable through the comprehensive data collection standards the policy proposes, would ensure that the public has access to misconduct records before police kill or use physical force on a civilian. The database could also spur necessary research into patterns of police misconduct and could strengthen the research base for the existing policies or generate new evidence-based proposals to reduce not only police brutality but misconduct generally.
A national police misconduct database would also assist in achieving a few of the 8 to Abolish campaign’s concrete reforms toward its overarching goals of both “defund[ing] the police” and “demilitariz[ing] communities.” Specifically, a misconduct database could provide the information that would allow members of the public to learn about the officers who police them and give them concrete data to pressure legislatures to regulate police officers and departments. This would provide the public with names to demand that departments “fire police officers who have any excessive force complaints” and stop departments from “rehir[ing] cops involved in use of excessive force,” two of the proposals within the defund the police goal. The public nature of the misconduct database would undermine the utility of the laws that “hide, excuse, or enable police misconduct,” achieving one of the “demilitarize communities” proposals to repeal these laws. Further, the database could provide the needed evidence to enjoin unconstitutional department policies through Monell suits, which could target “broken windows policing” and abusive “surveillance technologies,” addressing two more concrete proposals.
Transparency is not a solution unto itself. A national police misconduct database requires litigants, researchers, legislatures, and the public at large to use the information in the database to make desired changes, whether they be reform or abolition. But this transparency goal would begin to shift power out of the hands of police. It would prevent police from exercising discretion in responding to public records requests. It would eliminate the utility of police exerting influence on legislatures to make their records confidential. And it would stop police from using protective orders to keep misconduct records confidential, and as this Note argues, paying to maintain the secrecy status quo.
Under the current standards governing stipulated protective orders, police have control over if and how information about their behavior and policies reach the public. This Note presents suggestive evidence that police use protective orders to shield their misconduct records from the public and will pay a settlement premium to keep that information secret.
The police officers’ demonstrated policy of strict confidentiality breeds public distrust of police and prevents meaningful reform. As Chief Justice Burger explained, “[p]eople in an open society do not demand infallibility from their institutions, but it is difficult for them to accept what they are prohibited from observing.” Justice Burger’s concern highlights the problem with current legal standards that fail to consider the public interest during the discovery process. The courts and individual litigants do have the power to reintroduce public interest considerations in the discovery process where public health and safety, public officials, and public interest are at play. But the individual incentives and high standards for modification are not conducive to considering the public’s interest in civil suits against police.
The public harms from protective orders extend beyond the law enforcement sphere. This targeted legislative solution cannot address the other protective orders that keep public health concerns secret. New evidence undermines Arthur Miller’s claim that “assertions” about “protective orders keep[ing] information regarding public health and safety hidden” rely only on “anecdotal evidence [and] research or statistical data is completely nonexistent.” For example, as the opioid crisis claimed hundreds of thousands of lives, protective orders and judicial sealing concealed key evidence of pharmaceutical companies’ misleading advertising. Protective orders and sealing orders similarly shielded the public from information about the child sex abuse crisis in the Catholic Church. And as this Note’s correlative findings may suggest, police officers might use protective orders to keep evidence of police misconduct and dangerous departmental policies indefinitely confidential.
The Black Lives Matter protests have exposed the police brutality disproportionately borne by Black Americans as a pressing public health and moral crisis. Protective orders can and do keep these threats to public health and safety secret. Powerful, repeat players should not be allowed to manipulate the court system to obscure their wrongdoing at the expense of public interest. But under current legal standards, they can. A national database of police misconduct records would provide the public with the information they need to scrutinize the officers on the street and develop concrete reforms. Congress should act to create this database. The database would serve as a small but potentially useful tool toward reducing police brutality or overhauling the institution of policing.
Chelsea Hanlock: J.D., 2020, University of California, Berkeley, School of Law. I am grateful to Professor Rebecca Wexler, Joon Hwang, Paul Messick, Debbie Oh, Benjamin Webb, Ayyan Zubair and my classmates in Secrecy: The Use and Abuse of Information Control in the Courts for their comments and insights, which shaped this Note throughout the writing process. Many thanks to the editors of the California Law Review, especially Stephanie Tilden, for their invaluable feedback.
. Richard A. Oppel Jr., Derrick Bryson Taylor & Nicholas Bogel-Burroughs, What to Know About Breonna Taylor’s Death, N.Y. Times (Apr. 16, 2021), https://www.nytimes.com/article/breonna-taylor-police.html [https://perma.cc/5BBZ-TCM9].
. Kori Rumore & Chad Yoder, Minute by Minute: How Jason Van Dyke Shot Laquan McDonald, Chi. Trib. (Jan. 18, 2019), https://www.chicagotribune.com/news/laquan-mcdonald/ct-jason-vandyke-laquan-mcdonald-timeline-htmlstory.html [https://perma.cc/5AC9-D4PS].
. Al Baker, J. Goodman & Benjamin Mueller, Beyond the Chokehold: The Path to Eric Garner’s Death, N.Y. Times, (June 13, 2015), https://www.nytimes.com/2015/06/14/nyregion/eric-garner-police-chokehold-staten-island.html [https://perma.cc/XL3B-Q59B].
. Shalia Dewan & Richard A. Oppel Jr., In Tamir Rice Case, Many Errors by Cleveland Police, Then a Fatal One, N.Y. Times (Jan. 22, 2015), https://www.nytimes.com/2015/01/23/us/in-tamir-rice-shooting-in-cleveland-many-errors-by-police-then-a-fatal-one.html [https://perma.cc/T4AF-QBD8].
. Samuel Walker, Introduction to Records of the Wickersham Commission on Law Observance and Enforcement Part 1: Records of the Committee on Official Lawlessness, at v (Samuel Walker ed., 1997), http://www.lexisnexis.com/documents/academic/upa_cis/1965_WickershamCommPt1.pdf [https://perma.cc/N6BZ-LPH4].
. Katie Nodjimbadem, The Long, Painful History of Police Brutality in the U.S., Smithsonian Mag. (July 27, 2017), https://www.smithsonianmag.com/smithsonian-institution/long-painful-history-police-brutality-in-the-us-180964098/ [https://perma.cc/J4VR-2QKY]; Walker, supra note 7, at viii.
. David Montgomery, Sandra Bland, It Turns Out, Filmed Traffic Stop Confrontation Herself, N.Y. Times (May 7, 2019), https://www.nytimes.com/2019/05/07/us/sandra-bland-video-brian-encinia.html [https://perma.cc/6NK8-H5SJ].
. Joanna Stern, They Used Smartphone Cameras to Record Police Brutality—and Change History, Wall St. J. (June 13, 2020), https://www.wsj.com/articles/they-used-smartphone-cameras-to-record-police-brutalityand-change-history-11592020827 [https://perma.cc/6W79-JV4S].
. Audra D.S. Burch, Weiyi Cai, Gabriel Gianordoli, Morrigan McCarthy & Jugal K. Patel, How Black Lives Matter Reached Every Corner of America, N.Y. Times (June 13, 2020), https://www.nytimes.com/interactive/2020/06/13/us/george-floyd-protests-cities-photos.html [https://perma.cc/M7EP-XU9G].
. Robert Lewis, Noah Veltman & Xander Landen, Is Police Misconduct a Secret in Your State?, WNYC News (Oct. 15, 2015), https://www.wnyc.org/story/police-misconduct-records/ [https://perma.cc/W27T-BG8H] (reporting that police officer records are completely confidential in twenty-three states, have limited availability in fifteen states, and are public in only twelve states).
. See Caleb J. Robertson, Comment, Restoring Public Confidence in the Criminal Justice System: Policing Prosecutions When Prosecutors Prosecute Police, 67 Emory L.J. 853, 855 (2018) (“A recent pattern of high-profile non-indictments of police officers involved in the deaths of unarmed black Americans has led to renewed scrutiny of how police-suspects are treated by the criminal justice system.”); Asit S. Panwala, The Failure of Local and Federal Prosecutors to Curb Police Brutality, 30 Fordham Urb. L.J. 639, 642–43 (2003) (“FBI agents in Los Angeles [between 1984 and 1990] generally dismissed complaints on the basis of police reports which are often misleading or self-serving. The failure to thoroughly investigate these complaints by interviewing witnesses and examining relevant documents demonstrates that federal authorities have not taken police brutality seriously, thereby creating an environment where police violence is tolerated.”).
. Kyle Rozema & Max Schanzenbach, Good Cop, Bad Cop: Using Civilian Allegations to Predict Police Misconduct, 11 Am. Econ. J.: Econ. Pol’y 225, 233 (2019) (“Even if discovered, allegations [of police misconduct] are unlikely to be admissible. Under the Federal Rules of Evidence, evidence of prior bad acts is not generally admissible to show a defendant’s propensity to commit an act.”).
. See Robertson, supra note 16, at 859 (“When police are not charged for high-profile killings, the process appears to the public and the victims to be biased in favor of the police-suspects. On the other hand, local prosecutors who zealously pursue charges against police in high-profile cases face accusations that they are over-prosecuting police for political gain.”).
. See Fed. R. Civ. P. 26(b)(1) (“Parties may obtain discovery regarding any nonprivileged matter that is relevant to any party’s claim or defense and proportional to the needs of the case . . . .”); see also infra Part III.C (describing forty-eight protective orders explicitly covering police misconduct records in suits against NYPD officers that did not go to trial, making it likely that defendant-officers produced some of these records in discovery).
