research about crime statistics

An official website of the United States government

Here’s how you know

Official websites use .gov A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Find crime statistics

Federal, state, and local law enforcement agencies collect data about crime. Find crime statistics around the U.S. using the FBI’s Crime Data Explorer.

Use the Crime Data Explorer to find statistics about different types of crime nationally or in your state, county, or town.

LAST UPDATED: February 1, 2024

Have a question?

Ask a real person any government-related question for free. They will get you the answer or let you know where to find it.

New Data Shows Violent Crime Is Up… And Also Down.

Property crime and violence against young people are both up, recent federal data shows, but other crime trends are murkier..

T wo key crime reports released by the Justice Department this fall reveal a changing crime landscape, even when they diverge on year-over-year trends. Property crime rose in significant ways for the first time in years. Violent crime against young people doubled. As usual, most crimes go unreported. And as a major election season looms in 2024, the deviation between the reports on recent trends in violent crime could be read selectively to score political points.

The FBI’s crime reporting program and the Bureau of Justice Statistics’ crime victimization survey are generally seen as the Justice Department’s two major pillars of national crime statistics. While the FBI tries to collect crime data directly from more than 18,000 police agencies through its Uniform Crime Reporting program, the National Crime Victimization Survey (NCVS) interviews 150,000 families across the country. The NCVS asks questions that capture crimes that were not reported to police as well as ones that were, which are then weighted in order to estimate crime victimization for more than 120 million U.S. households.

Both reports show a return to pre-pandemic levels of violent crime since the COVID-19 pandemic broke out three years prior, and a similar pattern: The violent crime rate exhibited significant fluctuations during the pandemic, but by 2022, it had returned to pre-pandemic levels.

Violent crime and victimization rates return to pre-pandemic level

During the pandemic, the country saw a significant uptick in homicides, shootings and aggravated assaults , which likely contributed to the increase seen in data reported by law enforcement. At the same time, violent crime victimizations — which included less acute forms of violent crime, such as assaults without weapons that did not lead to serious injuries — dipped. These violent crime victimizations mostly went unreported.

Murders are a good measurement for the most serious violent crimes, partially because they are almost always reported to police. According to the FBI’s crime statistics, the number of murders dropped by 6.1% from 2021 to 2022, but is still higher than where it was prior to the pandemic.

For some observers, the jump in serious violent crime was always bound to stabilize, as the shock from the early pandemic wore off. “Violent crime rates had been trending down for at least a decade. And when the pandemic hit, unsurprisingly, with this once-in-a-century event, you saw things shift,” said Kim Foxx, state’s attorney for Cook County, Illinois. “As we are coming out of the pandemic, the fact that we are now seeing those numbers trend downward again is not surprising to me.”

Despite the long-term trends, the two reports differ more than ever before on the year-over-year change in violent crime.

Since the methodologies for these two reports are different, it’s not unusual for them to show different trends on a particular year, said Richard Rosenfeld, a criminology professor at the University of Missouri, who recently wrote about this divergence for the Council on Criminal Justice . Over the past 30 years, these two reports’ difference in year-over-year violence trends has never been as big as it was last year.

In 2021, the FBI changed how it collects data from police departments, and as a result, that year’s crime data missed nearly 40% of police agencies . Bureau analysts estimated the missing data with statistical modeling, but the change led to the most incomplete picture of national crime since the FBI began collecting data in the 1930s, which created confusion on how crime trends changed . Last year, the FBI reversed the change and revived the previously-retired data collection system. They also gave agencies that didn’t submit data for 2021 a chance to submit their data retrospectively. Nearly 2,500 agencies took the FBI’s offer and submitted crime data through the old system for 2022, but it’s unclear how many did for 2021.

Experts said the lingering effect of that transition could be why the 2021-2022 trend is unreliable: If the 2021 crime data remains incomplete, it is difficult to compare it with the 2022 data.

These data gaps and disagreements create more space for politicians to spin unsubstantiated, murky narratives. When Florida Gov. Ron DeSantis announced his run for president, for example, he touted that Florida’s crime rate has reached a record-low under his administration. But he failed to mention that he relied on a crime rate estimation that was missing data from about half of the state’s law enforcement agencies , which policed 40% of the state’s population.

The FBI said it cannot address the difference between its crime data and the BJS’ victimization survey because it “cannot comment on another agency's report.”

In an interview, BJS statisticians said there’s no single factor that can neatly explain the divergence — the victimization survey and police statistics are designed to complement each other, and often reflect different aspects of criminal justice and victim’s issues.

“It would be nice to know what's happening with violent crime rates,” said Richard Rosenfeld. “But having two contrasting reports both coming out of the Justice Department enables politicians, or anyone else who has a horse in the race, to just cherry-pick the estimate that fits best with their [priorities] and ignore the other.”

While the Justice Department’s two reports diverged on recent violent crime trends, both showed an uptick in property crime from 2021 to 2022: The FBI’s crime data showed a 7.1% increase in property crime, while the victimization survey showed a 14.5% jump.

Property crime and victimization increased in 2022

The FBI's crime data showed a 7% increase in property crime from 2021 to 2022, which came after a decades-long downward trend. The Justice Department's crime victimization survey shows similar trends. Experts said motor vehicle theft and soaring inflation could be behind the increases.

In both reports, a sharp increase in motor vehicle theft and larceny were the main drivers for the increase in property crime. The former, criminologists said, can be partially attributed to the “Kia Boyz,” whose videos on how to steal Hyundai and Kia vehicles went viral on social media platforms .

In Chicago, the number of stolen Kia and Hyundais jumped 35 times over a couple of months — from 45 cars stolen in May 2022 to more than 1,400 cars stolen that October, according to data compiled by Vice . Data from other major cities, like Los Angeles, Denver and Milwaukee, showed similar trends.

Relentless inflation can also lead to more property crimes like theft and larceny , Rosenfeld said. Last summer, the inflation rate reached 9.1% — the highest in decades — which led many people to trade down on where they shop, and some traded down to purchasing stolen goods. While the inflation rate has since dropped , the 2022 crime data doesn’t reflect its effect yet.

Another trend that both Justice Department reports show is that young people experienced more violence in 2022.

The FBI’s crime data shows that while fatal and non-fatal gun violence against adults declined in 2022, both increased by more than 10% for young people who are under the age of 18. Similarly, the crime victimization survey shows that the violent victimization rate for people between the ages of 12 and 17 doubled last year — from 13 to 27 violent crimes per 1,000 youth, representing the age group that saw the biggest increase in violent victimization.

Researchers have theories on why violence against young people jumped, but caution that little is definitive. Kim Smith, the Director of Programs at the University of Chicago Crime Lab, points to school enrollment as a factor that might be one piece of the puzzle. Nationwide, school enrollment rates fell during the pandemic and are still yet to recover. The Crime Lab has found that in Chicago, a startling 90% of youth shooting victims were not active or enrolled at school at the time they were shot. “Education is the most protective factor against future violence involvement,” Smith said.

Other analysts have pointed to the increase in gun purchases since the beginning of the pandemic, and the sense of danger that many young people feel in their neighborhoods. “We had an influx of guns during the COVID shutdown, and an enormous amount of guns that entered the country and ended up on the streets,” said Jamila Hodge, executive director of the national violence prevention group Equal Justice USA. “That influx is reflected in the numbers of rising violence against young people — it's access to guns.”

While the FBI continues trying to improve the national crime statistics, most property and violent crimes are not reported to the police, the victimization survey shows.

The survey asks victims if they reported crimes they experienced to the police. About 41% of violent crime victims said they reported the incident to the police. But the actual number of violent crimes police recorded is much lower.

More than half of crimes never got reported

The Justice Department's crime victimization survey shows millions of violent and property crimes were never reported to the police — often because the victim reported the incident to another authority, like a school official, or because they don't think police would take the matter seriously. Some crimes that are reported to the police never end up in the police report, and would not be counted by the FBI's crime stats.

While the BJS estimated more than 6 million violent crime victimizations, and estimated more than 2 million of those incidents were reported to the police, the FBI’s crime statistics only recorded 1.2 million incidents. In some incidents, a reported crime is not recorded by the police.

Our reporting has real impact on the criminal justice system

Our journalism establishes facts, exposes failures and examines solutions for a criminal justice system in crisis. If you believe in what we do, become a member today.

Weihua Li Twitter Email is a data reporter at The Marshall Project. She uses data analysis and visualization to tell stories about the criminal justice system. She studied journalism and comparative politics at Boston University and graduated from Columbia University with a master's degree in data journalism.

Jamiles Lartey Twitter Email is a New Orleans-based staff writer for The Marshall Project. Previously, he worked as a reporter for the Guardian covering issues of criminal justice, race and policing. Jamiles was a member of the team behind the award-winning online database “The Counted,” tracking police violence in 2015 and 2016. In 2016, he was named “Michael J. Feeney Emerging Journalist of the Year” by the National Association of Black Journalists.

Stay up to date on our reporting and analysis.

You are using an outdated browser. Please upgrade your browser to improve your experience.

Suggested Results

Informed citizens are our democracy’s best defense..

We respect your privacy .

• Research & Reports

Understanding the FBI’s 2021 Crime Data

Changes to the way the FBI reports national crime data may significantly complicate public understanding of recent crime trends.

• Accurate Crime Data

Click here for the latest FBI crime statistics >>

Policymakers, journalists, and the public look to data released annually by the Federal Bureau of Investigation (FBI) to understand state and national crime trends. Recently, the FBI rolled out a plan to modernize its reporting of crime data. Here’s what these changes mean for our understanding of recent crime trends.

What data does the FBI collect, and why does it matter? 

For  nearly a century , the FBI has collected data on offenses known to law enforcement from state and local agencies. Through its Uniform Crime Reporting program, the FBI aggregates these reports, applies quality control standards, uses  estimates  to fill gaps in their information, and then publishes both the raw data and trend analyses.

Researchers and policymakers have come to rely especially on an annual Uniform Crime Reporting publication,  Crime in the United States , for understanding the previous year’s crime trends. This report generally contains high-level tables tracking important statistics like state, regional, and national violent crime rates. Each year’s annual report tends to come out roughly nine months after the end of the year it describes, although the summary of 2021 crime data is coming out slightly later.

How has the FBI historically collected and reported this data? 

The FBI historically used two systems to collect crime data. The  older of the two , the Summary Reporting System,  tracked  “monthly counts of the number of crimes known to law enforcement.” While easy to use, the system tended to gloss over important nuances. Among other things, it only counted the most serious offense in an incident, applying a “hierarchy rule” to determine which offenses were more serious than others. The system also focused on a limited number of types of crime and failed to capture details like the number of people involved.

More recently, the FBI and reporting agencies began using the National Incident Based Reporting System. This system tracks information  in much greater detail  than the Summary Reporting System and covers  many additional types of crimes . The National Incident Based Reporting System also abandons the hierarchy rule, allowing law enforcement to report multiple offenses  in a single incident . Consider a hypothetical robbery where the victim, injured in the attack, dies of his injuries. Under the Summary Reporting System, this incident would be recorded solely as a murder — the robbery would vanish from the FBI’s data. Under the National Incident Based Reporting System, both the robbery and murder would be counted along with considerable details about both offenses.

What is changing about crime data collection and reporting?

For years, the FBI accepted data in both formats. On January 1, 2021, however, the FBI  stopped accepting data  through the Summary Reporting System.

Unfortunately, despite the advantages of the newer National Incident Based Reporting System, many state and local law enforcement agencies have yet to make the switch. Law enforcement agencies covering  just over half of the population  reported a full year’s worth of data to the FBI in 2021. By comparison, the FBI’s recent reports have been based on data from agencies covering  upwards of 95 percent  of the population. The  time and expense  involved in updating outdated government computer systems are likely partly to blame for the delayed adoption of the new system.

The FBI recently started releasing some crime data on a quarterly basis. Participation by police departments has been relatively low so far but is expected to increase as more agencies make the switch. These quarterly reports, available through an  online data dashboard , should eventually become a more reliable, regular source of crime data.

How will these changes affect crime data reported for 2021? 

With so many agencies failing to report a full year of data for 2021, this year’s annual crime data release will have significant blind spots.

We know that the agency’s annual  Crime in the United States  report  will feature  “state-level data” and “a trend study comparing 2020 and 2021 data,” the latter drawing on partial data supplemented by estimates. It  may also contain  national estimates of crime trends — that is, data on whether rates of murder and other offenses rose or fell at the national level. If the FBI ultimately publishes these national estimates, they will likely be expressed with “ confidence intervals ” to indicate uncertainty.

Some major cities will be absent from the data too. San Francisco, for example, does not plan to complete its transition to National Incident Based Reporting System  until 2025 , meaning it will be absent from FBI crime data until then. (Data on San Francisco crime trends will, of course, remain available directly from the city, but not in the new format.) And just  13 percent  of law enforcement agencies in New York State reported a full year of data to the FBI in 2021. New York City was not among them. On the other hand, some regions of the country with a high rate of National Incident Based Reporting System adoption —  like Michigan, Texas, and Virginia  — will have robust data available, allowing researchers to use data from the FBI to study crime trends in some states in great detail.

What challenges does this present for our understanding of crime trends? 

It will probably be impossible to speak of a precise “national” murder rate or “national” violent crime rate for 2021. Confidence intervals may make it difficult to determine whether the rates of some offenses rose or fell. Policymakers will have to exercise great care when using this limited data.

Some challenges will also remain even after National Incident Based Reporting System adoption is complete. Because it allows the reporting of multiple “offenses” per “incident,” comparing data from before and after the transition may create a false appearance of an increase in some offenses. Recall the example of the deadly robbery, above: the Summary Reporting System would count it as only one crime — a murder — while the National Incident Based Reporting System would count it as  two  violent crimes from that incident — a murder and a robbery. A  2019 study  of early adopters of the National Incident Based Reporting System suggests roughly 90 percent of incidents involve just one offense, and increases in offense counts related to the transition may be in the range of just 2 to 5 percent. But that could still be enough to affect perceptions of whether crime is “rising” or “falling.”

What about other sources of crime data for 2021?

Critically, the FBI is not the only source of information on public safety. For one, the Centers for Disease Control provides 3-month and 12-month  national homicide trends . This data, which uses a different methodology to track deaths but  aligned with the FBI estimates in 2020 , shows homicides rising at a slower rate than in 2020.

Researchers and think tanks have also produced their own crime analyses. Most notably, the Council on Criminal Justice publishes  regular reports on crime trends , including a  collection of year-end 2021 data  from 27 cities. Its findings track the CDC’s, suggesting that murder rose in 2021 but at a much slower pace than the previous year. Trends were mixed across other categories of crime, which are more challenging for private organizations to study. Crime data analyst Jeff Asher also tracks year-to-date murders in select cities on  a data dashboard , and his tallies currently show major-city murders declining in 2022.

Lastly, while the FBI tracks offenses known to law enforcement agencies, it cannot account for crimes that are never reported. The National Criminal Victimization Survey, published by the Bureau of Justice Statistics, seeks to account for the difference using a national survey to estimate the rate at which people  experience  non-fatal crime and violence. This survey shows rates of non-fatal violent crime  declining in 2020  and  increasing very slightly  in 2021.

Related Resources

U.S. Crime Rates and Trends — Analysis of FBI Crime Statistics

The FBI’s latest report on crime trends includes 2022 data for national, regional, and local levels.

Myths and Realities: Understanding Recent Trends in Violent Crime

Debunking the Myth of the ‘Migrant Crime Wave’

Data does not support claims that the United States is experiencing a surge in crime caused by immigrants.

Violent Crime Is Falling Nationwide — Here’s How We Know

Preliminary FBI data, however imperfect, confirms a sharp downward trend in crime, undercutting attempts to blame criminal justice reform for pandemic-era spikes in violence.

Why Inclusive Criminal Justice Research Matters

The hidden toll of new york city's misdemeanor system, myth vs. reality: trends in retail theft, fact-checking trump’s speech on crime and immigrants, trump misleads about crime and public safety, again, informed citizens are democracy’s best defense.

• Society ›
• Crime & Law Enforcement

Crime in the U.S.

State by state comparison, is crime increasing, key insights.

Detailed statistics

Most dangerous cities in the U.S. 2022, by violent crime rate

Least peaceful U.S. metropolitan areas 2012

Metropolitan areas - crime U.S. 2020, by type

Editor’s Picks Current statistics on this topic

Violent crime.

United States crime rate in 2022, by type of crime

Metropolitan areas - crime rate U.S. 2022, by type

Public assessment of crime as a serious issue U.S. 2000-2021

Further recommended statistics

• Basic Statistic Number of committed crimes in the U.S. 2022, by type
• Basic Statistic United States crime rate in 2022, by type of crime
• Basic Statistic Reported violent crime rate in the U.S. 2022, by state
• Premium Statistic Metropolitan areas - crime rate U.S. 2022, by type
• Basic Statistic Public assessment of crime as a serious issue U.S. 2000-2021
• Basic Statistic U.S. crime rate trend perception 1990-2022

Number of committed crimes in the U.S. 2022, by type

Number of committed crimes in the United States in 2022, by type of crime

Crime rate in the United States in 2022, by type of crime (per 100,000 inhabitants)

Reported violent crime rate in the U.S. 2022, by state

Reported violent crime rate in the United States in 2022, by state (per 100,000 of the population)

Crime rate in metropolitan areas in the United States in 2022, by type (per 100,000 inhabitants)

Percentage of the U.S. population that assesses crime as a serious problem from 2000 to 2021, nationwide and local

U.S. crime rate trend perception 1990-2022

Is there more crime in the U.S. than there was a year ago, or less?

