Reforms are warranted, abolition is not.
Fair and effective policing is required to both reduce shootings and ensure public support. Essential to the task is accurately identifying individuals who are most at risk of being shooters, gunshot wound victims or both. But the NYPD’s Criminal Group Database, colloquially known as the “gang database,” is a frequent target of reformers, who say it mistakenly sweeps in too many people.
The critics’ claim, specifically, is that New York City’s gang database produces enhanced targeting of people based on questionable allegations of gang membership. These concerns have motivated City Councilwoman Althea Stevens and Public Advocate Jumaane Williams to propose a bill intended to abolish the database. They have questioned, without rigorous empirical evidence, the violence-reduction efficacy of the Criminal Group Database, noted a lack of transparency in gang membership designations and decried the fact that 99% of people in the database were Black or Hispanic.
Disproportionality alone, however, does not establish bias. It must be evaluated against victimization and offending patterns.
It is well known that shootings are not a democratic crime when it comes to race and place. In 2024, New York City shooting victims were primarily Black (67%) or Hispanic (29%). Fewer than 3% of shooting victims were Asian or Pacific Islander, and about 1% were white. Shooting suspects (65% Black, 33% Hispanic) and arrestees (64% Black, 32% Hispanic) have the same racial demographics. These racial characteristics closely match the almost exclusively nonwhite entries in the NYPD Criminal Group Database.
In October 2025, the NYPD gang database had only 8,563 individuals identified, representing only 0.1% of New York City’s 8.5 million residents. By one estimate, upward of 65% of New York City shootings involve gang members as victims, perpetrators or both. Importantly, shootings tend to occur in Black and Brown neighborhoods characterized by concentrated disadvantage and disinvestment that lead to limited access to safe and affordable housing, quality education, gainful employment, clean air and water, and transportation. Preventing recurring gun violence in these neighborhoods is clearly a matter of social justice. Investments need to be made to address the underlying structural conditions that generate persistent gun violence.
It is equally important to recognize that these neighborhoods need immediate relief from the associated economic, social and psychological burdens of fatal and nonfatal shootings. Effective policing is key to the success of any comprehensive effort to reduce shootings in the near term.
Police departments can reduce shootings in cities by focusing on the gangs, drug-selling organizations, street robbery crews and other criminally active groups involved in a disproportionate share of fatal and nonfatal shootings. Group-involved violence stems from personal disputes, drug-market business, ongoing gang rivalries and other conflicts. A study of more than 20 cities found that, on average, less than 1% of a city’s population was involved in gangs, drug crews and other criminally active groups, but were connected to more than 50% of a city’s shootings and homicides. The public safety implication is clear: Shootings can be controlled by focusing on and changing the violent behaviors of a small number of very risky people.
The NYPD has been a pioneer of “precision policing” strategies that focus law enforcement attention on the small number of violent criminals who harm communities and seek to work with neighborhood residents to increase public safety in affected neighborhoods. Rigorous evidence suggests this approach works in preventing gun violence. An evaluation of precision-policing gang “takedowns” in New York City housing projects between 2011 and 2019 found that shootings in and around public housing communities fell by approximately one-third in the first year after the enforcement action was launched. No evidence was found of substantial displacement, which is to say, there was no sign that gun violence just moved elsewhere. This was a meaningful reduction.
The NYPD maintains that the Criminal Group Database is critical to such strategic efforts and important to solving violent gun crimes. That tracks with how the Department uses the database: After gang-involved shootings, the NYPD rapidly deploys resources to prevent retaliation by rival gangs and advance investigations to quickly hold shooters accountable for their crimes. The NYPD associates these actions with achieving its record low of 688 shooting incidents in 2025.
More broadly, gang databases are part of an evidence-based gun violence reduction strategy known as “focused deterrence.” These programs have been shown to be effective in many cities. The goal of focused deterrence is first to focus on high-risk individuals and then stop further violence through a strategic blend of law enforcement, social services and community-based action. For example, after a gang shooting, officers and their partners need to identify gang members who are likely to retaliate and who are likely to be shot. Knowing which gangs are involved and their membership is necessary to do this.
While no system is free from error, the research literature suggests that formalized gang databases with clear inclusion criteria and oversight may reduce — rather than exacerbate — the risks associated with purely individual, intuition-based determinations of gang involvement.