. Robert Timothy Reagan, Fed. Jud. Ctr., Confidential Discovery: A Pocket Guide on Protective Orders 4 (2012), https://www.fjc.gov/sites/default/files/2012/ConfidentialDisc.pdf [https://perma.cc/F6A6-2S79].
 Pub. Citizen v. Liggett Grp., Inc., 858 F.2d 775, 780 (1st Cir. 1988) (describing parties’ “first amendment freedoms with regard to information gained through discovery” absent a protective order); Sampel v. Livingston Cnty, No. 17-cv-06548, 2019 WL 6695916, at * 3 (W.D.N.Y. Dec. 9, 2019) (“[P]arties are free to disseminate discovery materials that are not placed under a protective order as they see fit.”); U.S. ex rel. Davis v. Prince, 753 F. Supp. 2d 561, 567 (E.D. Va. 2010) (“Many circuits have sensibly held that where discovery materials are not protected by a valid protective order, parties may use that information in whatever manner they see fit.”); Harris v. Amoco Prod. Co., 768 F.2d 669, 683–84 (5th Cir. 1985) (“A party may generally do what it wants with material obtained through the discovery process, as long as it wants to do something legal.”); Jepson, Inc. v. Makita Elec. Works, Ltd., 30 F.3d 854, 858 (7th Cir. 1994) (“Absent a protective order, parties to a law suit may disseminate materials obtained during discovery as they see fit.”); Educ. Station, LLC v. Crosby, No. 05-CV-812, 2005 WL 8176953, at *1 (E.D. Mo. Dec. 14, 2005) (“Absent a protective order, parties to a suit may disseminate materials obtained during discovery as they see fit”); In re Roman Catholic Archbishop of Portland in Or., 661 F.3d 417, 424 (9th Cir. 2011) (“It is well-established that the fruits of pretrial discovery are, in the absence of a court order to the contrary, presumptively public.” (quoting San Jose Mercury News, Inc. v. U.S. Dist. Ct., 187 F.3d 1096, 1103 (9th Cir. 1999))); Okla. Hosp. Ass’n v. Okla. Publ’g Co., 748 F.2d 1421, 1424 (10th Cir. 1984) (“[I]t may be conceded that parties to litigation have a constitutionally protected right to disseminate information obtained by them through the discovery process absent a valid protective order.”); Garcia v. Chapman, No. 12-21891-CIV, 2013 WL 12061867, at *7 (S.D. Fla. Oct. 22) (surveying the law across the circuits to conclude litigants may disclose pretrial discovery), report and recommendation adopted, No. 12-21891-CIV, 2013 WL 12061868 (S.D. Fla. Nov. 15, 2013).
. Pub. Citizen, 858 F.2d. at 780 (“[T]he Supreme Court has noted that parties have general first amendment freedoms with regard to information gained through discovery that, absent a valid court order to the contrary, they are entitled to disseminate . . . .”); Reagan, supra note 20, at 4.
. Newcomb v. Esurance Ins. Servs., No. 15-CV-02062, 2017 WL 11548655, at *2 (D. Colo. Oct. 25, 2017) (describing a litigant’s power to share pretrial discovery absent a valid court order to the contrary as “long settled law”).
. Donald J. Rendall, Jr., Comment, Protective Orders Prohibiting Dissemination of Discovery Information: The First Amendment and Good Cause, 1980 Duke L.J. 766, 770; see also 8A Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure, Protective Orders § 2044.1 (3d ed.), Westlaw (databased updated Apr. 2021) (“[E]nabl[ing] litigants to use information in other cases . . . can serve important efficiency and litigation fairness goals.”).
. Seattle Times Co. v. Rhinehart, 467 U.S. 20, 37 (1984) (“[W]here . . . a protective order is entered on a showing of good cause as required by Rule 26(c), is limited to the context of pretrial civil discovery, and does not restrict the dissemination of the information if gained from other sources, it does not offend the First Amendment.”).
. In re Avandia Mktg., Sales Pracs. & Prods. Liab. Litig., 924 F.3d 662, 671 (3d Cir. 2019) (quoting Pansy v. Borough of Stroudsburg, 23 F.3d 772, 786 (3d. Cir. 1994)); see also id. (setting forth seven non-exhaustive factors for assessing whether good cause exists that focus on (1) privacy interests, (2) parties’ motivations for the information, (3) embarrassment to litigants, (4) relevance of the information to public health and safety, (5) fairness and efficiency in discovery, (6) a party’s status as a public entity or official, and (7) public interest).
. In re Ohio Execution Protocol Litig., 845 F.3d 231, 236 (6th Cir. 2016) (“Good cause exists if ‘specific prejudice or harm will result’ [to the moving party] from the absence of a protective order.” (quoting In re Roman Catholic Archbishop of Portland in Or., 661 F.3d 417, 424 (9th Cir. 2011))).
. Letter from Patrick E. Higginbotham, Advisory Comm. on Civ. Rules Chair, to the Standing Comm. on Rules of Prac. & Proc. 1–2 (June 2, 1995), https://www.uscourts.gov/sites/default/files/fr_import/CV6-1995.pdf [https://perma.cc/7LM4-8UAW]; see also Draft Minutes from Advisory Comm. on Civ. Rules 9–10 (Apr. 20, 1995), https://www.uscourts.gov/sites/default/files/fr_import/CV04-1995-min.pdf [https://perma.cc/9K54-NYLW]. The Advisory Committee has updated Rule 26 in 2000, 2006, 2007, 2010, and 2015, but the Committee has not revisited the question of eliminating the good cause requirement for stipulated protective orders. See Fed. R. Civ. P. 26 advisory committee’s notes to 2000, 2006, 2007, 2010, and 2015 amendments.
. In re Violation of Rule 28(D), 635 F.3d 1352, 1358 (Fed. Cir. 2011) (using cases in which parties entered into stipulated protective orders without a good cause finding as “abuse [of] Rule 26(c)”); Pansy v. Borough of Stroudsburg, 23 F.3d 772, 785 n.14 (3d. Cir. 1994) (expressing concern that “agreements are reached by private parties and often involve materials and information that is never even presented to the court” but “[w]ith the signature of a federal judge . . . are converted into a powerful means of mainlining and enforcing secrecy” (quoting City of Hartford v. Chase, 942 F.2d 130, 137–38 (2d Cir. 1999) (Pratt, J., concurring)).
. Procter & Gamble Co. v. Bankers Tr. Co., 78 F.3d 219, 227 (6th Cir. 1996) (holding that a stipulated protective order that the district court permitted parties to enter into without a prior good cause finding allowed them “to adjudicate their own case based on their own self-interest” which violated Rule 26(c)).
. See Seattle Times Co. v. Rhinehart, 467 U.S. 20, 32 (1984) (“[T]he Rules authorizing discovery . . . are a matter of legislative grace.”); Pub. Citizen v. Liggett Grp., Inc., 858 F.2d 775, 780 (1st Cir. 1988) (“Certainly the public has no right to demand access to discovery materials which are solely in the hands of private party litigants.”); Bond v. Utreras, 585 F.3d 1061, 1066 (7th Cir. 2009) (“[T]here is no constitutional or common-law right of public access to discovery materials exchanged by the parties but not filed with the court.”).
. See e.g., United Nuclear Corp. v. Cranford Ins. Co., 905 F.2d 1424, 1427 (10th Cir. 1990) (finding that stipulated protective orders assist in alleviating the discovery complications that complex civil litigation presents by incentivizing “parties to make full disclosure in discovery without fear of public access to sensitive information and without the expense and delay of protracted disputes over every item of sensitive information”).
. Fed. R. Civ. P. 26(c) advisory committee’s note to 1970 amendment (citing Julius M. Ames Co. v. Bostitch, Inc., 235 F. Supp. 856 (S.D.N.Y. 1964). The court in Ames Co. ordered parties to exchange “confidential business information and trade secrets” under a protective order before Rule 26(c) was enacted to encourage “full discovery” while discouraging “improper and unfair use of the material.” Ames Co., 235 F. Supp. at 857.
. Chi. Trib. Co. v. Bridgestone/Firestone, Inc., 263 F.3d 1304, 1307 (11th Cir. 2001); see also In re Mirapex Prods. Liab. Litig., 246 F.R.D. 668, 672–73 (D. Minn. 2007) (“Protective orders are, obviously, an ever-expanding feature of modern litigation.”).
. Katsuya Endo, supra note 46, at 1252–53 (explaining that advocates of stipulated protective orders highlight “the reduction of barriers to production” and that “[w]hen a court issues a protective order preventing the sharing of discovery beyond the parties, a producing party is more likely to both share material and forgo expensive screening”).
. Id. at 371 (“Parties are free to enter into discovery confidentiality agreements [which are] generally . . . enforced as a matter of contract law . . . . The protective order adds two things to the parties’ confidentiality agreement: the power of contempt, and whatever symbolic power the court’s imprimatur carries.”).
. In re Mirapex Prods. Liab. Litig., 246 F.R.D. 668, 672–73 (D. Minn. 2007) (“Protective orders recognize that parties engaged in litigation do not sacrifice all aspects of privacy or their proprietary information simply because of a lawsuit.”); see also Miller, supra note 62, at 466 (“Litigants do not give up their privacy rights simply because they have walked, voluntarily or involuntarily, through the courthouse door.”).
. Id. at 483–84 (“The result of either increased discovery factiousness or resistance to settlement would be the expenditure of litigants’ time and money on matters that often have no bearing on the merits. In addition, the energies of that most precious systemic resource—our judges—would be dissipated, and their ability to handle large cases and litigation involving issues of significant social importance would be compromised.”).