• Basic Statistic U.S.: reported murder and nonnegligent manslaughter cases 1990-2022
• Basic Statistic USA - reported murder and nonnegligent manslaughter rate 1990-2022
• Basic Statistic U.S.: reported forcible rape cases 1990-2022
• Basic Statistic USA - reported forcible rape rate 1990-2022
• Basic Statistic U.S.: reported robbery cases 1990-2022
• Basic Statistic USA - reported robbery rate 1990-2022
• Basic Statistic U.S.: reported cases of aggravated assault 1990-2022
• Basic Statistic USA - reported aggravated assault rate 1990-2022

U.S.: reported murder and nonnegligent manslaughter cases 1990-2022

Number of reported murder and nonnegligent manslaughter cases in the United States from 1990 to 2022

USA - reported murder and nonnegligent manslaughter rate 1990-2022

Reported murder and nonnegligent manslaughter rate in the United States from 1990 to 2022 (per 100,000 of the population)

U.S.: reported forcible rape cases 1990-2022

Number of reported forcible rape cases in the United States from 1990 to 2022

USA - reported forcible rape rate 1990-2022

Reported forcible rape rate in the United States from 1990 to 2022 (per 100,000 of the population)

U.S.: reported robbery cases 1990-2022

Number of reported robbery cases in the United States from 1990 to 2022

USA - reported robbery rate 1990-2022

Reported robbery rate in the United States from 1990 to 2022 (per 100,000 of the population)

U.S.: reported cases of aggravated assault 1990-2022

Number of reported cases of aggravated assault in the United States from 1990 to 2022

USA - reported aggravated assault rate 1990-2022

Reported aggravated assault rate in the United States from 1990 to 2022 (per 100,000 of the population)

Property Crime

• Basic Statistic U.S.: reported larceny-theft cases 1990-2022
• Basic Statistic USA - reported larceny-theft rate 1990-2022
• Basic Statistic U.S.: reported burglary cases 1990-2022
• Basic Statistic USA - reported burglary rate 1990-2022
• Basic Statistic U.S.: reported motor vehicle theft cases 1990-2022
• Basic Statistic U.S.: reported motor vehicle theft rate 1990-2022

U.S.: reported larceny-theft cases 1990-2022

Number of reported larceny-theft cases in the United States from 1990 to 2022

USA - reported larceny-theft rate 1990-2022

Reported larceny-theft rate in the United States from 1990 to 2022 (per 100,000 of the population)

U.S.: reported burglary cases 1990-2022

Number of reported burglary cases in the United States from 1990 to 2022

USA - reported burglary rate 1990-2022

Reported burglary rate in the United States from 1990 to 2022 (per 100,000 of the population)

U.S.: reported motor vehicle theft cases 1990-2022

Number of reported motor vehicle theft cases in the United States from 1990 to 2022

U.S.: reported motor vehicle theft rate 1990-2022

Reported motor vehicle theft rate in the United States from 1990 to 2022 (per 100,000 of the population)

Crime Clearance & Arrests

• Basic Statistic Clearance rate - crime clearance rate U.S. 2022, by type
• Basic Statistic Regional distribution - crime clearance rate in the U.S. 2022
• Basic Statistic USA - number of arrests for all offenses 1990-2022
• Basic Statistic USA - arrest rate for all offenses 1990-2022
• Basic Statistic USA - number of arrests for violent offenses 1990-2022
• Basic Statistic USA - arrest rate for violent offenses 1990-2022
• Basic Statistic USA - number of arrests for property offenses 1990-2022
• Premium Statistic USA - arrest rate for property offenses 1990-2022

Clearance rate - crime clearance rate U.S. 2022, by type

Crime clearance rate in the United States in 2022, by type

Regional distribution - crime clearance rate in the U.S. 2022

Crime clearance rate in the United States in 2022, by region

USA - number of arrests for all offenses 1990-2022

Number of arrests for all offenses in the United States from 1990 to 2022

USA - arrest rate for all offenses 1990-2022

Arrest rate for all offenses in the United States from 1990 to 2022 (arrests per 100,000 people)

USA - number of arrests for violent offenses 1990-2022

Number of arrests for violent offenses in the United States from 1990 to 2022

USA - arrest rate for violent offenses 1990-2022

Arrest rate for violent offenses in the United States from 1990 to 2022 (arrests per 100,000 of the population)

USA - number of arrests for property offenses 1990-2022

Number of arrests for property offenses in United States from 1990 to 2022

USA - arrest rate for property offenses 1990-2022

Arrest rate for property offenses in the United States from 1990 to 2022 (per 100,000 of the population)

Further reports

Get the best reports to understand your industry.

Mon - Fri, 9am - 6pm (EST)

Mon - Fri, 9am - 5pm (SGT)

Mon - Fri, 10:00am - 6:00pm (JST)

Mon - Fri, 9:30am - 5pm (GMT)

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• SAGE - PMC COVID-19 Collection

Beyond Policing: The Problem of Crime in America

José luis morín.

1 John Jay College of Criminal Justice, CUNY, New York City, NY, USA

Photo by Francois Polito. Sculpture by Carl Fredrik Reuterswärd.

In 2020, the United States experienced the sharpest one-year rise in homicides on record. 1 In 2021, hate crimes also surged to their highest level in twelve years, with the largest increases being anti-Black crimes followed by anti-Asian crimes. 2 Pundits and politicians on the right have been quick to cite bail reform and “defunding” of police as reasons for the national rise in crime. Yet, the best available evidence points to other causes—among them, the massive social and economic dislocation resulting from the Covid-19 pandemic and the nationwide proliferation of guns along with the spread of racial and ethnic hatred and the violence it has roused.

While violent crime today [is] much lower . . . than in 1991, when [it] reached its highest point in recent history, public anxiety about crime is high.

While violent crime today registers at much lower levels than in 1991, when violent crime reached its highest point in recent history, 3 public anxiety about crime is high. Yet, more police and a redoubling of get-tough measures, however alluring, have not proven to be as effective as they appear. An examination of what is not driving the recent spike in crime as well as what probably is driving it—and revisiting the role that policing and the criminal justice system has played in U.S. society in reproducing racial, social, and economic inequalities—may move us closer to arriving at effective public safety solutions.

Starting with What Is Not: Bail Reform and Defunding the Police

The purpose of bail is to “provide reasonable assurance of court appearance or public safety,” 4 but, a 2022 briefing report by the U.S. Commission on Civil Rights suggests that the current cash bail system is also associated with producing deleterious racial disparities and economic inequities that undermine the presumption of innocence and worsen public safety. 5 The Commission report points out that low-income persons and people of color are disproportionately detained as a result of their inability to make bail, and persons of color are more often assigned higher bail amounts and considered more “dangerous” than whites. 6 To persons jailed simply because they could not afford bail, jail can result in severe trauma: loss of employment, housing, custody of a child; and economic hardship. 7 The think tank, Prison Policy Initiative, issued a report in 2016 documenting that cash bail “perpetuates an endless cycle of poverty and jail time.” 8 Nevertheless, with crime rates on the rise in 2020, bail reform became a political cudgel, and New York State’s law became a focal point of harsh condemnation from the GOP and conservative media outlets nationwide. But analyses of bail reform show no clear link between bail reform and spikes in crime. 9

To reduce unnecessary pretrial detention and ameliorate the harms associated with cash bail, New York State passed a bail reform law in 2019, ending cash bail for certain misdemeanors and most non-violent felony cases. Changes to the original law in 2020 and 2022 gave judges the ability to impose cash bail in more situations. 10 To date, research on the law shows no significant impact on crime rates. One study by the Times Union of Albany found that, of almost 100,000 cases, only a minimal number (2 percent) of individuals faced rearrest for a violent felony. As a result, as many as 80,000 people may have avoided incarceration while posing no documented threat to public safety. 11 And, looking at the national picture, the Brennan Center for Justice, a progressive law and public policy institute, points out that crime surged nationally, even in states that did not enact bail reform. 12

Another report—this one from the Office of New York City Comptroller Brad Lander—covered three calendar years and revealed that “pretrial rearrest rates remained nearly identical pre- and post-bail reform.” 13 The Comptroller’s report also warned that rollbacks to New York State’s bail reforms would “syphon money” from low-income communities. Indeed, families unable to make cash bail often turn to for-profit bail bond companies that require a non-refundable premium of 10 to 15 percent, even if no wrongdoing is found. Some form of collateral—such as a car or house—is also required. As the bail bond industry has become increasingly lucrative, growing numbers of indigent persons and their families face steep financial risks. 14 Critics of bail reform, by contrast, have produced little to no empirical evidence to support their position. Outspoken in its condemnation of bail reform, the New York Police Department has fallen short in backing up its assertions that bail reform was causing increases in gun violence. In 2021, when challenged by Albany legislators, NYPD Commissioner Dermot Shea, failed to provide any hard data to support his contention that bail reform is driving up crime. In the end, he was forced to retract his statements. 15

The term “defunding the police” has been variously understood. For the purpose of this discussion, I define the term not as a movement to eliminate police budgets, but as a call to lessen encounters with police by shifting funds away from aggressive and militarized forms of policing toward social services—such as mental health, addiction, education, and housing. In its most literal meaning, “defunding the police” is frequently cited as a reason for the surge in crime. As with bail, hard evidence to support this allegation has not materialized. Of twelve Democratic-led cities (including Austin, Louisville, Rochester, and St. Paul) cited by Republicans as examples of where crime purportedly rose due to police defunding, criminal justice scholars find no discernible link between defunding and crime. In fact, not all twelve cities defunded police; most did not substantially reduce police funding, and some actually increased their police budgets. 16

Funding quality educational programs, by comparison, has proven to be successful in diminishing crime. 17 For example, two studies—one in North Carolina and one in Michigan—showed that increased expenditures on primary schools helped to reduce adult crime by improving student academic success, which in turn provides a greater opportunity for socio-economic mobility. 18

Examining What Is: Covid-19, Guns, and the Rise in Hate

Prior to the Covid-19 pandemic, crime rates were relatively low. As the graph in Figure 1 demonstrates, the rate of violent crime offenses declined from a peak in 1991 of 758.2 per year to 398.5 per year in 2020. 19 The rate of homicide over the same period also dropped significantly, from its highest level in 1991 compared with 2020. 20 But the crime rate shows an uptick with the onset of the Covid-19 pandemic in 2020.

Rate of violent crime offenses by population in the United States: 1985-2020.

Source: Federal Bureau of Investigation, “Trend of Violent Crime from 1985 to 2020,” Crime Data Explorer, accessed September 19, 2022, https://crime-data-explorer.fr.cloud.gov/pages/explorer/crime/crime-trend .

Note. Rate per 100,000 people, by year.

The pandemic is widely understood as the cause of immense social and economic dislocation, disproportionately disadvantaging children, communities of color, immigrants, LGBTQIA+ (lesbian, gay, bisexual, transgender, queer or questioning, intersex, asexual, and more) youth, and persons with disabilities. 21 The pandemic also exposed and aggravated deeply entrenched inequities in health care, poverty, education, housing, and racial segregation. Its impact on mental health and psychosocial well-being, substance abuse, and domestic violence has become a focus of attention in relation to the rise in crime. All these factors are related to the rise in crime.

As gun violence became a major driver of crime nationwide, hate crimes also spiked. During the pandemic, reports of hate crimes reached a twelve-year high.

The proliferation of guns and gun violence nationwide appears to have contributed greatly to the spike in homicides. Sharp increases in gun purchases coincided with the start of the pandemic in 2020 and continued well into 2021. 22 The increased supply of guns as well as the types of guns—high-powered semi-automatic weapons, for instance—has been linked to the surge in gun violence. 23 Criminologists Philip J. Cook and Jens Ludwig deem gun violence and the fear of gun violence as devastating to the lives of children and families around the country, most especially in low-income neighborhoods and communities of color. In their estimation, public safety begins with addressing the needs of communities most vulnerable to gun violence, and that includes investments in social policies, such as summer jobs for teens, cleaning vacant lots, and spending more on social programs—all of which have been shown to reduce homicide rates. 24

As gun violence became a major driver of crime nationwide, hate crimes also spiked. During the pandemic, reports of hate crimes reached a twelve-year high (see Figure 2 ). 25 While anti-Black incidents topped the list of hate crimes based on race, people of Asian descent experienced a steep rise in anti-Asian violence and crime, with a 169 percent increase in reports of anti-Asian hate crimes in fifteen of America’s largest cities and counties in the first quarter of 2021 when compared with the first quarter of 2020. 26

Hate crimes in the United States: 1995-2020.

Source: Hate Crime in the United States Incident Analysis, 1995 to 2020,” Crime Data Explorer, accessed October 6, 2022, https://crime-data-explorer.fr.cloud.gov/pages/explorer/crime/hate-crime .

. . . [I]t is no surprise that the earliest form of policing in the United States was the slave patrol, established in 1704.

Even before the pandemic, former President Donald Trump’s xenophobic, racially inflammatory rhetoric and policies were understood as green-lighting racial and ethnic violence. But unfortunately, this is not unique in our history. Hate groups of different types—white nationalists, neo-Nazis, and anti-government paramilitary organizations—historically have played a major role in the spread of hatred and violence.

Policing: The Historical Context

The history of policing in the United States may help us determine the best policies and practices to advance public safety without subjecting communities to abusive police practices.

The institution of slavery—and its continuance—was integral to the founding of the nation. So, it is no surprise that the earliest form of policing in the United States was the slave patrol, established in 1704. 27 The patrols were designed to maintain the system of slave labor and to capture runaway slaves. Patrollers, often armed, used violence to terrorize slaves and deter rebellions. In 1787, the apprehension of slaves was codified in the U.S. Constitution in Article IV, Section 2, commonly referred to as the “Fugitive Slave Clause.” The intent of the clause, which passed unanimously, was “to require fugitive slaves and servants to be delivered up like criminals.” 28 The clause was nullified by passage of the Thirteenth Amendment in 1865. Slave patrols were also disbanded after the Civil War, only to be replaced by other forms of policing Black lives. These included the Ku Klux Klan and the institution of Jim Crow, which was maintained in Southern states by police who often used intimidation and terror to maintain a brutally oppressive system. 29

By the 1990s, the acquisition of military equipment by police forces across the country became ubiquitous . . .

Over time, police and other law enforcement agencies helped preserve and reproduce race and class inequality, as in the 1918 massacre of fifteen Mexican men in Porvenir, Texas and the 1921 Tulsa, Oklahoma massacre that resulted in destruction of a prosperous Black neighborhood. 30 Business and economic elites—such as Andrew Carnegie and Henry Clay Frick—also relied on police or private law enforcement agencies to employ deadly force against workers and union organizers. The massacre of strikers at the Homestead Steelworks in 1882 is one example; the 1897 massacre of coal miners in Lattimer, Pennsylvania is another. 31

The militarization of policing arose in the 1960s, amid cries for a “war on crime” and a “war on drugs.” In Los Angeles, Daryl F. Gates, then head of the Los Angeles Police Department, spearheaded an effort to outfit local police departments with military-grade armaments and equipment to handle emergencies, such as hostage situations and sniper shootings. By the 1990s, the acquisition of military equipment by police forces across the country became ubiquitous through a federal program that encouraged the militarization of law enforcement. 32 But, as a report from the American Civil Liberties Union documents, militarization, too, frequently came at the expense of individual civil liberties, particularly in Black and Latinx communities. 33

Police practices—including chokeholds, stop and frisks, and “broken windows” tactics—have come into question as the victims of police brutality have come into sharp focus. From George Floyd and Breonna Taylor to Eric Garner and Freddy Gray, their names are now familiar and—for some—synonymous with policing in the United States. Yet, despite the bright light shone on these cases, fatal police shootings continue to rise. According to the Washington Post , “2021 was the deadliest year for police shootings” since the newspaper began tracking such incidents in 2015. 34

Centering Communities to Advance Public Safety

Following incidents of excessive police force, municipalities commonly opt for police retraining. However, as some analysts observe, retraining is too often inadequate or ineffective in resolving or mitigating a recurrence of police misconduct. 35 Similarly, while there is merit to hiring police officers who resemble members of the communities they serve, a diverse police force does not necessarily decrease incidents of brutality against persons of color. Regrettably, research shows that a Black officer may be more inclined to use force in encounters with Black community members than white officers. 36 These officers often face the dilemma of how to fit into a police culture that commonly takes an “us against them” approach when patrolling communities. Aggressive behavior can be one way to prove that they belong. 37

Ensuring public safety requires attention to non-violent as well as violent situations. In the context of rising crime, expectations that police officers can resolve a wide array of concerns are high. Police are often called to aid unhoused people, assist individuals experiencing emotional difficulties, or settle domestic disputes. The police are not trained to handle such matters. Social workers or other trained professionals are much better equipped to deal with problems of this nature.

National data on homicide “clearance rates”—the rate at which homicide cases are resolved—also raise questions about the effectiveness of policing. In 2020, the clearance rate was just under 50 percent, representing a historic low and “a long, steady drop since the early 1980s, when police cleared about 70 percent of all homicides.” 38 The pandemic and the spike in violent crimes may help to explain the fall in clearance rates, but the data still beg the question of whether policing itself is sufficiently effective in meeting the public safety needs of contemporary society.

In determining the best approaches moving forward, the intractability of problems related to policing cannot be ignored. Policing remains a leading cause of death for young men in the United States. 39 People of color are most vulnerable, with Black men facing a one in one thousand risk of being killed by police. As we have seen, violent encounters with the police have profound effects on whole communities and neighborhoods, affecting the health and the life chances of individuals in those communities.

Community concerns about crime are real, especially in the most vulnerable communities of color. But the alternative of aggressive policing and mass incarceration has resulted in tremendous harm and cannot be the ultimate solution.

The high rate of recidivism—the rate at which persons released from prison are rearrested—does not point to a system that works well. The most recent Bureau of Justice Statistics covering a ten-year period (2008-2018) shows that “about 66 percent of prisoners released across 24 states in 2008 were rearrested within three years, and 82 percent were arrested within ten years.” 40 Recidivism rates this high should call into question the adequacy of the criminal justice system. It should also raise the issue of whether a system, focused on retribution rather than rehabilitation and public health, is actually serving the cause of public safety. These questions have found resonance with advocates of prison abolition. These abolitionists include scholars Angela Davis, Gina Dent, Ruth Wilson Gilmore, and Alex Vitale. In their view, the current structure of policing and incarceration is profoundly connected to systems of oppression. What is required, they believe, is a system that operates within a social-justice framework—one that substantively engages communities in maintaining their own safety. Such a system, they believe, would reaffirm the values of self-determination and community empowerment. 41 Rather than simply replicating punitive approaches that disproportionally and discriminatorily harm communities of color, abolitionists look to broader social solutions to the problem of crime—remedial measures such as restitution, reconciliation, rehabilitation, and restorative justice.

While a complete transformation of policing and the criminal justice system may not be on the immediate horizon, a variety of initiatives in recent years have sought to address the basic human needs of communities while minimizing negative interactions with police. In a 2021 experiment in Brooklyn, for example, the Brownsville Safety Alliance—a community-based organization—arranged for precinct police to disengage from their usual assignments in a two-block area for five days. In their place, trained violence interrupters and crisis management groups were charged with securing public safety. Although limited in duration, this pilot program has been praised by New York City officials as well as members of the community as “a model for the future.” 42 A range of other crime-reduction strategies that do not involve the deployment of police have also produced promising results. These include improvements in street lighting, clean-up of empty lots, provision of quality mental health and drug treatment services, and expansion of Medicaid services. 43

Community concerns about crime are real, especially in the most vulnerable communities of color. But the alternative of aggressive policing and mass incarceration has resulted in tremendous harm and cannot be the ultimate solution. The best, most promising option is to center communities and underlying social and economic inequality as the means to advance public safety.

Author Biography

José Luis Morín is a professor at John Jay College of Criminal Justice, City University of New York.

1 Jeff Asher, “Murder Rose by almost 30% in 2020: It’s Rising at a Slower Rate in 2021,” New York Times , September 22, 2021, updated November 15, 2021, available at https://www.nytimes.com/2021/09/22/upshot/murder-rise-2020.html .