People living in high-crime neighborhoods, particularly young men of color, bear a disproportionate policing burden. Though these young men are overwhelmingly uninvolved in crime, they are far more likely to be stopped and questioned by police than youth living in more affluent neighborhoods, certainly an infringement of their civil liberties. The NYPD remains under the supervision of a court-appointed monitor to ensure that its stop-and-frisk practices are lawful and not generating excessive racial disparities. Since 2011, NYPD stops have fallen by roughly 97%, sharply reducing the burden young men of color experience. But stop-and-frisks continue in smaller numbers, and the burden disproportionately falls on Black and Latino men.
A key question is whether abolishing the NYPD Criminal Group Database would lessen the undeserved policing burden on youth of color or, rather, unintentionally increase it. We know of no academic research that has studied this question for New York City or any other city.
Related academic work is possibly suggestive. An important study examined the effect of “banning the box,” that is restricting employers from asking about job applicants’ criminal histories. The authors used an audit study where fictitious black and white individuals applied online for jobs. Prior to ban-the-box implementation, they found that employers that asked about criminal history were 63% more likely to call applicants with no record. However, they also found that white applicants received 7% more callbacks before ban-the-box. After its implementation, this gap increased to a whopping 43%. In short, when employers were not able to ask about criminal history, they were far more likely to discriminate against Black candidates. Other studies have found similar patterns of racial discrimination in employment outcomes when prior criminal histories were not available to prospective employers.
Boston’s experience is also informative. In the early 1990s, and prior to that, Boston did not have an effective, much less focused, deterrence strategy for dealing with gang violence. The result was twofold: With the introduction of crack, the city experienced a period of uncontrolled gang violence; Black youth were indiscriminately and frequently stopped by police. The latter happened most dramatically after the Carol Stuart murder in 1989, when young Black men, no matter who they were or their circumstances, were, in the language of the police, “tipped upside down.” The Boston police had no idea who might have been responsible.
Similarly, without its Criminal Group Database, the NYPD would not have a reservoir of timely intelligence on gangs and gang dynamics to focus its deployments and investigations in the wake of shootings. The resulting harm could be twofold: A diminished capacity to control gun violence in traumatized neighborhoods and a broadening of police enforcement activities to identify and apprehend shooters.
Though we can’t say for certain, the evidence suggests that banning the NYPD’s gang database could have just the opposite effect that reformers intend. Without having specific knowledge about who is gang-involved, officers may well stop a far broader set of individuals, increasing the policing on innocent individuals, particularly young people of color.
Decades of psychological research on judgment and decision-making raise important concerns about relying solely on individual officers’ discretion to determine gang membership. When decisions must be made under time pressure, stress and incomplete information — conditions that typify street-level policing — human judgment is known to rely heavily on cognitive shortcuts and heuristics. A large body of research demonstrates that such circumstances increase susceptibility to confirmation bias and stereotype-based reasoning, even among well-intentioned professionals. Experimental studies further show that racial stereotypes can shape perceptions of criminality and threat when objective criteria are absent or unclear. Organizational research on street-level bureaucracy also finds that discretionary decision-making without standardized rules leads to inconsistency and “drift” across officers and units.
In contrast, extensive evidence from psychology and decision science indicates that structured, rule-based or actuarial classification systems consistently outperform unaided professional judgment in risk assessment and categorization tasks, producing more reliable and equitable outcomes. While no system is free from error, this literature suggests that formalized gang databases with clear inclusion criteria and oversight may reduce — rather than exacerbate — the risks associated with purely individual, intuition-based determinations of gang involvement.
None of this is to say that the NYPD’s gang database is beyond reproach. Research makes clear that gang databases suffer from vague inclusion criteria, weak oversight, a lack of due process and potential law enforcement biases about who is involved in gang activity. A recent study found young men of color were overrepresented in the California gang database (CalGang) and raised questions as to whether these disparities could be explained by legal factors and whether the database properly balanced public safety value with individual rights. The NYPD has faced similar questions. In response to a 2023 Office of the Inspector General report, the Department tightened its inclusion and removal criteria. It is certainly possible that the NYPD Criminal Group Database could benefit from further examination and reforms. A more focused and targeted database would increase police efficiency, lower gun violence rates even further and lessen the civil liberties burden — the primary benefits of a focused policing approach. But those who champion eliminating the gang database entirely might harm the very individuals they say they want to protect.