. Id. at 467–68 (describing the need to protect commercial information, like research and development information, from disclosure and concerns that absent protective orders, disclosure may harm the reputation of commercial entities and reduce profitability); Marcus, supra note 61 (evaluating the discovery confidentiality “controversy” in the early 1990s from the lens of product liability cases).
. Tennessee v. Garner, 471 U.S. 1 (1985) (evaluating whether the deadly force police used to seize a fleeing suspected felon violated the Fourth Amendment and rendered the officer liable under § 1983); Graham v. Connor, 490 U.S. 386, 395 n.10 (1989) (“[T]he Due Process Clause [of the Fourteenth Amendment] protects a pretrial detainee from the use of excessive force that amounts to punishment.”).
. Abram Chayes, Foreword: Public Law Litigation and the Burger Court, 96 Harv. L. Rev. 4, 5 (1982); Carl Tobias, Public Law Litigation and the Federal Rules of Civil Procedure, 74 Cornell L. Rev. 270, 279 (1989).
. Myriam E. Gilles, Reinventing Structural Reform Litigation: Deputizing Private Citizens in the Enforcement of Civil Rights, 100 Colum. L. Rev. 1384, 1386 (2000) (“Brown v. Board of Education, Hutto v. Finney, Roe v. Wade, Regents of the University of California v. Bakke, and scores of other landmark constitutional cases were driven by private plaintiffs who sought not only redress for themselves, but protection for society at large against the harms that they had personally suffered.”); see also Tobias, supra note 75, at 279–84 (explaining the rise of public law litigation in the 1960s and 1970s).
. Sarah Jorgensen, Settlement Reached in Police-Involved Shooting of Man in a Wheelchair, CNN (Dec. 16, 2016), https://www.cnn.com/2016/12/16/us/delaware-police-shoot-man-wheelchair-settlement/index.html [https://perma.cc/68D8-XWMW].
. Matt Lait, Controversial Police Restraint to Be Banned, L.A. Times (July 4, 1997), https://www.latimes.com/archives/la-xpm-1997-jul-04-me-9731-story.html [https://perma.cc/HKQ3-K632]. For more details on these cases and others, see Fact Sheet: Civil Lawsuits Lead to Better Safer Law Enforcement, Ctr. for Just. & Democracy at N.Y. L. Sch. (June 20, 2017), https://centerjd.org/content/fact-sheet-civil-lawsuits-lead-better-safer-law-enforcement [https://perma.cc/5TWV-4B7T].
. Lise Olsen & Blake Paterson, Council Approves Rare Settlement in HPD Shooting Death of Unarmed Man, Hous. Chron. (June 28, 2017), https://www.chron.com/news/houston-texas/houston/article/Council-to-review-settlement-in-HPD-11253001.php [https://perma.cc/ZX6Q-ZJ76].
. Joanna C. Schwartz, What Police Learn from Lawsuits, 33 Cardozo L. Rev. 841, 844 (2012); see also Joanna C. Schwartz, Myths and Mechanics of Deterrence: The Role of Lawsuits in Law Enforcement Decisionmaking, 57 UCLA L. Rev. 1023, 1067 (2010) (“Given the infrequency with which departments seek to gather information from lawsuits, and the barriers when they do try to gather this information, it seems fair to conclude that most law enforcement officials know little about lawsuits alleging misconduct by their officers.”).
. See Cynthia Conti-Cook, A New Balance: Weighing Harms of Hiding Police Misconduct Information from the Public, 22 CUNY L. Rev. 148 (2019) (describing how the NYPD’s policy of keeping police misconduct records confidential harms the public by continuing to traumatize victims, families, and communities that lack information about officers who use excessive force; damages the public’s trust in internal disciplinary procedures; and impedes public discourse about policing).
. Derek Hawkins, Officer Charged in George Floyd’s Death Used Fatal Force Before and Had History of Complaints, Wash. Post (May 29, 2020), https://www.washingtonpost.com/nation/2020/05/29/officer-charged-george-floyds-death-used-fatal-force-before-had-history-complaints/ [https://perma.cc/9DD4-GT38].
. Shalia Dewan & Serge F. Kovaleski, Thousands of Complaints Do Little to Change Police Ways, N.Y. Times (May 30, 2020), https://www.nytimes.com/2020/05/30/us/derek-chauvin-george-floyd.html?smid=url-share [https://perma.cc/3378-DA8Y].
. Civilian complaints against Minneapolis police officers are no longer available in Minneapolis’s open data portal. The city instead provides “a summary of complaints filed against officers in the Minneapolis Police Department.” Officer Complaint Data, Minneapolis, City of Lakes (Sept. 9, 2020), https://www.minneapolismn.gov/resident-services/public-safety/complaints-and-compliments/police-officer-complaint-process/officer-complaint-data/ [https://perma.cc/UD7F-23FZ].
. Dewan & Kovaleski supra note 97 (describing the discrepancy between the public complaint database, which listed twelve complaints for Mr. Chauvin, while the Internal Affairs summary released by the city after George Floyd’s killing listed seventeen).
. Johanna Wald, Chicago Cop Jason Van Dyke’s Record Was a Warning Sign, Marshall Project (Oct. 28, 2018) https://www.themarshallproject.org/2018/10/28/warning-signs-were-clear-before-laquan-mcdonald-s-murder [https://perma.cc/24S6-BS2S].
. Eliott C. McLaughlin, Chicago Officer Had History of Complaints Before Laquan McDonald Shooting, CNN (Nov. 26, 2015), https://www.cnn.com/2015/11/25/us/jason-van-dyke-previous-complaints-lawsuits/index.html [https://perma.cc/Y9VK-SA7N].
. James Queally, California Police Uphold Few Complaints of Officer Misconduct and Investigations Stay Secret. L.A. Times (Sept. 23, 2018), https://www.latimes.com/local/lanow/la-me-police-misconduct-complaints-20180923-story.html [https://perma.cc/MM6J-6P7M].
. Police Disciplinary Records Are Largely Kept Secret in US, WTTW (June 13, 2020), https://news.wttw.com/2020/06/13/police-disciplinary-records-are-largely-kept-secret-us [https://perma.cc/VJ2R-NS59] (“Phil Stinson, who has collected data on thousands of police charged, investigated or convicted of crimes, said that most officers go through their careers with few complaints against them, and that generally a small percentage of officers account for an outsize share of complaints.”).
. See, e.g., File a Complaint, City of N.Y., https://www1.nyc.gov/site/ccrb/complaints/file-complaint.page [https://perma.cc/KU9T-K8EX] (describing the complaint process and explaining that once a party submits a complaint against a NYPD officer, the complaining party “will be asked to come to our office so that an investigator can take a formal statement”); Investigative Process, City of Chi., http://copadev.wpengine.com/investigations/investigative-process/ (click COPA Investigative Process) [https://perma.cc/ZEN6-3NJ6] (explaining the process following a complaint, which includes “interviews with complainants, witnesses, and subjects.”); Report Employee Misconduct, L.A. Police Dep’t, https://www.lapdonline.org/our_communities/content_basic_view/9217 [https://perma.cc/9QAT-ZQ5T] (“Although it is not required, the Department encourages community members to make these reports in person so a supervisor has an opportunity to do a thorough initial assessment of your complaint. . . . [T]he investigation may include interviews of witnesses and officers; a review of Department records, policies, and procedures; an inspection of medical records, photographs and other evidence; and legal analysis.”).
. 436 U.S. 658 (1978); see also Los Angeles County v. Humphries, 562 U.S. 29, 37 (2010) (clarifying that Monell’s limits on municipal liability applied to both claims for damages and for prospective relief).
. See id.; see also City of Canton v. Harris, 489 U.S. 378, 379 (1989) (finding municipal liability under § 1983 where inadequacy of police amounted to deliberate indifference to constitutional rights).
. See Reuben Fischer-Baum, Allegations of Police Misconduct Rarely Result in Charges, FiveThirtyEight (Nov. 25, 2014), https://fivethirtyeight.com/features/allegations-of-police-misconduct-rarely-result-in-charges/ [https://perma.cc/KJL4-4RQM].
. Past 42 Officer-Involved Killings in Hennepin County Had No Indictment, Fox 9 KMSP (Mar. 29, 2016), https://www.fox9.com/news/past-42-officer-involved-killings-in-hennepin-county-had-no-indictments [https://perma.cc/2GXS-593M].
. Fed. R. Crim. P. 6(e)(2)(B) (prohibiting grand jurors, interpreters, court reporters, attorneys for the government or others who receive disclosures from “disclos[ing] a matter occurring before the grand jury”).
. Will Wright, Breonna Taylor Grand Juror Says Homicide Charges Were Not Presented, N.Y. Times (Oct. 20, 2020) https://www.nytimes.com/2020/10/20/us/breonna-taylor-grand-jury.html [https://perma.cc/PSL9-GDCB].
. Levine, supra note 130, at 1451, 1465 (“[Prosecutors rely] on the police in cases against civilian defendants in terms of arrests, evidence collection, and testimony. Such reliance on the police leads to a conflict of interest when it is an officer who must be prosecuted.”).
. Del. Code Ann. tit. 11, § 9200(c)(12) (2020) (“All [disciplinary] records shall be and remain confidential and shall not be released to the public.”); see also Miss. Code Ann. tit. 25, § 25-1-100 (2013) (exempting personnel records “in possession of a public body” from disclosure under Mississippi’s Public Records Act of 1983).