2 Federal Bureau of Investigation, “FBI Releases 2020 Hate Crime Statistics,” August 30, 2021, available at https://www.fbi.gov/news/press-releases/press-releases/fbi-releases-2020-hate-crime-statistics ; See also, Christina Carrega and Priya Krishnakumar, “Hate Crime Reports in US Surge to the Highest Level in 12 Years, FBI Says,” CNN , October 26, 2021, available at https://www.cnn.com/2021/08/30/us/fbi-report-hate-crimes-rose-2020/index.html .

3 Federal Bureau of Investigation, “Trend of Violent Crime from 1985 to 2020,” Crime Data Explorer, available at https://crime-data-explorer.fr.cloud.gov/pages/explorer/crime/crime-trend .

4 Timothy Schnacke, “Fundamentals of Bail: A Resource Guide for Pretrial Practitioners and a Framework for American Pretrial Reform,” National Institute of Corrections, September 2, 2014, available at https://s3.amazonaws.com/static.nicic.gov/Library/028360.pdf .

5 U.S. Commission on Civil Rights, “Civil Rights Implications of Cash Bail,” Briefing Report, Washington, DC, available at https://www.usccr.gov/reports/2021/civil-rights-implications-cash-bail .

6 U.S. Commission on Civil Rights, 10.

7 Adureh Onyekwere, “How Cash Bail Works,” Brennan Center for Justice at NYU Law, December 10, 2019, updated February 24, 2021, available at https://www.brennancenter.org/our-work/research-reports/how-cash-bail-works#:~:text=Cash%20bail%20is%20used%20as,is%20forfeited%20to%20the%20government ; See also, U.S. Commission on Civil Rights, 6-8.

8 Bernadette Rabuy and Daniel Kopf, “Detaining the Poor: How Money Bail Perpetuates an Endless Cycle of Poverty and Jail Time,” Prison Policy Initiative , May 10, 2016, available at https://www.prisonpolicy.org/reports/DetainingThePoor.pdf .

9 Ames Grawert and Noah Kim, “The Facts on Bail Reform and Crime Rates in New York State,” Brennan Center for Justice at NYU Law, March 22, 2022, available at https://www.brennancenter.org/our-work/research-reports/facts-bail-reform-and-crime-rates-new-york-state .

10 Taryn A. Merkl, “New York’s Latest Bail Law Changes Explained,” Brennan Center for Justice at NYU Law, April 16, 2020, available at https://www.brennancenter.org/our-work/analysis-opinion/new-yorks-latest-bail-law-changes-explained ; See also, Jon Campbell, “NY Lawmakers Pass $220B Budget that Changes Bail Reform, Approves Buffalo Bills Stadium Funding,” Gothamist, April 9, 2022, available at https://gothamist.com/news/ny-lawmakers-pass-220b-budget-that-changes-bail-reform-approves-buffalo-bills-stadium-funding?utm_source=sfmc&utm_medium=nypr-email&utm_campaign=Gothamist%20Daily%20Newsletter&utm_term=https%3A%2F%2Fgothamist.com%2Fnews%2Fny-lawmakers-pass-220b-budget-that-changes-bail-reform-approves-buffalo-bills-stadium-funding&utm_id=88591&sfmc_id=2849872&utm_content=202249 .

11 Grawert and Kim, “Bail Reform and Crime.”

12 Grawert and Kim, “Bail Reform and Crime.”

13 Office of New York City Comptroller Brad Lander, “NYC Bail Trends Since 2019,” available at https://comptroller.nyc.gov/reports/nyc-bail-trends-since-2019/ .

14 Gillian B. White, “Who Really Makes Money Off of Bail Bonds?” The Atlantic , May 12, 2017, available at https://www.theatlantic.com/business/archive/2017/05/bail-bonds/526542/ ; See also, Onyekwere, “How Cash Bail Works”; Rabuy and Kopf, “Detaining the Poor.”

15 “During Questioning in Albany, NYPD Commissioner Shea Backtracks on Bail Reform Law as Big Reason for Gun Violence,” CBS New York , October 14, 2021, available at https://www.cbsnews.com/newyork/news/bail-reform-nypd-commissioner-dermot-shea-assembly-hearing/ .

16 Daniel Funke, “Fact Check: No Evidence Defunding Police to Blame for Homicide Increases, Experts Say,” USA TODAY , January 28, 2022, available at https://www.usatoday.com/story/news/factcheck/2022/01/28/fact-check-police-funding-not-linked-homicide-spikes-experts-say/9054639002/ .

17 See, for example, Brian Bell, Rui Costa, and Stephen Machin, “Why Does Education Reduce Crime?” Journal of Political Economy 130, no. 3 (2022): 732-65.

18 David J. Deming, “Better Schools, Less Crime?” Quarterly Journal of Economics 126 (2011): 2063-115; See also, E. Jason Baron, Joshua M. Hyman, and Brittany N. Vasquez, “Public School Funding, School Quality, and Adult Crime,” National Bureau of Economic Research, Working Paper 29855, available at http://www.nber.org/papers/w29855 .

19 FBI, “Trend of Violent Crime from 1985 to 2020.”

20 FBI, “Trend of Homicide from 1985-2020,” Crime Data Explorer, available at https://crime-data-explorer.fr.cloud.gov/pages/explorer/crime/crime-trend .

21 Charles Oberg, H.R. Hodges, Sarah Gander, Rita Nathawad, and Diana Cutts. “The Impact of COVID-19 on Children’s Lives in the United States: Amplified Inequities and a Just Path to Recovery,” Current Problems in Pediatric and Adolescent Health Care 52, no. 7 (2022): 1-17.

22 Sabrina Tavernise, “An Arms Race in America: Gun Buying Spiked during the Pandemic. It’s Still Up,” New York Times , May 29, 2021, updated May 30, 2021, available at https://www.nytimes.com/2021/05/29/us/gun-purchases-ownership-pandemic.html .

23 Tavernise, “An Arms Race in America.”

24 Philip J. Cook and Jens Ludwig, “Gun Violence Is THE Crime Problem,” Vital City . March 2, 2022, available at https://www.vitalcitynyc.org/articles/gun-violence-is-the-crime-problem .

25 Carrega and Krishnakumar, “Hate Crime Reports in US Surge”; FBI, “Hate Crime in the United States Incident Analysis, 1995-2020,” Crime Data Explorer,” accessed October 6, 2022, https://crime-data-explorer.fr.cloud.gov/pages/explorer/crime/hate-crime .

26 Center for the Study of Hate & Extremism at California State University, “Report to the Nation: Anti-Asian Prejudice & Hate Crime 2021,” (2021), available at https://www.csusb.edu/sites/default/files/Report%20to%20the%20Nation%20-%20Anti-Asian%20Hate%202020%20Final%20Draft%20-%20As%20of%20Apr%2028%202021%2010%20AM%20corrected.pdf .

27 Chelsea Hansen, “Slave Patrols: An Early Form of American Policing,” National Law Enforcement Museum , July 10, 2019, available at https://lawenforcementmuseum.org/2019/07/10/slave-patrols-an-early-form-of-american-policing/ ; See also, Jill Lepore, “The Invention of the Police: Why Did American Policing Get so Big, so Fast? The Answer, Mainly, Is Slavery,” The New Yorker , July 13, 2020, available at https://www.newyorker.com/magazine/2020/07/20/the-invention-of-the-police .

28 Library of Congress, “The Fugitive Slave Clause,” Constitution Annotated, available at https://constitution.congress.gov/browse/essay/artIV-S2-C3-1/ALDE_00013571/ [“clause”].

29 Hansen, “Slave Patrols.”

30 Zinn Education Project, “Massacres in U.S. History,” available at https://www.zinnedproject.org/collection/massacres-us/ .

31 Gary Potter, “The History of Policing in the United States, Part 3,” EKU Online . Eastern Kentucky University. July 9, 2013, available at https://ekuonline.eku.edu/blog/police-studies/the-history-of-policing-in-the-united-states-part-3/ ; See also, Paul A. Shackel, “How a 1897 Massacre of Pennsylvania Coal Miners Morphed from a Galvanizing Crisis to Forgotten History,” Smithsonian Magazine , March 13, 2019, available at https://www.smithsonianmag.com/history/how-1897-massacre-pennsylvania-coal-miners-morphed-galvanizing-crisis-forgotten-history-180971695/ .

32 Radley Balko, Rise of the Warrior Cop: The Militarization of America’s Police Forces , (New York: Public Affairs, 2021).

33 American Civil Liberties Union, “War Comes Home: The Excessive Militarization of American Policing” (2014), available at https://www.aclu.org/sites/default/files/assets/jus14-warcomeshome-report-web-rel1.pdf .

34 The Marshall Project, “How Policing Has—and Hasn’t—Changed since George Floyd,” August 6, 2022, available at https://www.themarshallproject.org/2022/08/06/how-policing-has-and-hasn-t-changed-since-george-floyd .

35 See, for example, Alex S. Vitale, The End of Policing (London: Verso Books, 2017), 4-11.

36 Vitale, The End of Policing , 11-13; See also, “Does Diversifying Police Forces Reduce Brutality against Minorities?” NPR , June 22, 2020, available at https://www.npr.org/2020/06/22/881559659/does-diversifying-police-forces-reduce-brutality-against-minorities .

37 Vitale, The End of Policing ; See also, José Luis Morín, Latino/a Rights and Justice in the United States: Perspectives and Approaches , 2nd edition (Durham, NC: Carolina Academic Press, 2009), 106-15; Balko, Rise of the Warrior Cop .

38 Weihua Li and Jamiles Lartey, “As Murders Spiked, Police Solved about Half in 2020,” The Marshall Project , January 12, 2022, available at https://www.themarshallproject.org/2022/01/12/as-murders-spiked-police-solved-about-half-in-2020 .

39 Frank Edwards, Hedwig Lee, and Michael Esposito, “Risk of being Killed by Police Use of Force in the United States by Age, Race–Ethnicity, and Sex,” PNAS 116, no. 34 (2019): 16793-8.

40 Leonardo Antenangeli and Matthew R. Durose, Recidivism of Prisoners Released in 24 States in 2008: A 10-Year Follow-up Period (2008–2018), Bureau of Justice Statistics NCJ Number 256094 September 2021, available at https://bjs.ojp.gov/library/publications/recidivism-prisoners-released-24-states-2008-10-year-follow-period-2008-2018 .

41 See, for example, Ruth Wilson Gilmore, Golden Gulag: Prisons Surplus, Crisis, and Oppression in Globalizing California (Berkeley: University of California Press, 2007); Vitale, The End of Policing ; Derecka Purnell, Becoming Abolitionists: Police, Protests, and the Pursuit of Freedom (New York: Astra House, 2021); Mariame Kaba, We Do This ’Til We Free Us: Abolitionist Organizing and Transforming Justice (Chicago: Haymarket Books, 2021); Angela Y. Davis, Gina Dent, Erica R. Meiners, and Beth Richie. Abolition. Feminism. Now (Chicago: Haymarket Books, 2022).

42 Yoav Gonen and Eileen Grench, “Five Days without Cops: Could Brooklyn Policing Experiment Be a ‘Model for the Future’?” The City , January 3, 2021, available at https://www.thecity.nyc/2021/1/3/22211709/nypd-cops-brooklyn-brownsville-experiment-defund-police .

43 Shaila Dewan. “‘Re-fund the Police’? Why It might Not Reduce Crime,” New York Times , November 8, 2021, updated November 11, 2021, available at https://www.nytimes.com/2021/11/08/us/police-crime.html .

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.

Sorry, we did not find any matching results.

We frequently add data and we're interested in what would be useful to people. If you have a specific recommendation, you can reach us at [email protected] .

We are in the process of adding data at the state and local level. Sign up on our mailing list here to be the first to know when it is available.

Search tips:

• Check your spelling

• Try other search terms

• Use fewer words

How is crime measured in the US?

The FBI relied heavily on estimates for 2021 crime statistics due to low response rates from law enforcement agencies.

Published Wed, May 17, 2023 by the USAFacts Team

Is crime in the US increasing or decreasing? It’s a question that the Department of Justice tries to answer using two primary sources: victim surveys and administrative data from law enforcement agencies.

The Bureau of Justice Statistics’ National Crime Victimization Survey captures information directly from victims, while the FBI’s Uniform Crime Reporting (UCR) Program collects data from law enforcement agencies. These data sources, which account for crimes that are reported to the police as well as those that aren’t, together provide a more comprehensive understanding of crime in the US.

In 2021, however, with only 64% of the US population covered by law enforcement agency participation, the FBI relied on crime estimates to compensate for the missing data.

Local and state agencies are not required by federal law to submit crime data to the FBI, making their participation voluntary.

FBI adopts a more detailed reporting system, law enforcement participation drops

The FBI has been collecting crime data from local law enforcement agencies through the Summary Reporting System (SRS) since 1929. SRS was replaced in 2021 by the National Incident-Based Reporting System (NIBRS), a more detailed reporting system.

In addition to removing the SRS Hierarchy Rule, which only recorded the most serious offense in a crime incident, NIBRS captures more information than SRS by allowing the data to count up to 10 offenses per incident and including more offense categories.

Although the FBI created NIBRS in the 1980s , law enforcement agencies were given until January 1, 2021, to switch to the new system, and some continued to submit data to the SRS until the deadline.

The results of the transition have been mixed. While NIBRS collects data for 42 more types of offenses and collects more details about victims and offenders than SRS, technological and cost-related challenges meant some law enforcement agencies missed the deadline for the transition.

Most agencies take a year to transition to NIBRS, while larger or more complex ones may need two years. This involves certification, implementing new software, personnel training, and acquiring funding. Washington, DC’s, government spent $50,493 to implement NIBRS in 2021 , while the Maryland governor's office budgeted $2.5 million to ensure full NIBRS compliance for all local law enforcement agencies by the federal deadline.

Through the NIBRS Statewide Compliance Initiative, the Governor’s Office of Crime Prevention, Youth, and Victim Services has allocated approximately $2.5 million in Justice Department funding to support Maryland’s efforts to ensure 100% NIBRS compliance for all local law enforcement agencies by the federal deadline. The office anticipates making up to 50 awards. Priority will be given to agencies requesting technology enhancements.

Local law enforcement agency participation in SRS has historically covered greater than 90% of the US population , a number that dropped to 64% in 2021 with NIBRS. The FBI fell 16 percentage points short of its goal to have data from agencies serving more than 80% of the US population by 2021.

Notably, law enforcement agencies from New York City and Los Angeles, the two most populous cities in the US, did not submit data to NIBRS in 2021 . The entire state of Florida also did not submit data to NIBRS that year.

Although many cities didn't provide data to the FBI in 2021, the FBI employs statistical methods , such as weighting, to fill in data gaps and produce crime figures that are representative of the country.

For a fuller picture of crime in the US , read about which states have the least and most crime . Get USAFacts data in your inbox by subscribing to our weekly newsletter .

Explore more of USAFacts

Related articles, firearm background checks: explained.

The state of domestic terrorism in the US

The latest government data on school shootings

How many guns are made in the US?

Related Data

Firearm Deaths in the US: Statistics and Trends

Firearm background checks

38.88 million

Data delivered to your inbox

Keep up with the latest data and most popular content.

SIGN UP FOR THE NEWSLETTER

Crime Prevention Research Center

Dedicated to conducting academic quality research on the relationship between laws regulating the ownership or use of guns, crime, and public safety, latest research.

The Rate that Police Officers are Murdered with their Own Weapon

by johnrlott | May 23, 2024 | Original Research

Often, people point to the rate at which police officers are murdered with their guns. Over the 17 years from 2002 to 2018, there were 34 officers murdered with their gun compared to the 878 officers who were murdered — 3.87% of the total. But the relevant comparison...

How reliable are the FBI’s Report of Violent Crime Data? There are some major problems.

by johnrlott | Apr 20, 2024 | Original Research

The U.S. employs two distinct measures of crime. The Federal Bureau of Investigation’s Uniform Crime Reporting program counts the number of crimes reported to police annually. The Bureau of Justice Statistics, in its National Crime Victimization Survey (NCVS), asks...

The Latest Research

Get our latest research on crime sent straight to you. Find out what law enforcement and gun ownership policies will make people safer

Latest Columns

The Washington Times: Don’t defend yourself in New York City: If you do, you will be prosecuted

by johnrlott | May 29, 2024 | op-ed

Dr. John Lott has a new op-ed at the Washington Times. . The message in New York City is clear: don’t try defending yourself or anyone else from violence. If you do, you will be prosecuted by the city’s District Attorneys. . New Yorkers are likely to be very familiar...

At Townhall: The FBI’s Crime Data Have Real Problems

by johnrlott | May 15, 2024 | op-ed

Dr. John Lott has a new piece up at Townhall that continues our investigation into the problems with the FBI crime data. The news media relies almost exclusively on FBI data to report on changes in crime rates. But there is strong evidence that FBI data...

Letter Submitted to USA Today: Responding to “FBI data shows America is seeing a ‘considerable’ drop in crime. Trump says the opposite.”

by johnrlott | May 12, 2024 | op-ed

When USA Today had two articles claiming that Trump was wrong in claiming that crime was increasing, Dr. John Lott submitted the following letter to the editor, but the newspaper didn't run the letter. Dear Letters Editor: USA Today erroneously dismisses former...

At The Tennessean: My husband was killed in a ‘gun free zone.’ Arm teachers for safety and to save lives.

by johnrlott | May 1, 2024 | op-ed

The CPRC's Nikki Goeser has a new op-ed in the Tennessean. . My husband Ben and I used to run a mobile karaoke business in Nashville. . Every Thursday evening, we would load up our vehicle and head to a popular restaurant to help facilitate a night of good music and...

At the Wall Street Journal: The Media Say Crime Is Going Down. Don’t Believe It: The decline in reported crimes is a function of less reporting, not less crime.

by johnrlott | Apr 24, 2024 | Crime , op-ed

Dr. John Lott has a new op-ed at the Wall Street Journal. News outlets claim that Americans mistakenly believe violent crime is rising. A Gallup survey last year found that 92% of Republicans and 58% of Democrats thought crime was increasing. A February...

At Townhall: City where police emergency response time is 36 minutes wants to ban Civilians Carrying Guns for protection

by johnrlott | Apr 24, 2024 | op-ed

Dr. John Lott has a new piece at Townhall In 2023, New Orleans had the third highest per capita murder rate in the country according to the FBI. Last week, New Orleans Police Chief Anne Kirkpatrick claimed that the city can lower...

At the Missoulian: Duty to uphold the laws of the land rests on Tester’s shoulders

by johnrlott | Apr 15, 2024 | op-ed

Dr. John Lott has a new piece at the Missoulian. . Will Sen. Jon Tester provide the necessary vote this week to hold an impeachment trial for Homeland Security Secretary Alejandro Mayorkas? Democrats might think that they can protect vulnerable Democrats such as...