. Alaska Stat. tit. 39, § 39.25.080 (2020); Colo. Rev. Stat. tit. 24, § 24-72-204 (2020); Idaho Code. § 74-106 (2019); 5 Ill. Comp. Stat. 140/7 (Supp. 2020); Iowa Code § 22.7(11) (2021); Mass. Gen. Laws ch. 4, § 7(26)(c) (2019).
. Ark. Code Ann. § 25-19-105(c)(1) (2019) (making “employee evaluation or job performance records” public only when “the records form a basis for the decision to suspend or terminate the employee.”); Haw. Rev. Stat. § 92F-14 (2020) (exempting “personnel file[s]” from disclosure except for “information related to employment misconduct that results in an employee’s suspension or discharge”); Ind. Code § 5-14-3-4 (2020).
. Id. (claiming that “[p]olice disciplinary records are generally withheld under the privacy exception” in D.C. Code § 2-534(a)(2) (2001 & Supp. 2020). This statute provides, “[i]nformation of a personal nature where the public disclosure thereof would constitute a clearly unwarranted invasion of personal privacy” may be exempt from disclosure. Id.
. Md. Dep’t of State Police v. Dashiell, 443 Md. 435, 460 (2015) (“After having decided that the requested records are ‘personnel records’ and, thus, exempt from disclosure, we need not address Ms. Dashiell’s claim that, as the complainant, she is a ‘person in interest’ under the ‘investigatory records’ exemption of the Maryland Public Information Act.”).
. Ala. Code § 36-12-40 (2013); Ariz. Rev. Stat. Ann. §§ 39-121 to 39-128 (2011); Conn. Gen. Stat. §§ 1-210 (2019); Fla. Stat. § 112.533(2)(a) (2020); Ga. Code Ann. § 50-18-72(a)(8) (2016); Me. Stat. tit. 30-A, §§ 503(1)(B)(5)l, 2702(1)(B)(5) (2011), tit. 5, § 7070(2)(E) (2013); Minn. Stat. § 13.82 (2020); N.D. Cent. Code § 44-04-18 (2013 & Supp. 2019); Ohio Rev. Code Ann. § 149.43 (2016 & Supp. 2020); Utah Code Ann. § 63G-2-301 (West 2020); Wash. Rev. Code § 42.56.050 (2020); Wis. Stat. § 19.36(10)(b) (2020).
. Id. (citing State v. Jones, 59 A.3d 320 (Conn. App. Ct. 2013); In re Brooks, 548 S.E.2d 748, 755 (N.C. Ct. App. 2001); State v. Cano, 743 P.2d 956, 958 (Ariz. Ct. App. 1987); Jinks v. State, 274 S.E.2d 46 (Ga. Ct. App. 1980)).
. See NYC Administrative Code § 7-114, Civil Actions Regarding the Police Department, N.Y.C. L. Dep’t, https://www1.nyc.gov/site/law/public-resources/nyc-administrative-code-7-114.page [https://perma.cc/B8AC-VG8W] [hereinafter “NYC Administrative Code”].
. Meta S. Brown, City Governments Making Public Data Easier to Get: 90 Municipal Open Data Portals, Forbes (Apr. 29, 2018), https://www.forbes.com/sites/metabrown/2018/04/29/city-governments-making-public-data-easier-to-get-90-municipal-open-data-portals/?sh=3ad7a9495a0d [https://perma.cc/XDC8-ZQKJ].
. See NYC Administrative Code, supra note 172. Mayor Bill de Blasio signed Local Law 166 into effect in 2017. “The law mandates such reporting twice a year . . . regarding actions commenced in the preceding five year period.” Id. Accordingly, 2014 to 2018 were the years for which a full accounting of civil rights suits against the NYPD was available at the time of analysis.
. The remaining variables pre-populated in the NYPD Alleged Misconduct Matters dataset include matter name, plaintiff & firm, individual defendants, and tax #. See NYC Administrative Code, supra note 172.
. Local Civil Rule 83.10. Plan for Certain § 1983 Cases Against the City of New York (Southern District Only), S.D.N.Y. (Sept. 23, 2014), https://www.nysd.uscourts.gov/sites/default/files/pdf/Local-Civil-Rule-83.10.Final.pdf [https://perma.cc/5QM9-EKSM] [hereinafter “S.D.N.Y. Local Rule 83.10”]. This rule applies to cases filed by “represented plaintiff[s]” against the New York City Police Department and/or the City of New York for “excessive force, false arrest, or malicious prosecution” by NYPD officers. Id.
. Inferences drawn from a subset of an entire population can be generalizable to the full population if the sample is representative. “Generally the term representativeness is often used to indicate that a sample mirrors a population, reflecting all essential properties of the population in a correct way.” Definition Representativeness, Statista, https://www.statista.com/statistics-glossary/definition/361/representativeness/ [https://perma.cc/8WA7-N2HH].
. Confidentiality Stipulation and Protective Order, Hunte, v. City of New York, No. 16-cv-0188 at 1 (E.D.N.Y. Sept. 16, 2016) [hereinafter Hunte Order]; Stipulation and Order of Confidentiality, Loadholt v. Freeland, No. 14-cv-6904 at 1 (E.D.N.Y. May 11, 2015) [hereinafter Loadholt Order]; Protective Order Regarding Documents to be Produced on an Attorneys’-Eyes-Only-Basis, Marrero v. City of New York, No. 14-cv-9620 at 1 (S.D.N.Y. Aug. 31, 2015) [hereinafter Marrero Order]; Confidentiality Stipulation and Protective Order, Parker-El v. City of New York, No. 13-cv-6996 at 2 (S.D.N.Y. Nov. 7, 2016) [hereinafter Parker-El Order]; Confidentiality Stipulation and Protective Order, Simmons v. City of New York, No. 15-cv-06383 at 1 (E.D.N.Y. Feb. 26, 2016) [hereinafter Simmons Order]; Confidentiality Stipulation and Protective Order, Rodgers v. City of New York, No. 15-cv-6107 at 1 (E.D.N.Y. July 26, 2016) [hereinafter Rodgers Order]; Stipulation and Protective Order, Cirillo v. City of New York, No. 15-cv-200 at 1 (E.D.N.Y. Sep. 2, 2015) [hereinafter Cirillo Order]; Confidentiality Stipulation and Order, Charles v. City of New York, No. 15-cv-4510 at 1 (E.D.N.Y. Feb. 5, 2016) [hereinafter Charles Order]; Stipulation and Protective Order, Galarza v. City of New York, No. 14-cv-10039 at 1 (S.D.N.Y. Apr. 16, 2015) [hereinafter Galarza Order]; Stipulation of Confidentiality and Protective Order, Shaheed v. City of New York, No. 14-cv-7424 at 3 (S.D.N.Y. Dec. 5, 2016) [hereinafter Shaheed Order]; Stipulation and Order of Confidentiality, Accede v. City of New York, No: 16-cv-6222 at 1 (E.D.N.Y. Feb. 9, 2017) [hereinafter Accede Order]; Stipulation and Order of Confidentiality, Andolina v. City of New York, No. 19-cv-9211 at 1 (S.D.N.Y. Nov. 23, 2015) [hereinafter Andolina Order]; Stipulated Confidentiality Agreement and Protective Order, Ballou v. City of New York, No. 15-cv-1346 at 1 (E.D.N.Y. Oct. 7, 2015) [hereinafter Ballou Order]; Stipulation of Confidentiality and Protective Order, Breeden v. New York City, No. 18-cv-5048 at 1 (E.D.N.Y. Jan. 31, 2019) [hereinafter Breeden Order]; Stipulation and Order of Confidentiality, China v. City of New York, No. 16-cv-6699 at 1 (E.D.N.Y. Apr. 10, 2017) [hereinafter China Order]; Stipulation and Order of Confidentiality, Grant v. City of New York, No. 15-cv-3635 at 1 (E.D.N.Y. Apr. 4, 2016) [hereinafter Grant Order]; Stipulation and Order of Confidentiality, Onsoy v. Izzo, No. 15-cv-5574 at 1 (E.D.N.Y. Sept. 2, 2016) [hereinafter Onsoy Order]; Stipulation and Order of Confidentiality, Smith v. City of New York, No. 15-cv-1907 at 1 (E.D.N.Y. Jan. 20, 2016) [hereinafter Smith Order]; Stipulation and Protective Order, Moore v. City of New York, No. 15-cv-4365 (E.D.N.Y. Feb. 1, 2016) [hereinafter Moore Order]; Confidentiality Stipulation and Order, Simon v. City of New York, No. 16-cv-07189 at 1 (E.D.N.Y. Aug. 29, 2017) [hereinafter Simon Order]; Confidentiality Stipulation and Protective Order, Taylor v. City of New York, No. 18-cv-5413 at 1 (E.D.N.Y. Mar. 25, 2019) [hereinafter Taylor Order]; Stipulation and Order of Confidentiality, Fullerton v. City of New York, No. 14-cv-6029 at 1 (E.D.N.Y. Jan. 20, 2015) [hereinafter Fullerton Order]; Stipulation and Order of Confidentiality, Flit v. City of New York, No. 15-cv-3698 at 1 (E.D.N.Y. Sept. 1, 2015) [hereinafter Flit Order]; Stipulation of Confidentiality and Protective Order, McClain v. City of New York, No. 15-cv-6813 at 1 (E.D.N.Y. Nov. 18, 2016) [hereinafter McClain Order]; Stipulation and Order of Confidentiality, Martin v. City of New York, No. 18-cv-5935 at 1 (E.D.N.Y. Mar. 4, 2019) [hereinafter Martin Order]; Stipulation and Protective Order Governing Use of Disclosure of Confidential Materials, Brown v. City of New York, No. 15-cv-4913 at 1 (E.D.N.Y. Aug. 3, 2015) [hereinafter Brown Order]; Stipulation and Order of Confidentiality, Pearson v. City of New York, No. 16-cv-5798, 1 (E.D.N.Y. Mar. 22, 2017) [hereinafter Pearson Order]; Stipulation of Confidentiality and Protective Order, Windley v. City of New York, No. 16-cv-4529 at 1 (E.D.N.Y. Oct. 31, 2016) [hereinafter Windley Order]; Stipulation of Confidentiality and Protective Order, Brown v. Bab, No. 16-cv-3942 at 1 (E.D.N.Y. Mar. 10, 2017) [hereinafter Bab Order]; Stipulation of Confidentiality and Protective Order, Turner v. City of New York, No. 15-cv-00074 at 1 (E.D.N.Y. Mar. 11, 2016) [hereinafter Turner Order]; Stipulation of Confidentiality and Protective Order, Jackson v. City of New York, No. 14-cv-8975 at 1 (S.D.N.Y. June 30, 2016) [hereinafter Jackson Order]; Stipulation of Confidentiality and Protective Order, Pratt v. City of New York, No. 15-cv-04095 at 1 (E.D.N.Y. Feb. 1, 2016) [hereinafter Pratt Order]; Stipulation and Order of Confidentiality, Moya v. City of New York, No. 17-cv-4254 at 1 (E.D.N.Y. Mar. 12, 2018) [hereinafter Moya Order]; Stipulation and Order of Confidentiality, Santos v. Sim, No. 15-cv-1732 at 1 (E.D.N.Y. Dec. 1, 2016) [hereinafter Santos Order]; Stipulated Confidentiality Agreement and Protective Order Concerning the Individual Defendant’s Sensitive Records, McFadden v. City of New York, No. 14-cv-6940 at 1 (E.D.N.Y. June 2, 2015) [hereinafter McFadden Order]; Stipulation and Order of Confidentiality, Jose v. City of New York, No. 17-cv-05082 at 1 (E.D.N.Y. Jan. 12, 2018) [hereinafter Jose Order]; Stipulation and Order of Confidentiality, Callender v. City of New York, No. 16-cv-1706 at 1 (E.D.N.Y. July 5, 2017) [hereinafter Callender Order]; Stipulation and Order of Confidentiality, Freese v. Mattina, No. 17-cv-4390 at 1 (E.D.N.Y. May 15, 2018) [hereinafter Freese Order]; Protective Order Concerning Confidential Information, Pilipenko v. City of New York, No. 15-cv-6053 at 1 (E.D.N.Y. May 11, 2016) [hereinafter Pilipenko Order]; Confidentiality Stipulation and Protective Order, Betts v. Rodriquez, No. 15-cv-3836 at 2 (S.D.N.Y. Aug. 19, 2016) [hereinafter Betts Order]; Stipulation and Protective Order, Wilson v. City of New York, No. 17-cv-01960 at 1 (E.D.N.Y. Aug. 28, 2017) [hereinafter Wilson Order].
. Stipulation of Confidentiality and Protective Order, Moore v. Newton, No. 14-cv-6473 (E.D.N.Y. Mar. 26, 2015) [hereinafter Newton Order]; Freese Order, supra note 184, at 1. The order in Freese uses the standard “good cause” rationale in addition to mentioning a previous agreement. Id. Moore v. Newton deviated from standard language by saying the parties agreed that good cause existed. Newton Order, supra, at 1.
. Stipulation of Confidentiality and Protective Order, Hennis v. City of New York, No. 17-cv-3458 at 1 (E.D.N.Y. Jan 12, 2018) [hereinafter Hennis Order]; Stipulation of Confidentiality and Proposed Protective Order, Hoyte v. City of New York, No. 17-cv-1269 at 1 (E.D.N.Y. Nov. 14, 2017) [hereinafter Hoyte Order]; Stipulation and Order of Confidentiality, Taggart v. City of New York, No. 17-cv-5445 at 1 (E.D.N.Y. Jan. 16, 2018) [hereinafter Taggart Order]; Stipulation of Confidentiality and Protective Order, Alexander v. City of New York, No. 16-cv-5889 at 1 (E.D.N.Y. Sept. 26, 2017) [hereinafter Alexander Order].
. Order, Linton v. City of New York, No. 15-cv-2556 at 1 (E.D.N.Y. Jan. 25, 2016) [hereinafter Linton Order]; Order, Fedd v. City of New York, No. 15-cv-4015 at 1 (E.D.N.Y. May 10, 2016), ECF No. 20-1 [hereinafter Fedd Order].
. Alexander Order, supra note 186, at 1; Hunte Order, supra note 184, at 1; Marrero Order, supra note 184, at 1; Cirillo Order, supra note 184, at 1; Taggart Order, supra note 186, at 1; Windley Order, supra note 184, at 1; Charles Order, supra note 184, at 1; Wilson Order, supra note 184, at 1; Shaheed Order, supra note 184, at 2; Galarza Order, supra note 184, at 1; Turner Order, supra note 184, at 1; Betts Order, supra note 184, at 1; Accede Order, supra note 184, at 1; Andolina Order, supra note 184, at 2; Ballou Order, supra note 184, at 1; Breeden Order, supra note 184, at 1; Smith Order, supra note 184, at 1; Fullerton Order, supra note 184, at 1; Flit Order, supra note 184, at 1; Santos Order, supra note 184, at 1; McFadden Order, supra note 184, at 1; Jose Order, supra note 184, at 1; Rodgers Order, supra note 184, at 1; China Order, supra note 184, at 1; Grant Order, supra note 184, at 1; Hennis Order, supra note 186, at 1; Loadholt Order, supra note 184, at 1; Moore Order, supra note 184, at 1; Simon Order, supra note 184, at 1; McClain Order, supra note 184, at 1; Martin Order, supra note 184, at 1; Brown Order, supra note 184, at 1; Bab Order, supra note 184, at 1; Callender Order, supra note 184, at 1; Fedd Order, supra note 187, at 1.
. Hoyte Order, supra note 186, at 1; Onsoy Order, supra note 184, at 1; Parker-El Order, supra note 184, at 1; Simmons Order, supra note 184, at 1; Taylor Order, supra note 184, at 1; Pearson Order, supra note 184, at 1; Pratt Order, supra note 184, at 1; Moya Order, supra note 184, at 1; Freese Order, supra note 184, at 1.
. Alexander Order, supra note 186, at 1; Breeden Order, supra note 184, at 1; Taggart Order, supra note 186, at 1; Hennis Order, supra note 186, at 1; Rodgers Order, supra note 184, at 1; Charles Order, supra note 184, at 1; Galarza Order, supra note 184, at 1; Hunte Order, supra note 184, at 1; Marrero Order, supra note 184, at 1; Parker-El Order, supra note 184, at 1; Cirillo Order, supra note 184, at 1; Wilson Order, supra note 184, at 1; Windley Order, supra note 184, at 1; Shaheed Order, supra note 184, at 2; Fedd Order, supra note 187, at 1; Betts Order, supra note 184, at 1.
. Scheduling Order, Avila v. City of New York, No. 15-cv-09193 (S.D.N.Y. Feb. 11, 2016), ECF No. 24; Freese Order, supra note 184, at 1; Minute Entry, Ortiz v. City of New York, No. 17-cv-01386 (E.D.N.Y. Sept. 13, 2017), ECF No. 15; Order Granting Motion to Compel, Siemionko v. City of New York, No. 15-cv-04329 (E.D.N.Y. Mar. 11, 2016), ECF No. 18 (“I further direct defendants to disclose [information related to a complaining witness] to plaintiff’s counsel, subject to an ‘attorney’s eyes only’ protective order.”).
. Accede Order, supra note 184, at 2; Alexander Order, supra note 186, at 2; Andolina Order, supra note 184, at 1; Ballou Order, supra note 184, at 1–2; Breeden Order, supra note 184, at 2; China Order, supra note 184, at 2; Grant Order, supra note 184, at 2; Hennis Order, supra note 186, at 1–2; Hoyte Order, supra note 186, at 2; Hunte Order, supra note 184, at 1; Loadholt Order, supra note 184, at 1; Onsoy Order, supra note 184, at 1–2; Parker-El Order, supra note 184, at 2; Simmons Order, supra note 184, at 1; Smith Order, supra note 184, at 1–2; Rodgers Order, supra note 184, at 1; Cirillo Order, supra note 184, at 2; Moore Order, supra note 184, at 1; Simon Order, supra note 184, at 1; Taylor Order, supra note 184, at 2; Taggart Order, supra note 186, at 2; Fullerton Order, supra note 184, at 1; Charles Order, supra note 184, at 1; Flit Order, supra note 184, at 1–2; Wilson Order, supra note 184, at 2; McClain Order, supra note 184, at 1; Martin Order, supra note 184, at 1; Brown Order, supra note 184, at 1; Pearson Order, supra note 184, at 1; Windley Order, supra note 184, at 2; Bab Order, supra note 184, at 2; Turner Order, supra note 184, at 1; Jackson Order, supra note 184, at 1; Pratt Order, supra note 184, at 1; Moya Order, supra note 184, at 1–2; Galarza Order, supra note 184, at 2; Santos Order, supra note 184, at 1; McFadden Order, supra note 184, at 1; Jose Order, supra note 184, at 1; Callender Order, supra note 184, at 1; Fedd Order, supra note 187, at 1; Freese Order, supra note 184, at 2; Pilipenko Order, supra note 184, at 2; Betts Order, supra note 184, at 2; Shaheed Order, supra note 184, at 3.