At the Colorado Springs Gazette: Registries could be used to enforce gun bans

by johnrlott | Apr 14, 2024 | op-ed

John Lott and Colorado state Representative Ryan Armagost have a piece in Colorado's second largest newspaper the Colorado Springs Gazette. . Credit card companies soon might be tracking purchases of firearms and ammunition in Colorado. A bill that would flag...

Popular posts:

• Violent Crime Rates by Race
• UPDATED: Compiling Cases where concealed handgun…
• Massive errors in FBI’s Active Shooting…
• UPDATED: Mass Public Shootings keep occurring in…
• Twenty-five states have Constitutional Carry
• Massive errors in FBI’s Active Shooting Reports from…

Public Speech in São Paulo, Brazil on July 5th

by johnrlott | May 29, 2024 | Talk

Dr. John Lott will give talks in Santiago, Chile, at the end of June and then travel to Brazil, where he will give more talks.

Conference talk on the Collapse of Law Enforcement in the US

by RW | May 12, 2024 | Talk

https://youtu.be/b24Qep3c2nE Dr. John Lott gave a talk at a Berkeley Springs 2024 Conference on the collapse of law enforcement in the US. Dr. Lott disagreed with what he was told were the concerns of others at the conference regarding legal immigration and also...

Crime Prevention Research Center fellow Amanda Collins Johnson talked about being raped in college and her advocacy for “campus carry” gun laws

by johnrlott | Apr 4, 2024 | Talk

Crime Prevention Research Center fellow Amanda Collins Johnson talked about being raped in college and her advocacy for “campus carry” gun laws. This event was hosted on February 29th, 2024 by the Clare Boothe Luce Center for Conservative Women in Virginia. C-SPAN...

Media Appearances

On Iowa’s KXEL: To Discuss Crime Data

by RW | May 27, 2024 | Radio

Dr. John Lott appeared on Iowa’s giant 50,000-watt KXEL-AM radio station to discuss his new op-ed at the Wall Street Journal titled "The Media Say Crime Is Going Down. Don’t Believe It: The decline in reported crimes is a function of less reporting, not less crime."...

On WAVA FM’s The Drive Home Show: To Discuss AI’s Left-Wing Bias and Decline in Reported Crimes

by RW | May 26, 2024 | Radio

Dr. John Lott talked to Greg Seltz on WAVA's The Drive Home Show about AI's left-wing bias and the collapse in law enforcement. See also Dr. Lott's new op-ed at Real Clear Politics ("AI’s Left-wing Bias on Crime and Gun Control") and his latest piece at the Wall...

On The Rod Arquette Show in Salt Lake City: The Media Say Crime Is Going Down. Don’t Believe It.

by RW | May 25, 2024 | Radio

Dr. John Lott appeared on KNRS’s The Rod Arquette Show in Salt Lake City to discuss his new op-ed at the Wall Street Journal about why we should not believe the media when they say crime numbers are decreasing.(Thursday, April 25, 2024, from 6:05 to 6:15 PM)

All Our Posts

CPRC’s Nikki Goeser on Armed American Radio: Teachers carrying guns to protect students in Tennessee

by Nikki Goeser | May 29, 2024 | Campus carry , Media Bias

Senior Fellow Nikki Goeser speaks with host Mark Walters about Tennessee Teachers now being able to carry concealed to protect their students. Nikki also discusses some of the media bias she has experienced when standing up for gun rights. Sunday, May 19, 2024

Is there a massive change in whether both blacks and whites believe Republicans or Democrats have done the most for blacks?

by johnrlott | May 28, 2024 | Survey

Rasmussen Reports surveyed likely voters in 2020 and 2024 to determine whether they thought Republicans or Democrats had done the most for blacks. In May 2020, only 18% of blacks thought that Republicans had done the most, but that increased to 32% in May 2024. Large...

CPRC’s Nikki Goeser on the Tudor Dixon Podcast

by Nikki Goeser | May 28, 2024 | gun control , Gun Free Zones , Second Amendment

https://www.youtube.com/watch?v=RJAOvDaogUw May 6, 2024 In this episode, Tudor talks with Senior Fellow of the Crime Prevention Research Center Nikki Goeser, a victims' rights advocate, who shares her story of her husband being brutally murdered by her stalker. She...

The accuracy of crime statistics: assessing the impact of police data bias on geographic crime analysis

• Open access
• Published: 26 March 2021
• Volume 18 , pages 515–541, ( 2022 )

Cite this article

You have full access to this open access article

• David Buil-Gil   ORCID: orcid.org/0000-0002-7549-6317 1 ,
• Angelo Moretti   ORCID: orcid.org/0000-0001-6543-9418 2 &
• Samuel H. Langton   ORCID: orcid.org/0000-0002-1322-1553 3  

26k Accesses

25 Citations

32 Altmetric

Explore all metrics

Police-recorded crimes are used by police forces to document community differences in crime and design spatially targeted strategies. Nevertheless, crimes known to police are affected by selection biases driven by underreporting. This paper presents a simulation study to analyze if crime statistics aggregated at small spatial scales are affected by larger bias than maps produced for larger geographies.

Based on parameters obtained from the UK Census, we simulate a synthetic population consistent with the characteristics of Manchester. Then, based on parameters derived from the Crime Survey for England and Wales, we simulate crimes suffered by individuals, and their likelihood to be known to police. This allows comparing the difference between all crimes and police-recorded incidents at different scales.

Measures of dispersion of the relative difference between all crimes and police-recorded crimes are larger when incidents are aggregated to small geographies. The percentage of crimes unknown to police varies widely across small areas, underestimating crime in certain places while overestimating it in others.

Conclusions

Micro-level crime analysis is affected by a larger risk of bias than crimes aggregated at larger scales. These results raise awareness about an important shortcoming of micro-level mapping, and further efforts are needed to improve crime estimates.

Similar content being viewed by others

The Impact of Measurement Error in Regression Models Using Police Recorded Crime Rates

Too Fine to be Good? Issues of Granularity, Uniformity and Error in Spatial Crime Analysis

More places than crimes: implications for evaluating the law of crime concentration at place.

Avoid common mistakes on your manuscript.

Introduction

Police-recorded crimes are the main source of information used by police forces and government agencies to analyze crime patterns, investigate the geographic concentration of crime, and design and evaluate spatially targeted policing strategies and crime prevention policies (Bowers and Johnson 2014 ; Weisburd and Lum 2005 ). Police statistics are also used by criminologists to develop theories of crime and deviance (Bruinsma and Johnson 2018 ). Nevertheless, crimes known to police are affected by selection biases driven by unequal crime reporting rates across social groups and geographical areas (Buil-Gil et al. 2021 ; Goudriaan et al. 2006 ; Hart and Rennison 2003 ; Xie 2014 ; Xie and Baumer 2019a ). The level of police control (e.g., police patrols, surveillance) also varies across areas, which may affect victims’ willingness to report crimes to police and dictate the likelihood that police officers witness incidents in some places more than others (McCandless et al. 2016 ; Schnebly 2008 ). The sources of measurement error that affect the bias and precision of crime statistics is an issue that merits scrutiny, since it affects policing practices, criminal justice policies, and citizens’ daily lives. Yet, it is an understudied issue.

The implications of crime data biases for documenting and explaining community differences in crime and guiding policing operational decision-making processes are mostly unknown (Brantingham 2018 ; Gibson and Kim 2008 ; Kirkpatrick 2017 ). Moreover, police analyses and crime mapping are moving toward using increasingly fine-grained geographic units of analysis, such as street segments and micro-places containing highly homogeneous communities (Groff et al. 2010 ; Weisburd et al. 2009 , 2012 ). Geographic crime analysis based on police-recorded crime and calls for service data is used to identify the micro-places where crime is most prevalent in order to effectively target police resource (Braga et al. 2018 ). In this context, we define “micro-places” as very detailed spatial units of analysis such as addresses, street segments, or clusters of such units (Weisburd et al. 2009 ). Despite the increasing interest in small units of analysis, the extent to which such aggregations impact on the overall accuracy of statistical outputs and spatial analyses remains unknown (Ramos et al. 2020 ). In other words, we do not know whether aggregating crime data at such detailed levels of analysis increases the impact of biases introduced by underreporting. This article presents a simulation study to analyze the impact of data biases on geographic crime analysis conducted at different spatial scales. The open question that this research aims to address is whether aggregating crimes at smaller, more socially homogeneous spatial scales increases the risk of obtaining biased outputs compared with aggregating crimes at larger, more socially heterogeneous geographical levels.

Since the early 1830s, numerous researchers have expressed concern about the limitations of using official statistics to analyze crime patterns across space and time (Kitsuse and Cicourel 1963 ; Skogan 1974 ). Soon after the publication of the first judiciary statistics in France, Alphonse de Candolle ( 1987a [1830], 1987b [1832]) cautioned that the validity of these data was likely to be affected by various sources of measurement error. For instance, crimes may not be discovered by victims, some victims may not report crimes to the authorities, offenders’ identities may remain unknown, and legal procedures may not lead to conviction. Moreover, cross-sectional comparisons of the number of people convicted in court are likely to be affected by changes in prosecution activity, and the proportion of recorded crimes to unknown offences may vary between countries (Aebi and Linde 2012 ). De Candolle ( 1987b [1832]) argued that the number of persons accused of crime was a better indicator of crime incidence than the number of persons convicted, since the former is closer to crime events in terms of legal procedure. This rationale was later used to describe the so-called “Sellin’s dictum” (i.e., “the value of a crime rate for index purposes decreases as the distance from the crime itself in terms of procedure increases,” Sellin 1931 : 346), and it is the main reason why crime incidents known to the police are generally preferred over judiciary statistics when it comes to analyzing crime. Police-recorded crimes, however, are also subject to criticism over the validity of recording and reporting. So much so that such data lost the official designation of National Statistics in the UK in 2014 (UK Statistics Authority 2014 ).

A key issue of concern regarding the use of police records for crime analysis and mapping is the fact that crime reporting rates are unequally distributed across social groups and geographic areas. Crime reporting to police forces is known to be more common among female victims than male victims, and young citizens report crimes less often than adults (Hart and Rennison 2003 ; Tarling and Morris 2010 ). There are also contextual factors that affect crime reporting rates across areas, such as neighborhood economic deprivation, the degree of urbanization, the concentration of minorities, and social cohesion (Berg et al. 2013 ; Goudriaan et al. 2006 ; Slocum et al. 2010 ; Xie and Baumer 2019a , b ; Xie and Lauritsen 2012 ). The demographic and social characteristics of small areas are generally more homogeneous compared with larger scales (e.g., Brattbakk 2014 ; Weisburd et al. 2012 ). Thus, crime aggregates produced at the level of small geographies are more likely to be affected by unequal crime reporting rates across groups compared with aggregates and maps produced at larger, more heterogeneous spatial scales. For instance, Buil-Gil et al. ( 2021 ) show that the variation in the “dark figure of crime” (i.e., all crimes not shown in police statistics) between neighborhoods (within cities) is larger than the variation between cities. We expect the risk of police data bias to be especially large when aggregating crime records at the level of micro-places.

This paper is organized as follows: sect. “ The criminology of place ” introduces the move toward low-level crime analysis in criminology. Section “Geographic crime analysis and measurement error ” discusses the various sources of measurement error that may affect police records and introduce bias into our understanding of community differences in crime. Section “ Data and methods ” introduces the data, methods, and steps taken to generate the synthetic population for our simulation study, and methods used to assess the findings. Section “ Mapping the bias of police-recorded crimes ” reports the results of the simulation study. Finally, sect. “ Discussion and conclusions ” discusses the findings and presents the conclusions and limitations, along with suggestions for future research.

The criminology of place

In the 1980s, several researchers began analyzing the concentration of crime in places and found that a large proportion of crimes known to the police concentrated in a small number of micro-places. Pierce et al. ( 1988 ) showed that 50% of all calls for police services in Boston took place in just 2.6% of addresses, suggesting that a disproportionately large volume of total crime could be attributed to just a handful of places. A year later, Sherman et al. ( 1989 ) conducted similar research in Minneapolis, obtaining almost the same results: 2.5% of addresses in this city generated 50% of all crime calls to the police. These were only two of the first studies looking into the concentration of crime in places. Since then, many other researchers have published remarkably similar findings (see a review in Lee et al. 2017 ). Environmental criminologists argue that the social and contextual conditions that favor crime vary across micro-places, and that opportunities for crime are structured within very small geographic areas (Brantingham and Brantingham 1995 ; Weisburd et al. 2012 ).

Given the persistency of this finding across multiple study sites and countries, Weisburd ( 2015 : 138) argues for a so-called “law of crime concentration” at micro-places, namely, that “for a defined measure of crime at a specific microgeographic unit, the concentration of crime will fall within a narrow bandwidth of percentages for a defined cumulative proportion of crime.” This has served as a basis for police forces all over the world to develop place-based strategies that increase police control over those areas where crime is highly concentrated to efficiently reduce citywide crime (Braga et al. 2018 ; Groff et al. 2010 ; Kirkpatrick 2017 ).

However, the vast majority of research analyzing crime concentration, and evaluating the impact of place-based policing interventions, is based on data about crimes known to the police. For instance, 41 out of 44 studies examining the crime concentration at places reviewed by Lee et al. ( 2017 ) used crime incidents reported to police, and 4 out of 44 analyzed calls for police services (note that some studies used more than one source of data). Both these sources of data depend on citizens’ willingness to report crimes and cooperate with the police, which are known to be affected by the social and demographic characteristics of individuals, but also by variables that operate at the scales of small communities, such as concentrated disadvantage, perceived disorder, and collective efficacy (Jackson et al. 2013 ). Weisburd et al. ( 2012 : 5) argue that “the criminology of place [...] emphasizes the importance of micro-units of geography as social systems relevant to the crime problem.” And yet, these micro-level social systems may also be key in explaining why crime reporting rates—and thus the likelihood of crimes being known to police—are high in some places and low in others, and as such, we might expect that the sources of measurement error that affect police data will vary across micro-places.

Geographic crime analysis and measurement error

There are four primary sources of data bias that may affect the accuracy of community differences in crime documented through police statistics. First, the willingness of residents to report crimes to police is known to be associated with individual and contextual factors that vary across geographic areas (Hart and Rennison 2003 ). There are demographic, social, economic, and environmental factors that affect crime reporting rates. For example, the victims’ sex, age, employment status, education level, and ethnic group are all good predictors of their likelihood to report crimes to the police (Hart and Rennison 2003 ). Since some of these resident characteristics concentrate in particular areas, we also expect crime reporting rates to vary across areas. Generally, deprived neighborhoods and areas with large concentrations of immigrants have lower crime reporting rates than middle-class areas (Baumer 2002 ; Xie and Baumer 2019a ; Goudriaan et al. 2006 ), and crimes that take place in cohesive areas have a higher chance of being known to the police (Goudriaan et al. 2006 ; Jackson et al. 2013 ). Moreover, residents from rural areas are generally more willing to cooperate with police services than urban citizens (Hart and Rennison 2003 ). Research has also found that the incident seriousness and harm are very strongly linked to the reporting decision (Baumer 2002 ; Xie and Baumer 2019b ).

Second, studies have found that the overall crime rate and citizens’ perceptions about police forces, which also vary across areas, affect residents’ willingness to cooperate with the police (e.g., Xie 2014 ). Berg et al. ( 2013 ) show that the most important contextual factor in explaining crime reporting is the level of crime in the area. Jackson et al. ( 2013 ) argue that the level of trust in police fairness and residents’ perceptions of police legitimacy is key to predict the willingness to cooperate with police forces.

Third, unequal police control across areas may inflate crime statistics in some places but not others. Schnebly ( 2008 ) shows that cities with more police officers trained in community-oriented policing generally have higher rates of police notification, whereas McCandless et al. ( 2016 ) argue that poorly handled stop and search practices may discourage residents from engaging with the police.

Fourth, there may be differences between counting rules applied by different police forces (Aebi and Linde 2012 ). This is not expected to be a major source of error in England and Wales, since all 43 police forces follow common counting rules (National Crime Recording Standards and Home Office Counting Rules for Recorded Crime). Nevertheless, we note that, in 2014, Her Majesty’s Inspectorate of Constabulary and Fire & Rescue Services conducted an inspection about police statistics and concluded that the extent to which certain counting practices was followed varied between police forces (HMIC 2014 ).

Some of these sources of measurement error were mentioned by Skogan ( 1977 : 41) to argue that the dark figure of crime “limits the deterrent capability of the criminal justice system, contributes to the misallocation of police resources, renders victims ineligible for public and private benefits, affects insurance costs, and helps shape the police role in society.” Moreover, the UK public administration also acknowledges that “there is accumulating evidence that suggests the underlying data on crimes recorded by the police may not be reliable” (UK Statistics Authority 2014 : 2). As a consequence, in 2014, crime data were removed from the UK National Statistics designation.

Given that many of the factors generating disparities in the bias and precision of police-recorded crime data are non-uniformly distributed across space, even in the same city, it is plausible that the bias affecting crime data varies considerably between small areas. Indeed, issues of bias and precision may even be compounded as the geographic resolution becomes more fine-grained. Oberwittler and Wikström ( 2009 : 41) argue that, in order to analyze crime, “smaller geographical units are more homogeneous, and hence more accurately measure environments. In other words, smaller is better.” Smaller units of analysis are said to be better for explaining criminal behaviors since crime is determined by opportunities that occur in the immediate environment. However, smaller units of analysis may also be preferred to explain the amount of crime which remains hidden in police statistics (either because victims and witnesses fail to report or because the police fail to record). The “aggregation bias,” which argues that what is true for a group should also be true for individuals within such a group, tends to be used to justify the selection of smaller spatial units in geographic crime analysis due to this homogeneity in residential characteristics. And yet, high internal homogeneity and between-unit heterogeneity may generate greater variability in bias and precision between units. It would be paradoxical and self-defeating if, in seeking to avoid aggregation bias with the use of micro-scale units, studies increase the risk of crime statistics being affected by bias and imprecision. This would have significant repercussions for academic endeavor and policing practices that document and explain community differences in crime.

Data and methods

Simulation studies are computer experiments in which data is created via pseudo-random sampling in order to evaluate the bias and variance of estimators, compare estimators, investigate the impact of sample sizes on estimators’ performance, and select optimal sample sizes, among others (Moretti 2020 ). Brantingham and Brantingham ( 2004 ) recommend the use of computer simulations to understand crime patterns and provide policy guidance for crime control (see also Groff and Mazerolle 2008 ; Townsley and Birks 2008 ). In this study, we generate a synthetic dataset of crimes known and unknown to police in Manchester, UK, and aggregate crimes at different spatial scales. This permits an investigation into whether aggregates of crimes known to police at the micro-scale level suffer from a higher risk of bias compared with those at larger aggregations, such as neighborhoods and wards.