. Accede Order, supra note 184, at 2; Alexander Order, supra note 186, at 2; Andolina Order, supra note 184, at 1; Ballou Order, supra note 184, at 2; Breeden Order, supra note 184, at 2; China Order, supra note 184, at 2; Grant Order, supra note 184, at 2; Hennis Order, supra note 186, at 2; Hoyte Order, supra note 186, at 2; Hunte Order, supra note 184, at 1–2; Loadholt Order, supra note 184, at 1; Onsoy Order, supra note 184, at 2; Parker-El Order, supra note 184, at 2; Simmons Order, supra note 184, at 1–2.; Smith Order, supra note 184, at 2; Rodgers Order, supra note 184, at 1–2; Cirillo Order, supra note 184, at 2; Moore Order, supra note 184, at 1; Simon Order, supra note 184, at 1; Taylor Order, supra note 184, at 2; Taggart Order, supra note 186, at 2; Fullerton Order, supra note 184, at 1; Charles Order, supra note 184, at 1–2; Flit Order, supra note 184, at 1–2; Wilson Order, supra note 184, at 2; McClain Order, supra note 184, at 1; Martin Order, supra note 184, at 1; Brown Order, supra note 184, at 1–2; Pearson Order, supra note 184, at 1; Windley Order, supra note 184, at 2; Bab Order, supra note 184, at 2; Turner Order, supra note 184, at 1; Jackson Order, supra note 184, at 1; Pratt Order, supra note 184, at 1; Galarza Order, supra note 184, at 2; Santos Order, supra note 184, at 1–2; McFadden Order, supra note 184, at 1; Jose Order, supra note 184, at 2; Callender Order, supra note 184, at 1; Fedd Order, supra note 187, at 1; Freese Order, supra note 184, at 2; Betts Order, supra note 184, at 2; Shaheed Order, supra note 184, at 3; Marrero Order, supra note 184, at 2.
. S.D.N.Y. Local Rule 83.10, supra note 179 (presumptively protecting “New York City Police Department (“NYPD”) personnel and disciplinary-related records, and records of investigations regarding the conduct of Members of the service of the NYPD conducted by the NYPD, the Civilian Complaint Review Board, or other agencies”).
. Alexander Order, supra note 186, at 2; Hennis Order, supra note 186, at 2; Hoyte Order, supra note 186, at 2; Taylor Order, supra note 184, at 2; Taggart Order, supra note 186, at 2; Wilson Order, supra note 184, at 2; Windley Order, supra note 184, at 2; Bab Order, supra note 184, at 2.
. Alexander Order, supra note 186, at 2; Breeden Order, supra note 184, at 2; Taggart Order, supra note 186, at 2; Ballou Order, supra note 184, at 2; Grant Order, supra note 184, at 2; Onsoy Order, supra note 184, at 2; Bab Order, supra note 184, at 2; Moya Order, supra note 184, at 2.
. Andolina Order, supra note 184, at 2; Simmons Order, supra note 184, at 4; Rodgers Order, supra note 184, at 4; Taylor Order, supra note 184, at 2; Fedd Order, supra note 187, at 4; Pilipenko Order, supra note 184, at 2.
. Parker-El Order, supra note 184, at 2; Simmons Order, supra note 184, at 2; Pearson Order, supra note 184, at 1; Taylor Order, supra note 184, at 2; Jackson Order, supra note 184, at 1; Pratt Order, supra note 184, at 1; Santos Order, supra note 184, at 2.
. Ballou Order, supra note 184, at 2; Hoyte Order, supra note 186, at 2; Parker-El Order, supra note 184, at 2; Simmons Order, supra note 184, at 2; Charles Order, supra note 184, at 2; Pearson Order, supra note 184, at 2; Jackson Order, supra note 184, at 2.
. Accede Order, supra note 184, at 2; Alexander Order, supra note 186, at 2; Andolina Order, supra note 184, at 1–2; Breeden Order, supra note 184, at 2–3; China Order, supra note 184, at 2; Grant Order, supra note 184, at 2; Hennis Order, supra note 186, at 2; Hunte Order, supra note 184, at 2; Loadholt Order, supra note 184, at 2; Onsoy Order, supra note 184, at 2; Smith Order, supra note 184, at 2; Rodgers Order, supra note 184, at 2; Cirillo Order, supra note 184, at 2; Moore Order, supra note 184, at 2; Simon Order, supra note 184, at 1–2; Taggart Order, supra note 186, at 2; Fullerton Order, supra note 184, at 1–2; Flit Order, supra note 184, at 2; McClain Order, supra note 184, at 1–2; Martin Order, supra note 184, at 1–2; Brown Order, supra note 184, at 2; Bab Order, supra note 184, at 2; Turner Order, supra note 184, at 1–2; Pratt Order, supra note 184, at 1–2; Moya Order, supra note 184, at 2; Galarza Order, supra note 184, at 2; Santos Order, supra note 184, at 2; Jose Order, supra note 184, at 2; Callender Order, supra note 184, at 2; Fedd Order, supra note 187, at 2; Freese Order, supra note 184, at 2; Betts Order, supra note 184, at 2; Shaheed Order, supra note 184, at 3.
. Ballou Order, supra note 184, at 3; Hoyte Order, supra note 186, at 3; Onsoy Order, supra note 184, at 2; Parker-El Order, supra note 184, at 2; Simmons Order, supra note 184, at 2; McClain Order, supra note 184, at 2; Pearson Order, supra note 184, at 2; Turner Order, supra note 184, at 2; Jackson Order, supra note 184, at 2; Pratt Order, supra note 184, at 2; Moya Order, supra note 184, at 2; Freese Order, supra note 184, at 2; Grant Order, supra note 184, at 2.
. Accede Order, supra note 184, at 2; Andolina Order, supra note 184, at 2; Smith Order, supra note 184, at 2; Fullerton Order, supra note 184, at 2; Brown Order, supra note 184, at 2; Bab Order, supra note 184, at 2; Santos Order, supra note 184, at 2; Jose Order, supra note 184, at 2; Callender Order, supra note 184, at 2; Loadholt Order, supra note 184, at 2; Charles Order, supra note 184, at 2; Flit Order, supra note 184, at 2; Wilson Order, supra note 184, at 3; Martin Order, supra note 184, at 2.
. Alexander Order, supra note 186, at 3; Breeden Order, supra note 184, at 3; Hennis Order, supra note 186, at 3; Hunte Order, supra note 184, at 2; Marrero Order, supra note 184, at 3; Rodgers Order, supra note 184, at 2; Moore Order, supra note 184, at 2; Simon Order, supra note 184, at 2; Taggart Order, supra note 186, at 3; Windley Order, supra note 184, at 3; Shaheed Order, supra note 184, at 3; McFadden Order, supra note 184, at 2; Proposed Order, Fedd v. City of New York, No. 15-cv-4015 at 2 (May 10, 2016), ECF No. 20; Betts Order, supra note 184, at 2.
. Andolina Order, supra note 184, at 4–5; Ballou Order, supra note 184, at 5; Breeden Order, supra note 184, at 5–6; Hennis Order, supra note 186, at 5; Hoyte Order, supra note 186, at 5; Hunte Order, supra note 184, at 4; Marrero Order, supra note 184, at 3–4; Parker-El Order, supra note 184, at 5–6; Rodgers Order, supra note 184, at 4; Cirillo Order, supra note 184, at 4; Simon Order, supra note 184, at 3; Taggart Order, supra note 186, at 5–6; Fullerton Order, supra note 184, at 3; Charles Order, supra note 184, at 5; Wilson Order, supra note 184, at 5–6; Windley Order, supra note 184, at 6; Shaheed Order, supra note 184, at 5; Fedd Order, supra note 187, at 4; Betts Order, supra note 184, at 4.
. Accede Order, supra note 184, at 5; Flit Order, supra note 184, at 4–5; McClain Order, supra note 184, at 4–5; Martin Order, supra note 184, at 4–5; Brown Order, supra note 184, at 4; Pearson Order, supra note 184, at 4; Bab Order, supra note 184, at 6; Jackson Order, supra note 184, at 5; Pratt Order, supra note 184, at 5; Galarza Order, supra note 184, at 6; Santos Order, supra note 184, at 4–5; Turner Order, supra note 184, at 4–5; Jose Order, supra note 184, at 5; Callender Order, supra note 184, at 5; Freese Order, supra note 184, at 5; Pilipenko Order, supra note 184, at 5; Loadholt Order, supra note 184, at 4–5; Fullerton Order, supra note 184, at 5; Cirillo Order, supra note 184, at 5–6; Smith Order, supra note 184, at 4–5; China Order, supra note 184, at 6; Andolina Order, supra note 184, at 5.
. Ballou Order, supra note 184, at 6; Grant Order, supra note 184, at 6; Onsoy Order, supra note 184, at 6; Parker-El Order, supra note 184, at 6; Simmons Order, supra note 184, at 5; McFadden Order, supra note 184, at 4.
. Alexander Order, supra note 186; Breeden Order, supra note 184; Hennis Order, supra note 186; Hoyte Order, supra note 186; Hunte Order, supra note 184; Rodgers Order, supra note 184; Simon Order, supra note 184; Taylor Order, supra note 184; Taggart Order, supra note 186; Charles Order, supra note 184; Windley Order, supra note 184; Betts Order, supra note 184; Shaheed Order, supra note 184; Newton Order, supra note 185; Marrero Order, supra note 184; Wilson Order, supra note 184.
. Andrea Kuperman, Comm. on Rules of Prac. & Proc., Case Law on Entering Protective Orders, Entering Sealing Orders, and Modifying Protective Orders 1 (2010) (“[C]ourts maintain discretion to modify protective orders . . . .”).
. N.Y. Civ. Rights Law § 50–a (McKinney 2019) (repealed 2020) (“Personnel records of police officers . . . shall be considered confidential and not subject to inspection or review . . . .”); Cal. Penal Code § 832.7(a) (West 2019) (“[T]he personnel records of peace officers and custodial officers . . . are confidential and shall not be disclosed . . . .”).
. See, e.g., Cox v. N.Y.C. Hous. Auth., 482 N.Y.S.2d 5, 6 (N.Y. App. Div. 1984) (describing that the defendant police officer invoked New York Civil Rights Law § 50–a to refuse to comply with requested discovery for police officer’s personnel records because the statute made them “confidential”); City of San Diego v. Superior Ct., 136 Cal. Rptr. 112, 112 (Cal. Ct. App. 1981) (seeking to prevent “two defendant police officers” from disclosing their reprimand history on the ground that “the information is privileged under Penal Code sections 832.7 and 832.8” and “immune from discovery”).
. See Huthnance v. District of Columbia, 255 F.R.D. 285, 296 (D.D.C. 2008) (finding that “the District ha[d] made a sufficient showing of good cause” to justify a protective order because discovery would “constitute an unwarranted invasion of the privacy of the police officers”).
. Megargee v. Wittman, No. CV F 06 0684, 2007 WL 2462097, at *2 (E.D. Cal. Aug. 27, 2007); see also Joseph F. Anderson Jr., Hidden from the Public by Order of the Court: The Case Against Government-Enforced Secrecy, 55. S.C. L. Rev. 711, 729 (2004) (quoting a former judge recalling that she “routinely signed orders because they didn’t create any work for us and they resolved issues and there was no one around asking that anything else be done”).
. Id. at 867 (“When one begins to consider the amount of discretion on the part of supervisors, managers, and agencies that can go into making up a police officer’s disciplinary file, it becomes easier to understand why police officers might legitimately worry about such records becoming public.”).
. See Rozema & Schanzenbach, supra note 17, at 225; Robert E. Worden, Moonsun Kim, Christopher J. Harris, Mary Anne Pratte, Shelagh E. Dorn & Shelley S. Hyland, Intervention with Problem Officers: An Outcome Evaluation of an EIS Intervention, 40 Crim. Just. & Behav. 409, 428–29 (2013) (finding that officers trained in skills to improve police and citizen interactions and reduce numbers of citizen complaints “might have been deterred from a proactive approach to police work”); Kim Michelle Lersch, Are Citizen Complaints Just Another Measure of Officer Productivity? An Analysis of Citizen Complaints and Officer Activity Measures, 3 Police Prac. & Rsch. 135, 142–46 (2002) (“Highly productive experienced officers still received fewer citizen complaints than their less experienced peers. It may be that experience combined with maturity and additional hours of training are important underlying predictors of the number of citizen complaints.”).
. Rachel Moran & Jessica Hodge, Law Enforcement Perspectives on Public Access to Misconduct Records, 42 Cardozo L. Rev. (forthcoming) (manuscript at 18), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3552012 [https://perma.cc/ERU3-DAMT].
. See, e.g., In re Avandia Mktg., Sale Pracs. & Prods. Liab. Litig., 924 F.3d 662, 672 (3d Cir. 2019) (“In both criminal and civil cases, a common law right of access attaches ‘to judicial proceedings and records.’” (quoting In re Cendant Corp., 260 F.3d 183, 192(3d Cir. 2001)).
. See, e.g., Minn. Stat. § 471.44 (2020) (“[E]very city, town, or county of this state employing sheriffs, police officers, or peace officers shall be required to furnish legal counsel to defend any sheriff, deputy sheriff, police officer, or peace officer . . . .”); N.J. Stat. Ann. § 40A:14-155 (West 2019) (“Whenever a member or officer of a municipal police department or force is a defendant in any action or legal proceeding arising out of and directly related to the lawful exercise of police powers in the furtherance of his official duties, the governing body of the municipality shall provide said member or officer with necessary means for the defense of such action or proceeding . . . .”).
. See, e.g., City of Albuquerque and Albuquerque Police Officers Ass’n, Collective Bargaining Agreement ¶ 23.1.1 (2014), https://www.lris.com/wp-content/uploads/contracts/albuquerque_nm_police.pdf [https://perma.cc/3DGJ-NHVY] (“Should a police officer be sued in a civil action for any allegation arising out of the course and scope of the officer’s employment, the City will defend and indemnify that officer . . . .”); Rahm Emanuel & Garry F. McCarthy, Agreement Between the city of Chicago Department of Police and the Fraternal Order of Police Chicago Lodge No. 7 §22.2 (2012–2017), https://static1.squarespace.com/static/559fbf2be4b08ef197467542/t/55a26d14e4b02ee06b2a8625/1436708116462/Chicagopolicecontract.pdf [https://perma.cc/49N9-786B] (“Officers shall have legal representation by the [e]mployer in any civil cause of action brought against an [o]fficer resulting from or arising out of the performance of duties.”); Agreement Between the City of Corpus Christi and the Corpus Christi Police Officers’ Association 59 (2015–2019), https://static1.squarespace.com/static/559fbf2be4b08ef197467542/t/567718151115e0704eaeb01b/1450645525733/Corpus+Christie+Police+Contract.pdf [https://perma.cc/8ASK-KGPK] (“The City will provide a legal defense to any police officer in a civil lawsuit, on account of any action taken by such police officer while acting within the course and scope of the police officer’s employment . . . .”); Master Agreement Between the City of Detroit and the Detroit Police Officers Association ¶ 27 (2014–2019), https://static1.squarespace.com/static/559fbf2be4b08ef197467542/t/55a26d54e4b02ee06b2a86ed/1436708180775/Detroit+police+contract.pdf [https://perma.cc/RX34-N35N] (“The City will provide legal counsel and pay any costs and judgments that arise out of lawsuits filed against Employees alleging any act committed while said Employee was in the good faith performance of his duties.”). For more examples of these contracts, see Check the Police, Police Contracts Database, https://www.checkthepolice.org/database [https://perma.cc/Y3GD-ZWEC].
. See, e.g., State Auto. Mut. Ins. v. Davis, No. 06-cv-00630, 2007 WL 2670262, at *2 (S.D.W. Va. Sept. 7, 2007); Raytheon Co. v. Indigo Sys. Corp., No. 07-cv-109, 2008 WL 437169, at *2 (E.D. Tex. Sept. 18, 2008); Meyer Goldberg, Inc. v. Fisher Foods, Inc., 823 F.2d 159, 161, 164 (6th Cir. 1987); Foltz v. State Farm Mut. Auto. Ins., 331 F.3d 1122, 1133 (9th Cir. 2003); SRS Techs., Inc. v. Physitron, Inc., 216 F.R.D. 525, 526 (N.D. Ala. 2003).
. Id. at *4 (applying the In re Roman Catholic Archbishop of Portland in Oregon balancing test by weighing “(1) whether disclosure will violate any privacy interests; (2) whether the information is being sought for a legitimate purpose or for an improper purpose; (3) whether disclosure of the information will cause a party embarrassment; (4) whether confidentiality is being sought over information important to public health and safety; (5) whether the sharing of information among litigants will promote fairness and efficiency; (6) whether a party benefitting from the order of confidentiality is a public entity or official; and (7) whether the case involves issues important to the public.”); see also In re Roman Catholic Archbishop of Portland in Or., 661 F.3d 417, 424 (9th Cir. 2011) (describing the Ninth Circuit test for balancing public and private interests to determine if the harm from disclosure justifies further confidentiality).
. About, Anti Police-Terror Project (2020), https://www.antipoliceterrorproject.org/about [https://perma.cc/8M8S-EDBU]; Patrisse Khan-Cullors, We Didn’t Start a Movement. We Started a Network., Medium.com (Feb. 22, 2016), https://medium.com/@patrissemariecullorsbrignac/we-didn-t-start-a-movement-we-started-a-network-90f9b5717668 [https://perma.cc/S7AR-LZLK] (describing how Patrisse Khan-Cullors, Alicia Garza, and Opal Tometi founded the “Black-centered political will and movement building project called #BlackLivesMatter.”).
. California Senate Advances “Right to Know” Police Transparency Legislation with Bipartisan Support, ACLU Southern California (May 30, 2018), https://www.aclusocal.org/en/press-releases/california-senate-advances-right-know-police-transparency-legislation-bipartisan [https://perma.cc/8P32-6VG2] (describing “communities desperate for answers about what really happened in police shootings” and “stories of people being killed and mistreated by police”).
. Press Release, Nancy Skinner, Cal. Sen., Dist. 9, Senator Skinner Introduces SB 1421 to Open Law Enforcement Records (Apr. 2, 2018), https://sd09.senate.ca.gov/news/20180402-senator-skinner-introduces-sb-1421-open-law-enforcement-records [https://perma.cc/SBW3-FM8G] (explaining the rationale for introducing SB 1421 as necessary to alleviate one of “the most secretive [confidentiality rules] in the country”).