Based on parameters obtained from the UK Census 2011 and Index of Multiple Deprivation (IMD) 2010, we simulate a synthetic individual-level population consistent with the characteristics of Manchester. The simulated population reflects the real distributions and parameters of variables related to individuals residing in each area of the city (i.e., mean, proportion, and variance of the citizens’ age, sex, employment status, education level, ethnicity, marriage status, and country of birth). The measure of multiple deprivation captures the overall level of poverty in each area. Then, based on parameters derived from the Crime Survey for England and Wales (CSEW) 2011/2012, we simulate the victimization of these individuals across social groups and areas and predict the likelihood of these crimes being known to the police. This allows us to compare the relative difference between all crimes and police-recorded incidents at the different spatial scales.

The main motivation for using a simulation study with synthetic data, instead of simply using crime records, is because the absolute number of crimes in places is an unknown figure, regardless which source of data we use (see sect. “ Geographic crime analysis and measurement error ”). Police records are affected by a diverse array of sources of error which vary between areas, and the CSEW sample is only designed to allow the production of reliable estimates at the level of police force areas (smaller areas are unplanned domains with very small sample sizes for which analyses based on direct estimates lead to unreliable outputs; Buil-Gil et al. 2021 ). Nevertheless, the analytical steps followed in this article are designed to provide an answer to our research question (namely, whether micro-level aggregates of police-recorded crime are affected by a larger risk of bias compared with larger scales), rather than producing unbiased estimates of crime in places. Future research will explore if the method used here is also a good way to produce accurate estimates of crime in places and compare these estimates with model-based estimates of crime indicators obtained from more traditional methods in small area estimation (Buil-Gil et al. 2021 ). Indeed, unbiased estimates of crime in places are needed to guide evidence-based policing and research.

In this section, we describe the data and methods used to generate the synthetic population of crimes known and unknown to police and evaluate differences between spatial scales. Section “ Generating the population and simulation steps ” outlines the data-generating mechanism and the steps of our simulation study, and in sect. “ Empirical evaluation of simulated dataset of crimes ,” we provide an empirical evaluation of the simulated dataset. We discuss methods to assess the results in sect. “ Assessing the results .”

Generating the population and simulation steps

The simulation of our synthetic population involves three steps which are described in detail below. All analyses have been programmed in R (R Core Team 2020 ), and all data and code used for this simulation study are available from a public GitHub repository (see https://github.com/davidbuilgil/crime_simulation2 ).

Step 1. Simulating a synthetic population from census data

The first step is to generate a synthetic population consistent with the social, demographic, and spatial characteristics of Manchester. We download aggregated data about residents at the output area (OA) level from the Nomis website ( https://www.nomisweb.co.uk/census/2011 ), which publishes data recorded by the UK Census 2011. For consistency, we will conduct all our analyses using information collected in 2011. From Nomis, we obtain census parameters of various variables in each OA in Manchester. OAs are the smallest geographic units for which census data are openly published in the UK. The minimum population size per OA is 40 households and 100 residents, but the average size is 125 households. We will also use other units of geography in further steps: lower layer super output areas (LSOAs), that generally contain between four and six OAs with an average population size of 1500; and middle layer super output areas (MSOAs), which have an average population size of 7200. The largest scale used are wards. In Manchester local authority, there are 1530 OAs, 282 LSOAs, 57 MSOAs, and 32 wards.

Although UK census data achieve nearly complete coverage of the population, and measurement error arising from using these data is likely to be very small, Census data are not problem-free. For instance, census non-response rates vary between age, sex, and ethnic groups (e.g., while more than 97% of females above 55 responded the census, the response rate for males aged 25 to 29 was 86%), and questionnaire items (e.g., non-response rates were 0.4% and 0.6% for sex and age questions, respectively, and 3%, 4%, and 5.7% for ethnicity, employment status, and qualifications questions). In Manchester, the census response rate was 89%. In order to adjust for non-response in census data, the Office for National Statistics used an edit and imputation system and coverage assessment and adjustment process before publishing data in Nomis (Compton et al. 2017 ; Office for National Statistics 2015 ). Census data are widely used as empirical values of demographic domains in areas for academic research and policy (Gale et al. 2017 ). From the census, we obtain the number of citizens living in each OA (i.e., resident population size), the mean and standard deviation of age by OA, and the proportion of citizens in each area with the following characteristics defined by binary variables (in parentheses, we detail the reference category): sex (male), ethnicity (white), employment status (population without any income), education (higher education or more), marriage status (married), and country of birth (born in the UK). We use this information to simulate our synthetic individual-level population and their corresponding social-demographic characteristics within each OA. Moreover, we attach the known IMD 2010 decile in each OA. This ensures that we account for both individual and area-level measures in our simulation. The IMD is a measure of multiple deprivation calculated by the UK Government from indicators of income, employment, health, education, barriers to housing and services, and crime and living environment at the small area level (McLennan et al. 2011 ). Generating these values allows us, in subsequent steps, to simulate crimes experienced by citizens, as well as the likelihood of each crime being known to the police, based on parameters obtained from survey data. We use these specific variables since these are known to be associated with crime victimization and crime reporting rates (see sect. “ Geographic crime analysis and measurement error ”). Thus, the selection of census parameters is driven by the literature review and the availability of data recorded by the census and IMD.

The variables are generated for d  = 1, …, D OAs and i  = 1, …, N d individual citizens according to the distributions detailed below, where N d denotes the population dimension in the d th OA:

\( \mathrm{Ag}{\mathrm{e}}_{di}\sim N\left({\mu}_d^{\mathrm{Age}},{\sigma}_d^{2,\mathrm{Age}}\right),\kern0.5em \) where \( {\mu}_d^{\mathrm{Age}} \) and \( {\sigma}_d^{2,\mathrm{Age}} \) denote the mean and variance of age for the d th OA.

\( \mathrm{Se}{\mathrm{x}}_{di}\sim \mathrm{Bernoulli}\left({\pi}_d^{\mathrm{Male}}\right) \) , where \( {\pi}_d^{\mathrm{Male}} \) denotes the proportion of males in d th OA.

\( {\mathrm{NoInc}}_{di}\sim \mathrm{Bernoulli}\left({\pi}_d^{\mathrm{NoInc}}\right) \) , where \( {\pi}_d^{\mathrm{NoInc}} \) denotes the proportion of citizens without any income in the d th OA.

\( {\mathrm{HE}}_{di}\sim \mathrm{Bernoulli}\left({\pi}_d^{\mathrm{HE}}\right) \) , where \( {\pi}_d^{\mathrm{HE}} \) denotes the proportion of citizens with high education (holding a university degree) in the d th OA.

\( \mathrm{Whit}{\mathrm{e}}_{di}\sim \mathrm{Bernoulli}\left({\pi}_d^{\mathrm{White}}\right) \) , where \( {\pi}_d^{\mathrm{White}} \) denotes the proportion of white citizens in the d th OA.

\( {\mathrm{Married}}_{di}\sim \mathrm{Bernoulli}\left({\pi}_d^{\mathrm{Married}}\right) \) , where \( {\pi}_d^{\mathrm{Married}} \) denotes the proportion of married population in the d th OA.

\( {\mathrm{BornUK}}_{di}\sim \mathrm{Bernoulli}\left({\pi}_d^{BornUK}\right) \) , where \( {\pi}_d^{\mathrm{BornUK}} \) denotes the proportion of population born in the UK in the d th OA.

Thus, we generate N = 503,127 units with their individual and contextual characteristics across D = 1,530 OAs in Manchester. Given that we simulate all individual information based on population parameters obtained from the census using small spatial units of analysis (i.e., OAs), our synthetic population is very similar (in terms of distributions and ranking) to the empirical population of each OA. The Spearman’s rank correlation coefficient of the mean of age, sex, income, higher education, ethnicity, marriage status, and country of birth across areas in census data and our simulated dataset is almost perfect (i.e., larger than 0.99 for all variables).

Step 2. Simulating crime victimization from CSEW data

We use parameters obtained from the CSEW 2011/2012 to generate the crimes experienced by each individual citizen. The CSEW is an annual victimization survey conducted in England and Wales. Its sampling design consists of a multistage stratified random sample by which a randomly selected adult (aged 16 or more) from a randomly selected household is asked about experienced victimization in the last 12 months (Office for National Statistics 2013 ). The survey also includes questions about crime reporting to the police and whether each crime took place in the local area, among others. The main part of the survey is completed face-to-face in respondents’ households, although some questions (about drugs and alcohol use, and domestic abuse) are administered via computer-assisted personal interviewing. The CSEW sample size in 2011/2012 was 46,031 respondents.

In order to simulate the number of crimes faced by each individual unit within our synthetic population of Manchester residents, we first estimate negative binomial regression models of crime victimization from CSEW data and then use the model parameter estimates to predict crime incidence within our simulated population. Given that different crime types are known to be associated with different social and contextual variables (Andresen and Linning 2012 ; Quick et al. 2018 ), and the variables associated with crime reporting to the police also vary according to crime type (Baumer 2002 ; Hart and Rennison 2003 ; Tarling and Morris 2010 ), we estimate one negative binomial regression model by each of four groups of crime types:

Vehicle crimes: includes the number of (a) thefts of motor vehicles, (b) things stolen off vehicles, and (c) vehicles tampered or damaged, all during the last 12 months.

Residence crimes: number of times (a) someone entered a residence without permission to steal, (b) someone entered a residence without permission to cause damage, (c) someone tried to enter a residence without permission to steal or cause damage, (d) anything got stolen from a residence, (e) anything stolen from outside a residence (garden, doorstep, garage), and (f) anything damaged outside a residence. These refer to events happening both at the current and previous households during the last 12 months.

Theft and property crimes (excluding burglary) : number of times (a) something stolen out of hands, pockets, bags, or cases; (b) someone tried to steal something out of hands, pockets, bags, or cases; (c) something stolen from a cloakroom, office, car or anywhere else; and (d) bicycle stolen, all during the last 12 months.

Violent crimes: number of times (a) someone deliberately hit the person with fists or weapon or used force or violence in any way, (b) someone threatened to damage or use violence on the person or things belonging to the person, (c) someone sexually assaulted or attacked the person, and (d) some member of the household hit or used weapon, or kicked, or used force in any way on the person, all during the last 12 months.

Thus, this approach assumes that distributions and slopes observed in the CSEW at a national level apply to crimes that take place in Manchester local authority. The CSEW sample for Manchester is not large enough to estimate accurate regression models, and thus, we use models estimated at a national level to estimate parameters used to generate crimes at a local level. The implications of taking this approach are further discussed in sect. “ Empirical evaluation of simulated dataset of crimes ”. To alleviate the concern about this potential limitation, we show in Appendix Table 7 that the negative binomial regression model of crime victimization estimated from respondents residing in urban and metropolitan areas (excluding London) shows very similar results to model results estimated from all respondents in England and Wales.

The negative binomial regression model is a widely adopted model in this context, which has been proven to adjust well to the skewness of crime count variables (Britt et al. 2018 ; Chaiken and Rolph 1981 ). To estimate the negative binomial regression models, we use the same independent variables described in step 1 (i.e., age, sex, employment status, education level, ethnic group, marriage status, country of birth, IMD decile). However, in this step, these are taken from the CSEW. This allows us to obtain the regression model coefficient estimates and dispersion parameter estimates (Table 1 ), denoted by \( {\hat{\ \alpha}}_p \) for a generic p independent variable and \( \hat{\ \theta } \) , respectively, that will be used to generate the crime counts per person in the synthetic population. Thus, regression models consider individual and area-level variables typically associated with crime victimization risk and crime reporting, but these do not account for other area-level contextual attributes associated with crime and crime reporting, such as the presence of crime generators and attractors in the area (Brantingham and Brantingham 1995 ). Since this is a new methodological approach, we include only a small number of variables recorded in the census and IMD to keep the model parsimonious, avoid multicollinearity, and improve the model accuracy. Models do not consider other important factors, such as individuals’ routine activities and alcohol consumption, because these are not recorded in the census.

Table 1 shows the negative binomial regression models used to estimate crime victimization from CSEW 2011/2012 data. Measures of pseudo- R 2 and normalized root mean squared error (NRMSE) indicate a good fit and accuracy of our models. We use the estimated regression coefficients to generate our synthetic population of crimes, but these also provide some information about which individual characteristics are associated with a higher or lower risk of victimization by crime type. For example, age is negatively associated with crime victimization in all crime types. Being male is a good predictor of suffering vehicle and property crimes, but not residence or violent crimes. With regards to income levels, those with some type of income have a higher risk of victimization by vehicle and violent crimes, whereas respondents without any income have a higher risk of suffering residence crimes. Citizens with a higher education degree generally suffer more property and vehicle crimes than residents without university qualifications, whereas those without higher education certificates are at a higher risk of suffering violent crimes. Married citizens tend to suffer more vehicle crimes, while non-married suffer more property and violent crimes. Citizens born in the UK experience more residence and vehicle crimes than immigrants. And areas with high values of deprivation concentrate more vehicle, residence, and property crimes.

Crime victimization counts for each unit in the simulated population are generated following a negative binomial regression model using the regression coefficient and dispersion parameter estimates obtained from the CSEW (Table 1 ) and the independent variables simulated in step 1. For example, we predict the number of vehicle crimes (Vehi i ) suffered by a given individual i as follows:

where NB denotes the negative binomial distribution, and:

We repeat this procedure for all four crime types. Thus, the variability and relationships between variables observed in the CSEW are reproduced in our simulated population, and we assume that these values represent the true extent of crime victimization in the population of Manchester. We evaluate the quality of the synthetic population of crimes in sect. “ Empirical evaluation of simulated dataset of crimes .”

Step 3. Simulating crimes known to police from CSEW data

The third step consists of estimating whether each simulated crime is known to the police or not. This allows us to analyze the difference between all crimes (generated in step 2), and those crimes known to the police (to be estimated in step 3) for each area in Manchester. First, we create a new dataset in which every crime generated in step 2 becomes the observational unit. Here, our units of analysis are crimes in places, instead of individual citizens; therefore, some residents may be represented more than once (i.e., those who suffered multiple forms of victimization).

In order to estimate the likelihood of each crime being known to the police, we follow a similar procedure as in step 2, but in this case, we make use of logistic regression models for binary outcomes, which are better described by the Bernoulli distribution of crime reporting. First, we estimate a logistic regression model of whether crimes are known to police or not. We use the CSEW dataset of crimes ( n  = 14,758), and fit the model using the same independent variables as in step 2 to estimate the likelihood of crimes being known to the police (see the results of logistic regression models in Table 2 ). We estimate one regression model per crime types to account for the fact that the crime type and incident seriousness are strongly linked to crime reporting (Baumer 2002 ; Xie and Baumer 2019b ). The CSEW asks each victim of each crime whether “Did the police come to know about the matter?” We use this measure to estimate our regression models. Thus, here, we estimate if the police knows about each crime, which is not always due to crime reporting (i.e., estimates from the CSEW 2011/2012 indicate that 32.2% of crimes known to the police were reported by another person, 2.3% were witnessed by the police and 2.2% were discovered by the police by another way).

Second, we estimate whether each crime in our simulated dataset is known to the police, following a Bernoulli distribution from the regression coefficient estimates shown in Table 2 and the independent variables simulated in step 1. As in the previous case, we repeat this procedure for each crime type, since some variables may affect some crime types in a different way than others (Xie and Baumer 2019a ). For example, to estimate whether each vehicle crime j , suffered by an individual i , is known to police (KVehi ji ), we calculate:

\( {\hat{\gamma}}_p \) denotes the regression model coefficient estimate for a p independent variable, and J denotes all simulated crimes. Measures of pseudo-R 2 show a good fit of models.

One important constraint of crime estimates produced from the CSEW is that these provide information about area victimization rates (i.e., number of crimes suffered by citizens living in one area, regardless of where crimes took place), instead of area offence rates (i.e., number of crimes taking place in each area). This may complicate efforts to compare and combine survey-based estimates with police records. Given that our simulated dataset of crimes is based on CSEW parameters and census data about residential population characteristics, our synthetic dataset of crimes is also likely to be affected by this limitation. In order to mitigate the impact of this shortcoming on any results drawn from our study, we follow similar steps as in step 3 in order to estimate whether each crime took place in the residents’ local area or somewhere else and remove from the study all those crimes that do not take place within 15-min walking distance from the citizens’ household (see Appendix 2). Our final sample size is 452,604 crimes distributed across 1530 OAs in Manchester. This facilitates efforts to compare our simulated dataset of crimes with police-recorded incidents, but we note that our synthetic dataset does not account for those crimes that take place in an area but are suffered by persons living in any other place. According to estimates drawn from the CSEW 2011/2012, this represents 26.0% of all crimes, which are likely to be overrepresented in commercial areas and business districts in the city center, where the difference between the workday population and the number of residents is generally very large (e.g., 490.2% in Manchester city center; Manchester City Council 2011 ). We return to this point in the discussion section to discuss ways in which this shortcoming may be further addressed in future research.

Empirical evaluation of simulated dataset of crimes

Once all synthetic data are generated, we use victimization data recorded by the CSEW and data about crimes known to Greater Manchester Police (GMP) to empirically evaluate whether our simulated dataset of crimes matches the empirical values of crime. This is used to evaluate the quality of our synthetically generated dataset of crimes.

First, Table 3 compares the average number of crimes suffered by individuals across socio-demographic groups as recorded by the CSEW 2011/2012 and our simulated dataset. The distribution of the synthetic dataset of crimes is very similar to that of the CSEW, but values appear to be slightly larger in the synthetic population than in the survey data. For instance, citizens younger than 35 suffer the most crimes in both datasets, and males suffer more vehicle, residence, and property crimes. Crime victimization differences by ethnicity, employment status, education level, marriage status, country of birth, and IMD decile shown in the CSEW are also observed in the simulated dataset of crimes. In the case of residence crimes, incidences in our simulated population appear to be slightly larger than those observed in the CSEW. We note that our simulated dataset refers to crimes taking place in Manchester local authority, whereas the CSEW reports data for all England and Wales. In 2011/2012, the overall rate of crimes known to police per 1000 citizens was notably larger in Manchester than in the rest of England and Wales (Office for National Statistics 2019 ), and the Crime Severity Score for 2011/2012 (an index that ranks the severity of crimes in each local authority) was 104.6% larger in Manchester than the average of England and Wales (Office for National Statistics 2020 ). Therefore, the differences observed between CSEW and our synthetic population of crimes are likely to reflect true variations between the crime levels in Manchester and England and Wales as a whole.

Second, Table 4 presents the proportion of crimes that are known to the police grouped by the socio-demographic and contextual characteristics of victims in CSEW and our simulated data. By looking at the table, we see that the proportions related to the CSEW are very similar to the ones obtained on the simulated data. This shows that modeling results are consistent, thus preserving relationships between variables.

Third, we download crime data recorded by GMP ( https://data.police.uk/ ) and compare area-level aggregates of crimes known to GMP with our synthetic dataset of crimes known to the police. To do this, we only consider those simulated crimes that were estimated as being known to police and taking place in the local area. Spearman’s rank correlation and Global Moran’s I coefficients between the area-level aggregates of our synthetic dataset of crimes and crimes known to GMP are reported in Table 5 . Tiefelsdorf’s ( 2000 ) exact approximation of the Global Moran’s I test is used as a measure of spatial dependency between the two measures, to analyze if the number of crimes in our simulated dataset is explained by the value of crimes known to GMP in surrounding areas (Bivand et al. 2009 ).

We aggregate all crimes known to police to each spatial unit using the “sf” package in R (Pebesma 2018 ). Out of the 87,457 crimes known to GMP, 642 could not be geocoded. We note that we obtained slightly different results using two different analytical approaches to aggregating crimes in areas (i.e., counting crimes in OAs and then aggregating from OAs to LSOA, MSOAs, and wards using a lookup table, versus counting crimes in OAs, LSOAs, MSOAs, and wards, respectively), which may be due to errors arising from the aggregation process or inconsistencies in the lookup table. We chose the second approach (i.e., counting points in polygons at the different scales), since, on average, a larger number of offences were registered in each area using this method. Tompson et al. ( 2015 ) demonstrate that open crime data published in England and Wales is spatially precise at the levels of LSOA and MSOA, but that the spatial noise added to these data for the purposes of anonymity means that OA-level maps often have inadequate precision. Thus, we only present and discuss the results obtained at LSOA and larger spatial levels.

Table 5 shows positive and statistically significant coefficients of Spearman’s rank correlation for all crime types at the LSOA level. The index of Global Moran’s I is also statistically significant and positive in all cases. At the MSOA and ward levels, the coefficients of Spearman’s correlation for vehicle crimes are not statistically significant. This is likely to be explained by the small number of MSOAs and wards under study (56 and 32, respectively). Generally speaking, our simulated dataset of synthetic crimes is a good indicator of crimes known to police, although both datasets are not perfectly aligned. Our synthetic dataset of crimes may underestimate crimes known to police in areas with a large difference between workday and residential populations, but it appears to be a precise indicator of crimes known to police in residential areas. In the discussion section, we present some thoughts about how to address this in future research.

Assessing the results

In order to assess the extent to which the number of simulated crimes known to police varies from all simulated crimes at the different spatial scales, we calculate the absolute percentage relative difference (RD) and the percentage relative bias (RB) between these two values for each crime type in each area at four spatial scales.

First, RD is calculated for every area d in the specified level of geography (i.e., Geo = {OA, LSOA, MSOA, wards}), as follows:

where E d denotes the count of all crimes in area d , and K d is the count of crimes known to police in the same area.

Second, RB is computed as follows:

We evaluate the average RD and RB at the different spatial scales, but also their spread, to establish if the measures of dispersion across areas become larger when the geographic scale becomes smaller. This permits a demonstration not just of the mean differences between all crimes and crimes known to police at different spatial scales but also the variability in these differences, to help shed light on whether there is higher variability at fine-grained spatial scales. This is investigated via the standard deviation (SD), minimum, maximum, and mean of the RD and RB at the different scales. In addition, boxplots and maps are shown to visualize outputs.

Mapping the bias of police-recorded crimes

This section presents the results of the simulation study. More specifically, we analyze the mean, minimum, maximum, and SD of the RD and RB between all simulated crimes and those synthetic crimes known to the police. We present analyses at the levels of OAs, LSOAs, MSOAs, and wards for four different crime types, in order to establish if the variability of the RD and RB becomes larger at more fine-grained spatial scales.

First, Table 6 presents the summary statistics of RD and RB for all crime types across the four spatial scales. On average, the RD is close to 62% at all the spatial scales (i.e., on average, 62% of crimes are unknown to police at each spatial scale), but the measures of dispersion—and the minimum and maximum values—vary considerably depending on the spatial level under study. The SD of the RD between all crimes and police-recorded offences is the largest at the level of OAs, whereas it is much smaller when crimes are aggregated at the LSOA level. It becomes almost zero at the level of MSOAs and wards. In other words, the RD has a large variability across small areas, but it is minimal when using larger geographies. In one OA, the police might be aware of the vast majority of crimes, and in another one, very few. Thus, geographic crime analysis produced solely from police records at highly localized spatial scales, such as OAs, and may show high concentrations of crime in some areas, but simply as an artefact of the variability in the crimes known to police. By contrast, the police know roughly the same proportion of crimes in all MSOAs and wards, with little variation around the mean. This is also observed in the minimum and maximum values. As such, documenting community differences in crimes based on police records aggregated at these larger scales will reduce the risk of mistakenly classifying some areas as high-crime density, but not others.

Similarly, the mean RB between all crimes and crimes known to police is roughly the same across all spatial scales, but the SD of the RB varies across levels of analysis. The SD is very large when crimes are aggregated at the level of OAs compared with larger scales.

Results shown in Table 6 , nevertheless, are produced from all crime types merged together and thus are likely to hide important heterogeneity depending on each type of crime under study. Crime research shows that different crime types are affected by different individual and contextual predictors (Andresen and Linning 2012 ; Quick et al. 2018 ), and there are also differences in terms of crime reporting to the police (Tarling and Morris 2010 ). Therefore, some crime types may be less affected by data biases than others, and it may be beneficial to disaggregate results by crime type in order to observe differences that may otherwise remain hidden.

Figure 1 shows boxplots of the RD between all crimes (known and unknown to police) and police-recorded crimes across crime types and spatial scales. Detailed results on this are also shown in Appendix Table 9 . We observe that, on average, the RD is lower for violent crimes than any other crime type. Thus, the proportion of total crime known to police is generally larger in the case of violent crimes. We also see that the measures of dispersion in the RD are much larger in the case of property crimes than all other crime types, while the variance of the RD of residence crimes appears to be the smallest. In the case of property crimes, for example, we observe that there is one OA with a RD equal to zero and another area with a RD equal to 100. In other words, in one OA, all property crimes were known to the police, while in the other small area not a single crime was known to police forces. Regardless of the crime type, larger levels of geography are associated with a smaller variance in the RD between areas, whereas the difference between the RD of crime aggregates for MSOAs or wards is generally small. In summary, geographic analysis produced from police records at larger spatial scales may show a more valid representation of the geographic distribution of crimes (known and unknown to police) than analysis produced for small areas.

Boxplots of RD% between all crimes and crimes known to police at the different spatial scales (simulated dataset)

In order to better illustrate the impact of selection bias on maps produced at the different spatial scales, Fig. 2 visualizes the values of RD between all property crimes and property crimes known to the police at the level of OAs, LSOAs, MSOAs, and wards in Manchester. We produce maps of property crimes since it is the crime type with the most extreme measures of dispersion in terms of RD, but similar—less extreme—results are also observed for other crime types. Figure 2 shows that the RD varies widely across OAs (i.e., in some areas, no crimes are known to police, and in others, nearly every crime is known to the police), while the RD between all crimes and police-recorded crimes becomes very homogeneous when crimes are aggregated at the scales of MSOAs and wards.

Maps of RD% between all property crimes and property crimes known to police at the different spatial scales (simulated dataset). Breaks based on equal intervals

Discussion and conclusions

Crime analysis and crime mapping researchers are moving toward increasingly fine-grained geographic resolutions to study the urban crime problem and to design spatially targeted policing strategies (Braga et al. 2018 ; Groff et al. 2010 ; Kirkpatrick 2017 ; Weisburd et al. 2012 ). Researchers document and explain community differences in crime to generate knowledge about crime patterns, test ideas, and assess interventions. Nevertheless, aggregating crimes known to police at such detailed levels of analysis increases the risk that the data biases inherent in police records reduce the accuracy of research outputs. These biases may contribute to the misallocation of police resources, and ultimately have an impact on the lives of those who reside in places mistakenly defined as high-crime-density or low-crime-density areas (Skogan 1977 ). They may also affect the validity of analyses which test theoretical explanations for the geographic distribution of crime (Gibson and Kim 2008 ).

This issue around the bias of police-recorded crime data largely depends on residents’ willingness to report crimes to police, and the police capacity to control places. Both are known to be affected by social and contextual conditions that are more prevalent in some areas than others (Berg et al. 2013 ; Goudriaan et al. 2006 ; Jackson et al. 2013 ; Slocum et al. 2010 ; Xie and Lauritsen 2012 ). The demographic and social characteristics of micro-places are usually very homogeneous (Brattbakk 2014 ; Oberwittler and Wikström 2009 ), which means that populations unwilling to report crime and cooperate with the police will concentrate in particular places, while other areas may contain social groups that are much more inclined to report crime and work with the police. The influence of these factors is reduced when crimes are aggregated to meso- and macro-levels of spatial analysis with more heterogeneous populations. Our simulation study shows that aggregates of police-recorded crime produced for neighborhoods and wards show a much more accurate—less biased—image of the geography of crime compared with those aggregated to small areas. This can be attributed to greater variability (i.e., between-unit heterogeneity) in the proportion of crimes known to police at fine-grained spatial scales. This study also demonstrates that some crime types are affected by data bias differently, which demonstrates the need to disaggregate analyses by crime types.

However, our simulation study is also affected by some limitations that could be addressed in future research. Namely, our simulated dataset of crimes captures area victimization rates instead of area crime rates and, as a consequence, the empirical evaluation when comparing synthetically generated crimes with actual crimes known to GMP showed that our synthetic dataset could be further improved in those areas with a large difference between workday and residential populations. In order to mitigate against this shortcoming, future research should investigate replicating this analysis using census data for workday populations instead of census data for residential populations. This may allow for the generation of more accurate crime counts, especially in non-residential places where crime is prevalent, such as the city center and commercial districts. Moreover, since the CSEW sample in Manchester is very small, our approach assumed that slopes observed in regressions estimated from the CSEW at a national level apply to crimes in Manchester. Future research may merge several editions of the CSEW to obtain a large enough sample in Manchester. Nevertheless, in such a case, survey and census data would refer to different time periods, and there would be a risk of repeated respondents in the CSEW. There are three further limitations that may have more difficult solutions: (a) the CSEW and most victimization surveys do not record information of so-called victimless crimes (e.g., drug-related offences, corporate crimes) and homicides, for which generating synthetic estimates may be more complicated; (b) the sample of the CSEW consists of adults aged 16 or more, and thus it may be difficult to accurately generate crimes faced by individuals younger than 16 years; and (c) the census is only conducted every 10 years and generating periodic synthetic populations to estimate crime will require the implementation of novel techniques (e.g., spatial microsimulation models; Morris and Clark 2017 ). Future research will also explore the use of other individual and contextual variables recorded in the census and other data sources to further improve the precision of synthetic crime data. Moreover, this approach could be applied to other urban areas with available local crime surveys (e.g., Islington Crime Survey, Metropolitan Police Public Attitudes Survey) which would allow for an empirical evaluation of synthetic crime data generated in each local area.

Those who advocate the need for documenting and explaining micro-level community differences in crime have well-sustained arguments to claim that aggregating crimes at fine-grained levels of spatial analysis allows for better explanations of crime, and more targeted operational policing practices. To mention only a few of their arguments, Oberwittler and Wikström ( 2009 ) show that between-neighborhood crime variance and the statistical power of research outputs increase when smaller units of analysis are used; Steenbeek and Weisburd ( 2016 ) show that most temporal variability in crimes known to police can be attributed to micro-scales; Braga et al. ( 2018 ) show that increasing police control in high-crime-density areas reduces the overall prevalence and incidence of crimes; and Weisburd et al. ( 2012 ) argue that the social systems relevant to understanding the crime problem concentrate in small units of geography. It is not our intention to dismiss the merits of micro-level geographic crime analysis, nor do we directly assess whether the claims made by the advocates of micro-level mapping remain verifiable when analyzing unbiased datasets of crime (this is, perhaps, an area for future research). That said, the results reported in this paper serve to raise awareness about an important shortcoming of micro-level crime analysis. There is a clear need for academics and police administrations to evaluate whether crime rates are associated with conditions external to victimization. In particular, there is a need to make this evaluation with consideration for the spatial scale being used (Ramos et al. 2020 ). The potential sources of bias in police-recorded crime data should always be investigated and acknowledged with this in mind. Further efforts might focus on developing techniques which mitigate against these sources of bias to ensure that geographic crime analysis remains an effective tool in understanding and tackling the crime problem.

Aebi, M. F., & Linde, A. (2012). Conviction statistics as an indicator of crime trends in Europe from 1990 to 2006. European Journal on Criminal Policy and Research, 18 , 103–144.

Article   Google Scholar  

Andresen, M. A., & Linning, S. J. (2012). The (in)appropriateness of aggregating across crime types. Applied Geography, 35 (1-2), 275–282.

Baumer, E. P. (2002). Neighborhood disadvantage and police notification by victims of violence. Criminology, 40 (3), 579–616.

Berg, M. T., Slocum, L. A., & Loeber, R. (2013). Illegal behaviour, neighborhood context, and police reporting by victims of violence. Journal of Research in Crime and Delinquency, 50 (1), 75–103.

Bivand, R., Müller, W. G., & Reder, M. (2009). Power calculations for global and local Moran’s I. Computational Statistics & Data Analysis, 53 (8), 2859–2872.

Bowers, K., & Johnson, S. D. (2014). Crime mapping as a tool for security and crime prevention. In M. Gill (Ed.), The Handbook of Security (2nd ed., pp. 566–587). London: Palgrave Macmillan.

Chapter   Google Scholar  

Braga, A. A., Weisburd, D., & Turchan, B. (2018). Focused deterrence strategies and crime control: An updated systematic review and meta-analysis of the empirical evidence. Criminology & Public Policy, 17 (1), 205–250.

Brantingham, P. J. (2018). The logic of data bias and its impact on place-based predictive policing. Ohio State Journal of Criminal Law, 15 (2), 473–486.

Google Scholar  

Brantingham, P. L., & Brantingham, P. J. (1995). Criminality of place. Crime generators and crime attractors. European Journal of Criminal Policy and Research, 3 , 5–26.

Brantingham, P. L., & Brantingham, P. J. (2004). Computer simulation as a tool for environmental criminologists. Security Journal, 17 , 21–30.

Brattbakk, I. (2014). Block, neighbourhood or district? The importance of geographical scale for area effects on educational attainment. Geografiska Annaler: Series B, Human Geography, 96 (2), 109–125.

Britt, C. L., Rocque, M., & Zimmerman, G. M. (2018). The analysis of bounded count data in criminology. Journal of Quantitative Criminology, 34 , 591–607.

Bruinsma, G. J. N., & Johnson, S. D. (Eds.). (2018). The Oxford handbook of environmental criminology . New York: Oxford University Press.

Buil-Gil, D., Medina, J., & Shlomo, N. (2021). Measuring the dark figure of crime in geographic areas. Small area estimation from the Crime Survey for England and Wales. British Journal of Criminology, 61 (2), 364–388.

Chaiken, J. M. & Rolph, J. E. (1981). Methods for estimating crime rates of individuals. The Rand Corporation. Report R-2730//1-NIJ . Retrieved from https://www.rand.org/content/dam/rand/pubs/reports/2009/R2730.1.pdf .

Compton, G., Wilson, A., & French, B. (2017). The 2011 census: From preparation to publication. In J. Stillwell (Ed.), The Routledge handbook of census resources, methods and applications (pp. 33–53). London: Routledge.

de Candolle, A. (1830 [1987a]). Considérations sur la statistique des délits. Déviance et Société , 11 (4), 352–355.

de Candolle, A. (1832 [1987b]). De la statistique criminelle. Déviance et Société , 11 (4), 356-363.

Gale, C., Singleton, A., & Longley, P. (2017). Creating a new open geodemographic classification of the UK using 2011 census data. In J. Stillwell (Ed.), The Routledge handbook of census resources, methods and applications (pp. 213–229). London: Routledge.

Gibson, J., & Kim, B. (2008). The effect of reporting errors on the cross-country relationship between inequality and crime. Journal of Development Economics, 87 (2), 247–254.

Goudriaan, H., Wittebrood, K., & Nieuwbeerta, P. (2006). Neighbourhood characteristics and reporting crime: Effects of social cohesion, confidence in police effectiveness and socio-economic disadvantage. British Journal of Criminology, 46 (4), 719–742.

Groff, E. R., & Mazerolle, L. (2008). Simulated experiments and their potential role in criminology and criminal justice. Journal of Experimental Criminology, 4 , 187.

Groff, E. R., Weisburd, D., & Yang, S. M. (2010). Is it important to examine crime trends at a local “micro” level?: A longitudinal analysis of street to street variability in crime trajectories. Journal of Quantitative Criminology, 26 , 7–32.

Hart, T. C. & Rennison, C. (2003). Reporting crime to the police, 1992-2000. Special Report. Bureau of Justice Statistics . Retrieved from https://static.prisonpolicy.org/scans/bjs/rcp00.pdf .

HMIC. (2014). Crime-recording: Making the victim count. The final report of an inspection of crime data integrity in police forces in England and Wales. HMIC Report . Retrieved from https://www.justiceinspectorates.gov.uk/hmicfrs/wp-content/uploads/crime-recording-making-the-victim-count.pdf

Jackson, J., Bradford, B., Stanko, B., & Hohl, K. (2013). Just authority? Trust in the police in England and Wales . Abingdon: Routledge.

Kirkpatrick, K. (2017). It’s not the algorithm, it’s the data. Communications of the ACM, 60 (2), 21–23.

Kitsuse, J. I., & Cicourel, A. V. (1963). A note on the uses of official statistics. Social Problems, 11 (2), 131–139.

Lee, Y., Eck, J. E., & O, S. & Martinez, N. N. (2017). How concentrated is crime at places? A systematic review from 1970 to 2015. Crime Science, 6 (6).

Manchester City Council. (2011). Workday population summary: 2011 census. Manchester City Council Report . Retrieved from https://www.manchester.gov.uk/download/downloads/id/25545/q05q_2011_census_summary_-_workday_population.pdf .

McCandless, R., Feist, A., Allan, J. & Morgan, N. (2016). Do initiatives involving substantial increases in stop and search reduce crime? Assessing the impact of Operation BLUNT 2. Home Office Report . Retrieved from https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/508661/stop-search-operation-blunt-2.pdf

McLennan, D., Barnes, H., Noble, M., Davies, J., Garratt, E. & Dibben, C. (2011). The English Indices of Deprivation 2010. Department for Communities and Local Government . Retrieved from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/6320/1870718.pdf

Moretti, A. (2020). Simulation studies. In I. P. Atkinson, S. Delamont, A. Cernat, R. Williams, & J. Sakshaug (Eds.), SAGE Research Methods Foundations . SAGE.