. Perkins v. City of Modesto, No. 19-cv-00126, 2020 WL 4748273, at *5 (E.D. Cal. Aug. 17, 2020) (“[T]he Court notes that its finding is consistent with that of the State of California which, in balancing competing public and private interests, determined that files like the subject IA and personnel files should not be disclosed to the public.”).
. Sukey Lewis, Thomas Peele, Annie Gilbertson & Maya Lau, California Cops Are Withholding Public Records Dispute New Law Saying They Can’t, Desert Sun (July 5, 2019) https://www.desertsun.com/story/news/2019/07/05/california-police-still-withholding-misconduct-records-despite-law-sb-1421-shooting-sexual-dishonest/1659748001/ [https://perma.cc/KV73-BSUR].
. Mercury News & E. Bay Times Ed. Bds., Editorial, Becerra, Coddler of Bad Cops, Offers Hypocritical Post-Floyd Reforms, Mercury News (June 21, 2020), https://www.mercurynews.com/2020/06/21/editorial-ag-becerra-police-offer-hypocritical-post-floyd-reform-vows/ [https://perma.cc/6797-WN6A].
. Press Release, Nancy Skinner, Cal. Sen., Dist. 9, CA Lawmaker Introduces New Bills to Further Lift the Veil on Police Misconduct and “Reimagine” Policing (July 1, 2020), https://sd09.senate.ca.gov/news/20200701-ca-lawmaker-introduces-new-bills-further-lift-veil-police-misconduct-and-%E2%80%9Creimagine%E2%80%9D [https://perma.cc/P49G-ETZC].
. Thomas Peele, Bill Would Broaden and Speed Up Access to California Police Disciplinary Records and Make Complaints About Racist Cops Public, Mercury News (June 29, 2020), https://www.mercurynews.com/2020/06/29/bill-would-broaden-and-speed-up-access-to-police-disciplinary-records-and-make-complaints-about-racist-cops-public/ [https://perma.cc/2AM4-VURB].
. See S. 8496, 2019–2020 Leg., Reg. Sess. (N.Y. 2020), https://www.nysenate.gov/legislation/bills/2019/s8496 [https://perma.cc/EAM2-9QNN] (“Due to the interpretation of § 50-a, records of complaints or findings of law enforcement misconduct that have not resulted in criminal charges against an officer are almost entirely inaccessible to the public or to victims of police brutality, excessive use of force, or other misconduct.”); People v. Gissendanner, 399 N.E.2d 924, 929 (N.Y. 1979) (explaining that under section 50-a, criminal defendants seeking police personnel records for impeachment must provide “a clear showing of facts sufficient to warrant the judge to request records” and that the records will only be disclosed if a judge, after in camera review, “determines that the records contain matter that is relevant and material”).
. Stephanie Wykstra, The Fight for Transparency in Police Misconduct, Explained, Vox (June 16, 2020), https://www.vox.com/2020/6/16/21291595/new-york-section-50-a-police-misconduct [https://perma.cc/5D5D-CHMP].
. N.Y.C.L. Union v. N.Y.C. Police Dep’t, 118 N.E.3d 847, 849 (N.Y. 2018) (holding that police personal records “are exempt from disclosure pursuant to . . . Civil Rights Law § 50-a” because 50-a “requires that police officer personnel records be kept confidential”).
. State of N.Y. Dep’t of State Comm. on Open Gov’t, Annual Report to the Governor and State Legislature 3 (2014), http://web.archive.org/web/20210425092006/https://www.dos.ny.gov/coog/pdfs/2014AnnualReport.pdf.
. Press Release, Andrew M. Cuomo, Governor, N.Y. State, Governor Cuomo Signs ‘Say Their Name’ Reform Agenda Package (June 12, 2020), https://www.governor.ny.gov/news/governor-cuomo-signs-say-their-name-reform-agenda-package [https://perma.cc/5HJN-H944].
. Innocence Staff, In a Historic Victory, Governor Cuomo Signs Repeal of 50-A into Law, Innocence Project (June 6, 2020), https://innocenceproject.org/in-a-historic-victory-the-new-york-legislature-repeals-50-a-requiring-full-disclosure-of-police-disciplinary-records/ [https://perma.cc/7X8E-NXW7].
. Rushin, supra note 138, at 1222 (discussing how police union contracts limit accountability, for instance, where many contracts “call for the destruction of officer personnel records after a set period of time”).
. Officer Involved Shootings, Police Data Initiative, https://www.policedatainitiative.org/datasets/officer-involved-shootings/ [https://perma.cc/HNU9-ZUP9] (providing links to datasets for the Fairfax County Police Department; Jacksonville Sheriff’s Office; Los Angeles County Sheriff; Vermont State Police; Seattle Police Department; Tacoma Police Department; Atlanta Police Department; Bloomington Police Department; Cincinnati Police Department; Dallas Police Department; Denver Police Department; Hampton Police Division; Hartford Police Department; Indianapolis Police Department; Knoxville Police Department; Orlando Police Department; Philadelphia Police Department; Redondo Beach Police Department; Sparks Police Department; Portland Police Bureau; Norwich Police Department; Tucson Police Department; Louisville Metro Police Department; Charlotte-Mecklenburg Police Department; Springfield Police Department; and San Francisco Police Department).
. Use of Force, Police Data Initiative, https://www.policedatainitiative.org/datasets/use-of-force/ [https://perma.cc/2AAJ-HK8E] (providing links to datasets for the Austin Police Department; Baltimore Police Department; Bedford Police Department; Beloit Police Department; Bloomington Police Department; Cincinnati Police Department; Dallas Police Department; Indianapolis Police Department; Lincoln Department of Public Safety; New Orleans Police Department; North Bergen Police Department; Northampton Police Department; Norwich Police Department; Orlando Police Department; Portland Police Bureau; Seattle Police Department; South Bend Police Department; and University of Delaware Police Department).
. Complaints, Police Data Initiative, https://www.policedatainitiative.org/datasets/complaints/ [https://perma.cc/CG6T-MBMN] (providing links to datasets for the Tacoma Police Department; Bloomington Police Department; Indianapolis Police Department; North Bergen Police Department; Philadelphia Police Department; and South Bend Police Department).
. Eric Durkin, NYC to Publish Police Disciplinary Records Online, Politico (June 17, 2020), https://www.politico.com/states/new-york/albany/story/2020/06/17/nyc-to-publish-police-disciplinary-records-online-1293508 [https://perma.cc/FDJ9-ASSB].
. David Cruz, Police and Fire Unions Sue to Block Disciplinary Records from the Public, Gothamist (July 16, 2020), https://gothamist.com/news/police-and-fire-unions-sue-block-disciplinary-records-public [https://perma.cc/7ZNQ-VEDW].
. Tom McParland, 2nd Circuit Panel Stays Release of Police Disciplinary Records Pending Appeal, N.Y.L.J. (Sept. 17, 2020), https://www.law.com/newyorklawjournal/2020/09/17/2nd-circuit-panel-stays-release-of-police-disciplinary-records-pending-appeal/ [https://perma.cc/RAP6-XSSQ].
. Fatal Force, Wash. Post, https://www.washingtonpost.com/graphics/investigations/police-shootings-database/ [https://perma.cc/382H-V79Z]; The Counted: People Killed by Police in the US, Guardian, theguardian.com/us-news/ng-interactive/2015/jun/01/the-counted-police-killings-us-database [https://perma.cc/NVC2-QF68].
. John Kelly & Mark Nichols, Search the List of More Than 30,000 Police Officers Banned by 44 States, USA Today (Apr. 24, 2019), https://www.usatoday.com/in-depth/news/investigations/2019/04/24/biggest-collection-police-accountability-records-ever-assembled/2299127002/ [https://perma.cc/5ETG-C2CH].
. Catherine E. Lhamon, Letter of Transmittal of U.S. Comm’n on C.R., Police Use of Force: An Examination of Modern Policing Practices (2018) https://www.usccr.gov/pubs/2018/11-15-Police-Force.pdf [https://perma.cc/TRA9-L6ZM].
. Mark Tran, FBI Chief: ‘Unacceptable’ That Guardian Has Better Data on Police Violence, Guardian (Oct. 8, 2015), https://www.theguardian.com/us-news/2015/oct/08/fbi-chief-says-ridiculous-guardian-washington-post-better-information-police-shootings [https://perma.cc/WM3B-LMCR].
. Li Zhou & Ella Nilsen, The House Just Passed a Sweeping Police Reform Bill, Vox (June 25, 2020), https://www.vox.com/2020/6/25/21303005/police-reform-bill-house-democrats-senate-republicans [https://perma.cc/H7SS-29BF].
. 8 to Abolition: Abolitionist Policy Changes to Demand from Your City Officials (2020), https://static1.squarespace.com/static/5edbf321b6026b073fef97d4/t/5ee0817c955eaa484011b8fe/1591771519433/8toAbolition_V2.pdf [https://perma.cc/FCH3-MDTH].
. Campaign Zero, Model Use of Force Policy 9–10, https://static1.squarespace.com/static/55ad38b1e4b0185f0285195f/t/5deffeb7e827c13873eaf07c/1576009400070/Campaign+Zero+Model+Use+of+Force+Policy.pdf [https://perma.cc/BMC7-K6VQ].
. See Benjamin Lesser, Dan Levine, Lis Girion & Jami Dowdell, How Judges Added to the Grim Toll of Opioids, Reuters Investigates (June 25, 2019) https://www.reuters.com/investigates/special-report/usa-courts-secrecy-judges/ [https://perma.cc/5JBM-DJND].