Morris, M. A., & Clark, S. (2017). A big data application of spatial microsimulation for neighborhoods in England and Wales. In L. Schintler & Z. Chen (Eds.), Big data for regional science . London: Routledge.

Oberwittler, D., & Wikström, P. O. H. (2009). Why small is better: Advancing the study of the role of behavioral contexts in crime causation. In D. Weisburd, W. Bernasco, & G. J. N. Bruinsma (Eds.), Putting crime in its place: Units of analysis in geographic criminology (pp. 35–60). New York: Springer.

Office for National Statistics. (2013). Crime Survey for England and Wales, 2011-2012 . UK Data Service, SN 7252. https://doi.org/10.5255/UKDA-SN-7252-2

Office for National Statistics. (2015). 2011 Census: General report for England and Wales. Office for National Statistics . Retrieved from https://www.ons.gov.uk/file?uri=/census/2011census/howourcensusworks/howdidwedoin2011/2011censusgeneralreport/2011censusgeneralreportforenglandandwalesfullreport_tcm77-384987.pdf .

Office for National Statistics. (2019). Recorded crime data at Community Safety Partnership and local authority level. Office for National Statistics . Retrieved from https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/datasets/recordedcrimedataatcommunitysafetypartnershiplocalauthoritylevel .

Office for National Statistics (2020). Crime Severity Score (experimental statistics). Office for National Statistics . Retrieved from https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/datasets/crimeseverityscoreexperimentalstatistics .

Pebesma, E. (2018). Simple features for R: standardized support for spatial vector data. The R Journal, 10 (1), 439–446.

Pierce, G. L., Spaar, S., & Briggs, L. R. (1988). The character of police work: Strategic and tactical implications . Boston: Northeastern University.

Quick, M., Li, G., & Brunton-Smith, I. (2018). Crime-general and crime-specific spatial patterns: A multivariate spatial analysis of four crime types at the small-area scale. Journal of Criminal Justice, 58 , 22–32.

R Core Team. (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing . Retrieved from https://www.R-project.org/ .

Ramos, R. G., Silva, B. F. A., Clarke, K. C., & Prates, M. (2020). Too fine to be good? Issues of granularity, uniformity and error in spatial crime analysis. Journal of Quantitative Criminology .

Schnebly, S. (2008). The influence of community-oriented policing on crime-reporting behaviour. Justice Quarterly, 25 (2), 223–251.

Sellin, T. (1931). The basis of a crime index. Journal of Criminal Law and Criminology, 22 (3), 335–356.

Sherman, L. W., Gartin, P. R., & Buerger, M. E. (1989). Hot spots of predatory crime: Routine activities and the criminology of place. Criminology, 27 (1), 27–55.

Skogan, W. S. (1974). The validity of official crime statistics: An empirical investigation. Social Science Quarterly, 55 , 25–38.

Skogan, W. G. (1977). Dimensions of the dark figure of unreported crime. Crime & Delinquency, 23 (1), 41–50.

Slocum, L. A., Taylor, T. J., Brick, B. T., & Esbensen, F. A. (2010). Neighborhood structural characteristics, individual-level attitudes, and youths’ crime reporting intentions. Criminology, 48 (4), 1063–1100.

Steenbeek, W., & Weisburd, D. (2016). Where the action is in crime? An examination of variability of crime across different spatial units in The Hague, 2001–2009. Journal of Quantitative Criminology, 32 , 449–469.

Tarling, R., & Morris, K. (2010). Reporting crime to the police. British Journal of Criminology, 50 (3), 474–490.

Tiefelsdorf, M. (2000). Modelling spatial processes. The identification and analysis of spatial relationships in regression residuals by means of Moran’s I . Berlin: Springer.

Tompson, L., Johnson, S., Ashby, M., Perkins, C., & Edwards, P. (2015). UK open source crime data: accuracy and possibilities for research. Cartography and Geographic Information Science, 42 (2), 97–111.

Townsley, M., & Birks, D. J. (2008). Building better crime simulations: systematic replication and the introduction of incremental complexity. Journal of Experimental Criminology, 4 , 309–333.

UK Statistics Authority. (2014). Assessment of compliance with the code of practice for official statistics. Statistics on crime in England and Wales. Assessment Report 268 . Retrieved from https://www.statisticsauthority.gov.uk/wp-content/uploads/2015/12/images-assessmentreport268statisticsoncrimeinenglandandwale_tcm97-43508-1.pdf .

Weisburd, D. (2015). The law of crime concentration and the criminology of place. Criminology, 53 (2), 133–157.

Weisburd, D., Bruinsma, G. J. N., & Bernasco, W. (2009). Units of analysis in geographic criminology: Historical development, critical issues, and open questions. In D. Weisburd, W. Bernasco, & G. J. N. Bruinsma (Eds.), Putting crime in its place: Units of analysis in geographic criminology (pp. 3–34). New York: Springer.

Weisburd, D., Groff, E. R., & Yang, S. M. (2012). The criminology of place. Street segments and our understanding of the crime problem . New York: Oxford University Press.

Book   Google Scholar  

Weisburd, D., & Lum, C. (2005). The diffusion of computerized crime mapping in policing: Linking research and practice. Police Practice and Research, 6 (5), 419–434.

Xie, M. (2014). Area differences and time trends in crime reporting: Comparing New York with other metropolitan areas. Justice Quarterly, 31 (1), 43–73.

Xie, M., & Baumer, E. P. (2019a). Neighborhood immigrant concentration and violent crime reporting to the police: A multilevel analysis of data from the National Crime Victimization Survey. Criminology, 57 (2), 237–267.

Xie, M., & Baumer, E. R. (2019b). Crime victims’ decisions to call the police: Past research and new directions. Annual Review of Criminology, 2 , 217–240.

Xie, M., & Lauritsen, J. (2012). Racial context and crime reporting: A test of Black’s stratification hypothesis. Journal of Quantitative Criminology, 28 , 265–293.

Download references

Acknowledgements

The authors would like to thank Reka Solymosi for comments that greatly improved the manuscript.

This work is supported by the Campion Grant of the Manchester Statistical Society (project title: “Mapping the bias of police records”).

Author information

Authors and affiliations.

School of Social Sciences, University of Manchester, Humanities Bridgeford Street building, 176 Oxford Road, Manchester, M13 9PL, UK

David Buil-Gil

Manchester Metropolitan University, Manchester, UK

Angelo Moretti

University of Leeds, Leeds, UK

Samuel H. Langton

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to David Buil-Gil .

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

To estimate whether each crime took place in the victims’ local area or somewhere else, we follow the same procedure as in step 3. First, we estimate a logistic regression model of crimes happening in the local area (as opposed to crimes happening elsewhere) from the CSEW dataset of crimes. We use the same individual independent variables as above (see model results in Table 8 ). Second, we estimate whether each simulated crime took place in the resident’s local area or somewhere else following a Bernoulli distribution from the regression coefficient estimates presented in Table 8 and the independent variables simulated in Step 1. For example, to estimate whether vehicle crime j suffered by person i took place in local area, denoted by AVehi ji , we compute:

where \( {\hat{\beta}}_p \) is the regression model coefficient estimate for a p independent variable.

Then, we remove all those offences that did not take place in the local area from our synthetic dataset of crimes.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Buil-Gil, D., Moretti, A. & Langton, S.H. The accuracy of crime statistics: assessing the impact of police data bias on geographic crime analysis. J Exp Criminol 18 , 515–541 (2022). https://doi.org/10.1007/s11292-021-09457-y

Download citation

Accepted : 12 January 2021

Published : 26 March 2021

Issue Date : September 2022

DOI : https://doi.org/10.1007/s11292-021-09457-y

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Crime analysis
• Official Statistics
• Unreliability
• Simulation experiment
• Find a journal
• Publish with us
• Track your research

How the Pandemic Reshaped American Gun Violence

By Robert Gebeloff ,  K.K. Rebecca Lai ,  Eli Murray ,  Josh Williams and Rebecca Lieberman

Taking a stroll around the neighborhood is a routine activity for many Americans. Yet for 47 million people — about one in seven — such a walk would pass near the location of a recent gun homicide.

The number of people living this close to fatal violence grew drastically during the pandemic years, a New York Times analysis has found, as a surge in killings not only worsened gun violence in neighborhoods that were already suffering but also spread into new places.

To assess the impact of the pandemic years, The New York Times created a map of every gun homicide in the United States since 2020, using data collected from the police and news media accounts by the nonprofit Gun Violence Archive . For every block where Americans resided, The Times then drew a quarter-mile circle to determine how many people lived in close proximity to the killings.

Often, it was not just one killing, but two or three. In extreme cases, a dozen fatal shootings or more fell within those circles.

Enter your address to see how many fatal shootings took place near you:

Note: Addresses are matched to census blocks. Shootings are counted from the census internal point of each block. Base map data: © Mapbox © OpenStreetMap

“There are a lot more guns on the street and when people get angry and frustrated, instead of getting into a fistfight, they get into a gun fight,” said Dr. Regan Williams, an emergency room director at a Memphis children’s hospital who has seen a spike in young shooting victims.

Though the level of violence has fallen since the worst days of the pandemic, Americans are still shooting and killing one another more frequently than they did in the years before the coronavirus arrived. The long-term impact of the surge in violence is being felt in many corners of the nation, and researchers will undoubtedly study it for years to come.

“We’re taking a few steps back from the cliff,” said Dr. Garen J. Wintemute, an emergency room doctor who directs a violence prevention research program at the University of California, Davis. “But there are some ominous developments. What happens in a society that is increasingly violent, increasingly mistrustful, increasingly polarized, increasingly indulgent in hate rhetoric?”

The rate of fatal shootings per 100,000 residents remains above pre-pandemic levels in many places{potentialBut}

Note: Chart does not show rates for places with a population of fewer than 10,000 residents.

The Times mapped homicides to better understand not only the numbers of direct victims but also the communities most exposed. The analysis revealed that gun deaths spread into new neighborhoods during the pandemic: An additional 8.7 million Americans now live on a block near a gun homicide — a 23 percent increase from the prepandemic years.

But even as the geography of fatal shootings expanded, killings also rose sharply in the nation’s existing centers of violence. These neighborhoods saw the worst of the surge, perpetuating a pattern of concentrated violence that long predated the pandemic. More than half of all gun homicides still occurred in neighborhoods where just 6 percent of Americans live.

“You don’t want people to think that everywhere is so dangerous in a way that it’s not,” said John MacDonald, a criminologist at the University of Pennsylvania who reviewed the Times analysis. “On the other hand, you don’t want people to think that, oh, this is just somebody else’s problem. It’s not happening in my neighborhood.”

Base map data: © Mapbox © OpenStreetMap

One thing the pandemic did not change is the sharp racial disparity in the communities most exposed to fatal shootings. Black people were five times as likely to live near a gun homicide as white people, while Latinos were three times as likely, Asian Americans were twice as likely, and Native Americans were 1.4 times as likely. The violence mostly followed patterns of housing segregation, which often leaves people of color living in poorer neighborhoods where crime rates are often higher.

Gun suicides , which outnumber homicides and were not part of the Times analysis, have been rising steadily for years and reached a record number in 2022. The demography of gun suicides is vastly different, with rates higher for white men and in rural areas.

An Expanding Footprint

Criminologists have offered several explanations for the drastic rise in the number of fatal shootings during the pandemic:

A rise in gun ownership made it more likely for violent disputes to become deadly. An increase in drug use, and drug dealing, made violent conflicts more probable. The disruption of public schools abetted an expansion of youth gang activity. And an upheaval in policing led to reduced enforcement in many cities.

The police say many of these factors contributed to what happened in Senator Henry M. Jackson Park in Everett, Wash., over Thanksgiving weekend last year. Mayor Cassie Franklin was awakened at 2:30 a.m. by the sounds of a gun battle near her home, which the police attributed to a turf war between two street gangs. Police recovered more than 50 shell casings and the body of a 17-year-old boy.

Everett is a city of 110,000 north of Seattle that is a hub for aerospace manufacturing. It is one of many smaller American cities where the number of fatal shootings both increased and spread during the pandemic years.

The share of residents who lived near at least one fatal shooting rose in most communities{potentialBut}

Note: Figures may not sum due to rounding.

“Most people weren’t focused on violent crime because it was only impacting a small demographic, a small portion of our community,” Ms. Franklin said. “Now we’re seeing violent crime throughout different parts of Everett and more of our community is starting to pay attention and care about it.”

City officials are now prioritizing combating gun violence but say they have challenges: The fentanyl epidemic has spun out of control, and the police force is understaffed after state lawmakers tightened regulations on how the police can engage with criminal suspects.

Everett, Wash.

Basemap data: © Mapbox © OpenStreetMap

When George Floyd was murdered by a Minneapolis police officer early in the pandemic, it set off an anti-police protest movement around the nation. In Everett and elsewhere, the result was more difficulty in recruiting police officers to do the kind of work necessary to curb crime, said the chief of the city’s Police Department, John DeRousse.

“We were one of many states where officers became really reluctant to do their jobs,” he said. “Police departments were losing officers at rates higher than they’ve ever seen before.”

Down the West coast, the California city of Vallejo has also confronted a big spike in violence, with policing at the center of the discussion. Local officials say the city has too few officers, which has allowed gang activity to flourish. But community leaders blame long-running mistrust of police as central to the crime problem. In April, the state attorney general reached a settlement with the Vallejo force requiring a broad range of reforms , a situation spurred by years of allegations about police misconduct .

“If you were to compare us to Oakland or San Francisco, we don’t have the level of support or the same level of resources,” said Andrea Sorce, an economics professor who is running for mayor of Vallejo. “So, yeah, when something hits like the pandemic, we do get hit hard.”

Vallejo, Calif.

South Vallejo

Askari Sowonde, a professional event planner and community activist, said residents are concerned with crime but still wary of the police.

“People are angry about both,” she said. “We don’t like the fact that some of these other people are killing each other and we have to talk about that, too. But let’s also deal with these police officers. Let’s not push that away.”

Overall, the footprint of violence spread in four out of five major U.S. cities. In Atlanta, the percentage of residents exposed to nearby gun violence rose to 58 percent during the pandemic years, up from 36 percent in the four prior years. In Columbus, Ohio, the exposure went to 41 percent from 28 percent.

Pockets of Violence

Even as violence spread in cities where it had been relatively low before the pandemic, it also intensified and spread in the places that already had high homicide rates.

Memphis is one example. Fatal shootings hit a new high in 2023, and in November, a former city council member, now a state senator, wrote to the governor and asked for enforcement help, saying the city was “ under siege .”

Binghampton

Dr. Williams runs the trauma unit at Le Bonheur Children’s Hospital and said the number of children and teenagers wounded by gunfire more than doubled during the pandemic, including 96 children 5 and under who suffered gunshot wounds.

“During Covid, we were so worried about the effect that it had on older people,” she said. “But we failed to recognize the effect of our children being out of school, and being out of normal socialization.”

In poorer communities, children rely on public institutions like schools and recreation departments to provide structure, and when that support was cut back during the pandemic, poorer children were more likely to suffer the consequences. Dr. Williams said many young people dropped out of the school system when society shut down, and never rejoined.

“There’s just a lot more children in the community that don’t have any way to stay busy and be occupied, and that’s getting them into trouble,” she said.

Memphis had more than a thousand homicide victims during the pandemic but the impact was even broader, since more than 335,000 people lived on blocks in close proximity to the violence — 83 percent of them Black or Hispanic. Some researchers believe more attention should be paid to these indirect victims.

“Neighborhoods that have persistently elevated levels of violence have lots of trauma across many people,” said Nicole Kravitz-Wirtz, a sociologist on the California, Davis, violence prevention project. “That impacts relationships between neighbors and translates into collective senses of fear.”

People in poor neighborhoods and those with a large Black population were most likely to experience violence.

Population exposed to a shooting.

Black population

Note: Neighborhood data is based on census tracts. Poverty is measured using the C.D.C.’s Social Vulnerability Index.

Most major cities contained both mostly safe areas and pockets of violence. Chicago has a national reputation for high gun violence, but on the ground, nearly a third of the city’s population lived in neighborhoods with very few shootings, while more than a quarter of the residents lived on blocks where the violence was extreme.

New York and Los Angeles, meanwhile, had relatively low homicide rates overall, but those figures masked the presence of some of the nation’s most dangerous neighborhoods.

This geographic disparity was reflected in the large differences in exposure to violence for people of different races. Black people were already far more likely to live near shootings before the pandemic, so when violence spiked, they were most likely to be affected.

In Milwaukee, for example, where shootings are so frequent that more than a third of white residents lived near one, their Black neighbors had it far worse: 83 percent lived near a gun homicide.

The racial demographics are far different in the least and most violent neighborhoods.

“These disparities become especially stark when we start talking about more than one incident of gun violence during the past year,” Ms. Kravitz-Wirtz said.

Debate Over Reforms

While homicide rates are falling in many parts of the country, they are still higher than prepandemic levels, and in some places they are still going up. The policy implications are still playing out in two primary areas: the battle over gun regulations and the debate over the role of policing.

The national gun homicide rate has fallen but remains higher than at any time since the 1990s.

Source: C.D.C. | Note: The rate for 2023 is estimated using data from the Gun Violence Archive.

Congress responded to the crisis by passing bipartisan “Safer Communities” legislation in 2022. It expanded background checks for gun buyers 21 and younger, incentivized states to enact “red flag” laws to temporarily confiscate guns from people deemed to be dangerous to society, and provided hundreds of millions of dollars to community-based, street-level gun violence prevention efforts.

In addition, dozens of states strengthened gun-safety bills, such as laws requiring a permit to purchase new firearms.

“There’s been tremendous progress at the state level,” said Kelly Drane, the research director at the Giffords Law Center, a gun violence prevention group that advocates for stronger regulations. “And there is also at the same time, this competing reality that there are states that are actively weakening their laws right now.”

Many local leaders have sought to confront long-simmering tensions over policing and the way suspects are prosecuted. More than 140 justice reform bills passed in 30 states in 2020 and 2021, measures that are still controversial in some jurisdictions .

“Everything seemed to be getting at making crime less costly to commit or making law enforcement more costly to do,” said Rafael A. Mangual, a fellow at the right-leaning Manhattan Institute who studies criminal justice. “I think people are unwilling to sacrifice the level of safety that was clearly sacrificed during the pandemic years.”

Whatever happens with the law and policing, researchers worry that the pandemic has left the nation more prone to gun violence than before.

“We are, as a society, experiencing long Covid,” said Dr. Wintemute, the University of California epidemiologist. “I don’t mean the physical effects of having the illness. We are only beginning to come to terms with the social damage that this pandemic has done.”

He added: “Many people’s futures, many people’s trajectories were altered by the pandemic, very few of them for the better. We’re going to be dealing with this for a long time.”

Methodology

Except where noted, data for this analysis comes from the Gun Violence Archive, a nonprofit that collects information on nearly every fatal shooting in the United States. The archive also collects data on gun suicides and nonfatal shootings, but this data is less complete and The Times excluded these cases from the analysis. (For more information on the archive’s data, see its methodology page.)

The location of every fatal shooting episode was plotted on a map, and then analyzed using Census Bureau data to determine the spread of violence and the racial disparity in shootings.

Location and incident information were the best available from the archive as of Jan. 11, 2024. Cases for which a precise location could not be determined are not shown on the maps but are included in summary statistics.

For the maps of shootings, every census block is color-coded by the number of shooting episodes within a quarter mile of the center of that block or within the boundaries of the block. In summary statistics, fatal shooting counts may not match other published totals because they are based on cases within city boundaries, which may differ from local police jurisdictions, and on the number of cases, not the number of victims.

Exposure is measured by the share of the population living in blocks where there was at least one fatal shooting within a quarter mile during the pandemic years. Population figures are based on the 2020 census.

The Asian category includes people who are Native Hawaiian and other Pacific Islanders, and the Native American category includes Alaska Natives.

Change-over-time figures compare the pandemic years, 2020 through 2023, with the four preceding years, 2016 through 2019.

• Share full article

Advertisement

Numbers, Facts and Trends Shaping Your World

Read our research on:

Full Topic List

Regions & Countries

• Publications
• Our Methods
• Short Reads
• Tools & Resources

Read Our Research On:

What the data says about gun deaths in the U.S.

More Americans died of gun-related injuries in 2021 than in any other year on record, according to the latest available statistics from the Centers for Disease Control and Prevention (CDC). That included record numbers of both gun murders and gun suicides. Despite the increase in such fatalities, the rate of gun deaths – a statistic that accounts for the nation’s growing population – remained below the levels of earlier decades.

Here’s a closer look at gun deaths in the United States, based on a Pew Research Center analysis of data from the CDC, the FBI and other sources. You can also read key public opinion findings about U.S. gun violence and gun policy .

This Pew Research Center analysis examines the changing number and rate of gun deaths in the United States. It is based primarily on data from the Centers for Disease Control and Prevention (CDC) and the Federal Bureau of Investigation (FBI). The CDC’s statistics are based on information contained in official death certificates, while the FBI’s figures are based on information voluntarily submitted by thousands of police departments around the country.

For the number and rate of gun deaths over time, we relied on mortality statistics in the CDC’s WONDER database covering four distinct time periods:  1968 to 1978 ,  1979 to 1998 ,  1999 to 2020 , and 2021 . While these statistics are mostly comparable for the full 1968-2021 period, gun murders and suicides between 1968 and 1978 are classified by the CDC as involving firearms  and  explosives; those between 1979 and 2021 are classified as involving firearms only. Similarly, gun deaths involving law enforcement between 1968 and 1978 exclude those caused by “operations of war”; those between 1979 and 2021 include that category, which refers to gun deaths among military personnel or civilians  due to war or civil insurrection in the U.S . All CDC gun death estimates in this analysis are adjusted to account for age differences over time and across states.

The FBI’s statistics about the types of firearms used in gun murders in 2020 come from the bureau’s  Crime Data Explorer website . Specifically, they are drawn from the expanded homicide tables of the agency’s  2020 Crime in the United States report . The FBI’s statistics include murders and non-negligent manslaughters involving firearms.

How many people die from gun-related injuries in the U.S. each year?

In 2021, the most recent year for which complete data is available, 48,830 people died from gun-related injuries in the U.S., according to the CDC. That figure includes gun murders and gun suicides, along with three less common types of gun-related deaths tracked by the CDC: those that were accidental, those that involved law enforcement and those whose circumstances could not be determined. The total excludes deaths in which gunshot injuries played a contributing, but not principal, role. (CDC fatality statistics are based on information contained in official death certificates, which identify a single cause of death.)

What share of U.S. gun deaths are murders and what share are suicides?

Though they tend to get less public attention than gun-related murders, suicides have long accounted for the majority of U.S. gun deaths . In 2021, 54% of all gun-related deaths in the U.S. were suicides (26,328), while 43% were murders (20,958), according to the CDC. The remaining gun deaths that year were accidental (549), involved law enforcement (537) or had undetermined circumstances (458).

What share of all murders and suicides in the U.S. involve a gun?

About eight-in-ten U.S. murders in 2021 – 20,958 out of 26,031, or 81% – involved a firearm. That marked the highest percentage since at least 1968, the earliest year for which the CDC has online records. More than half of all suicides in 2021 – 26,328 out of 48,183, or 55% – also involved a gun, the highest percentage since 2001.

How has the number of U.S. gun deaths changed over time?

The record 48,830 total gun deaths in 2021 reflect a 23% increase since 2019, before the onset of the coronavirus pandemic .

Gun murders, in particular, have climbed sharply during the pandemic, increasing 45% between 2019 and 2021, while the number of gun suicides rose 10% during that span.

The overall increase in U.S. gun deaths since the beginning of the pandemic includes an especially stark rise in such fatalities among children and teens under the age of 18. Gun deaths among children and teens rose 50% in just two years , from 1,732 in 2019 to 2,590 in 2021.

How has the rate of U.S. gun deaths changed over time?

While 2021 saw the highest total number of gun deaths in the U.S., this statistic does not take into account the nation’s growing population. On a per capita basis, there were 14.6 gun deaths per 100,000 people in 2021 – the highest rate since the early 1990s, but still well below the peak of 16.3 gun deaths per 100,000 people in 1974.

The gun murder rate in the U.S. remains below its peak level despite rising sharply during the pandemic. There were 6.7 gun murders per 100,000 people in 2021, below the 7.2 recorded in 1974.

The gun suicide rate, on the other hand, is now on par with its historical peak. There were 7.5 gun suicides per 100,000 people in 2021, statistically similar to the 7.7 measured in 1977. (One caveat when considering the 1970s figures: In the CDC’s database, gun murders and gun suicides between 1968 and 1978 are classified as those caused by firearms and explosives. In subsequent years, they are classified as deaths involving firearms only.)

Which states have the highest and lowest gun death rates in the U.S.?

The rate of gun fatalities varies widely from state to state. In 2021, the states with the highest total rates of gun-related deaths – counting murders, suicides and all other categories tracked by the CDC – included Mississippi (33.9 per 100,000 people), Louisiana (29.1), New Mexico (27.8), Alabama (26.4) and Wyoming (26.1). The states with the lowest total rates included Massachusetts (3.4), Hawaii (4.8), New Jersey (5.2), New York (5.4) and Rhode Island (5.6).

The results are somewhat different when looking at gun murder and gun suicide rates separately. The places with the highest gun murder rates in 2021 included the District of Columbia (22.3 per 100,000 people), Mississippi (21.2), Louisiana (18.4), Alabama (13.9) and New Mexico (11.7). Those with the lowest gun murder rates included Massachusetts (1.5), Idaho (1.5), Hawaii (1.6), Utah (2.1) and Iowa (2.2). Rate estimates are not available for Maine, New Hampshire, Vermont or Wyoming.

The states with the highest gun suicide rates in 2021 included Wyoming (22.8 per 100,000 people), Montana (21.1), Alaska (19.9), New Mexico (13.9) and Oklahoma (13.7). The states with the lowest gun suicide rates were Massachusetts (1.7), New Jersey (1.9), New York (2.0), Hawaii (2.8) and Connecticut (2.9). Rate estimates are not available for the District of Columbia.

How does the gun death rate in the U.S. compare with other countries?

The gun death rate in the U.S. is much higher than in most other nations, particularly developed nations. But it is still far below the rates in several Latin American countries, according to a 2018 study of 195 countries and territories by researchers at the Institute for Health Metrics and Evaluation at the University of Washington.

The U.S. gun death rate was 10.6 per 100,000 people in 2016, the most recent year in the study, which used a somewhat different methodology from the CDC. That was far higher than in countries such as Canada (2.1 per 100,000) and Australia (1.0), as well as European nations such as France (2.7), Germany (0.9) and Spain (0.6). But the rate in the U.S. was much lower than in El Salvador (39.2 per 100,000 people), Venezuela (38.7), Guatemala (32.3), Colombia (25.9) and Honduras (22.5), the study found. Overall, the U.S. ranked 20th in its gun fatality rate that year .

How many people are killed in mass shootings in the U.S. every year?

This is a difficult question to answer because there is no single, agreed-upon definition of the term “mass shooting.” Definitions can vary depending on factors including the number of victims and the circumstances of the shooting.

The FBI collects data on “active shooter incidents,” which it defines as “one or more individuals actively engaged in killing or attempting to kill people in a populated area.” Using the FBI’s definition, 103 people – excluding the shooters – died in such incidents in 2021 .

The Gun Violence Archive, an online database of gun violence incidents in the U.S., defines mass shootings as incidents in which four or more people are shot, even if no one was killed (again excluding the shooters). Using this definition, 706 people died in these incidents in 2021 .

Regardless of the definition being used, fatalities in mass shooting incidents in the U.S. account for a small fraction of all gun murders that occur nationwide each year.

How has the number of mass shootings in the U.S. changed over time?

The same definitional issue that makes it challenging to calculate mass shooting fatalities comes into play when trying to determine the frequency of U.S. mass shootings over time. The unpredictability of these incidents also complicates matters: As Rand Corp. noted in a research brief , “Chance variability in the annual number of mass shooting incidents makes it challenging to discern a clear trend, and trend estimates will be sensitive to outliers and to the time frame chosen for analysis.”

The FBI found an increase in active shooter incidents between 2000 and 2021. There were three such incidents in 2000. By 2021, that figure had increased to 61.

Which types of firearms are most commonly used in gun murders in the U.S.?

In 2020, the most recent year for which the FBI has published data, handguns were involved in 59% of the 13,620 U.S. gun murders and non-negligent manslaughters for which data is available. Rifles – the category that includes guns sometimes referred to as “assault weapons” – were involved in 3% of firearm murders. Shotguns were involved in 1%. The remainder of gun homicides and non-negligent manslaughters (36%) involved other kinds of firearms or those classified as “type not stated.”

It’s important to note that the FBI’s statistics do not capture the details on all gun murders in the U.S. each year. The FBI’s data is based on information voluntarily submitted by police departments around the country, and not all agencies participate or provide complete information each year.

Note: This is an update of a post originally published on Aug. 16, 2019.

• Partisanship & Issues
• Political Issues
• Politics & Policy

John Gramlich is an associate director at Pew Research Center .

About 1 in 4 U.S. teachers say their school went into a gun-related lockdown in the last school year

Striking findings from 2023, key facts about americans and guns, for most u.s. gun owners, protection is the main reason they own a gun, gun violence widely viewed as a major – and growing – national problem, most popular.

1615 L St. NW, Suite 800 Washington, DC 20036 USA (+1) 202-419-4300 | Main (+1) 202-857-8562 | Fax (+1) 202-419-4372 |  Media Inquiries

Research Topics

• Email Newsletters

ABOUT PEW RESEARCH CENTER  Pew Research Center is a nonpartisan fact tank that informs the public about the issues, attitudes and trends shaping the world. It conducts public opinion polling, demographic research, media content analysis and other empirical social science research. Pew Research Center does not take policy positions. It is a subsidiary of  The Pew Charitable Trusts .

© 2024 Pew Research Center

An official website of the United States government, Department of Justice.

Here's how you know

Official websites use .gov A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS A lock ( Lock A locked padlock ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

National Crime Victimization Survey

The  National Crime Victimization Survey  ( NCVS) is the nation's primary source of information on criminal victimization. Each year, data are obtained from a nationally representative sample of about 240,000 persons in about 150,000 households. Persons are interviewed on the frequency, characteristics, and consequences of criminal victimization in the United States.

The NCVS collects information on nonfatal personal crimes (i.e., rape or sexual assault, robbery, aggravated and simple assault, and personal larceny) and household property crimes (i.e., burglary/trespassing, motor vehicle theft, and other types of theft) both reported and not reported to the police. Survey respondents provide information about themselves (e.g., age, sex, race and Hispanic origin, marital status, education level, and income) and whether they experienced a victimization.

For each victimization incident, the NCVS collects information about:

• the offender, including age, race and Hispanic origin, sex, and victim-offender relationship
• characteristics of the crime, including time and place of occurrence, if weapons were used, the nature of any injury sustained by a victim, and economic consequences to the victim related to their victimization
• whether the crime was reported to police
• reasons the crime was or was not reported
• and victim experiences with the criminal justice system.

The survey has been ongoing since 1973. For more information, see these  NCVS  data collection pages. 

Criminal Victimization Publication Series

These annual publications, which began in 1973, provide official estimates of criminal victimizations reported and not reported to police, based on data from the NCVS.

• Criminal Victimization  publication series

Subnational Program 

While the NCVS was originally designed to provide national-level estimates of criminal victimization, BJS has developed multiple strategies for producing subnational victimization estimates. Through the NCVS Subnational Program , BJS has examined each of these approaches, including the relative benefits, limitations, and application.

Supplement Surveys

NCVS supplement surveys allow BJS to capture the emerging landscape of crime while preserving consistency with the main NCVS. Interviews using supplement surveys are typically in the field for 6 months, either January-June or July-December. A supplement interview is conducted immediately after the NCVS interview for the sample person and before proceeding to the next eligible household member’s NCVS interview. In most cases, the reference period is 12 months, and estimates generated by supplement data are prevalence-based.

National Crime Victimization Survey (NCVS) Dashboard (N-DASH) Tool 

The National Crime Victimization Survey (NCVS) Dashboard (N-DASH) Tool  modernizes public access to NCVS data with new, interactive online data visualizations. The N-DASH tool enhances the core functionality of the NCVS Victimization Analysis Tool (NVAT), increases the speed of conducting analyses, contains new data elements, and enables custom graphics and other modern features. The dashboard also provides direct and user-friendly access to the nation’s primary source of data on criminal victimization, beginning with 1993. This tool replaced the NVAT in early 2022.

NCVS Instrument Redesign  New

The NCVS instrument redesign is a BJS-initiated multiyear effort to improve the efficiency, reliability, and utility of the entire NCVS, which includes a household roster, victimization screener, and detailed crime incident report.

NCVS Sample Design and Redesign New

The Bureau of Justice Statistics’ (NCVS) collects information from a sample of U.S. households that represents the nation. The NCVS sample is a two-stage stratified sample of housing units and group quarters. Every ten years, the  NCVS undergoes a sample redesign  to ensure that the NCVS sample reflects the population distributions identified through the most recent decennial census. The next NCVS sample redesign to reflect changes in the U.S. population is scheduled for 2026, based on the 2020 decennial census.

Events and Trainings

• 50th Anniversary Celebration of the National Crime Victimization Survey
• Update on the NCVS Instrument Redesign: Juvenile Testing Efforts
• University of Maryland NCVS Research Forum Session 3: NCVS User Workshop

Learn More about the NCVS (Adult Participants, English)

Review the YouTube Terms of Service and the Google Privacy Policy

Learn More about the NCVS (Youth Participants, English)

Learn More about the NCVS (Adult Participants, Spanish)

Learn More about the NCVS (Youth Participants, Spanish)

View related videos

How many persons are victims of crime in a given year?

The number and percentage of unique persons who are victims of violent crime (prevalence of violent crime) and the number and percentage of unique households that are victims of property crime (prevalence of property crime) can be found using the National Crime Victimization Survey (NCVS) . See the Criminal Victimization series for prevalence rates by year. See A New Measure of Prevalence for the National Crime Victimization Survey for more information on the measurement of prevalence using the NCVS.

What is a victimization?

A victimization is a single victim or household that experiences a criminal incident. Criminal incidents or crimes are distinguished from victimizations in that one criminal incident may have multiple victims or victimizations. For violent crimes (rape/sexual assault, robbery, aggravated assault, and simple assault) and for personal theft/larceny, the count of victimizations is the number of individuals who experienced a violent crime. For crimes against households (burglary, trespassing, other theft, and motor vehicle theft), each household affected by a crime is counted as a single victimization.

What is the National Crime Victimization Survey?

The Bureau of Justice Statistics' (BJS) National Crime Victimization Survey (NCVS)  is the nation's primary source of information on criminal victimization. Each year, data are obtained from a nationally representative sample of about 240,000 persons in about 150,000 households on the frequency, characteristics, and consequences of criminal victimization in the United States. The NCVS collects information on nonfatal personal crimes (rape or sexual assault, robbery, aggravated and simple assault, and personal larceny) and household property crimes (burglary, motor vehicle theft, and other theft) both reported and not reported to police. Survey respondents provide information about themselves (e.g., age, sex, race and Hispanic origin, marital status, education level, and income) and whether they experienced a victimization. For each victimization incident, the NCVS collects information about the offender (e.g., age, race and Hispanic origin, sex, and victim-offender relationship), characteristics of the crime (including time and place of occurrence, use of weapons, nature of injury, and economic consequences), whether the crime was reported to police, reasons the crime was or was not reported, and victim experiences with the criminal justice system. 

Data Collections

• National Crime Victimization Survey (NCVS)
• Police-Public Contact Survey (PPCS)
• City-Level Survey of Crime Victimization and Citizen Attitudes

Recent Publications

• Crimes Involving Juveniles, 1993–2022
• Data Breach Notifications and Identity Theft, 2021
• The National Crime Victimization Survey and National Incident-Based Reporting System: A complementary picture of crime in 2022
• View related awards

Related Links

• Find sites with information related to: National Crime Victimization Survey (NCVS)

Alcohol and guns are everywhere in American society. Researchers say it's a deadly combo.

A fight broke out in a bar and spilled out onto the street early one Saturday morning last month in Melbourne, Florida, leaving three people injured, windows shattered and a bullet lodged in the child car seat of a nearby vehicle.

"This is unacceptable," Melbourne Mayor Paul Alfrey said after the shooting. The incident spurred city leaders to consider requiring bars and restaurants to obtain extended-hours permits that emphasize security measures in order to serve alcohol.

The shooting in Melbourne is just one of the many incidents of gun violence happening in the U.S. nearly every day — many with a connection to alcohol.

"Alcohol misuse is a risk factor for gun violence," said Silvia Villarreal, director of research translation at the Johns Hopkins Center for Gun Violence Solutions and lead author of a new report on the intersecting public health issues.

The report by the Consortium for Risk-Based Firearm Policy, a group of more than 30 gun violence experts, and the Center for Gun Violence Solutions recommends policymakers take action to limit access to firearms by people with a documented history of alcohol misuse and restrict firearms at locations where alcohol is consumed.

"One huge part of why we did this is there’s really not a lot of awareness of the data and how many people are affected by this," Villarreal said in a panel discussion on the report Thursday.

What to know about alcohol and gun violence in the U.S.

Alcohol kills 140,000 people annually, and guns kill more than 48,000, according to data from the U.S. Centers for Disease Control and Prevention.

The two often intersect: An estimated 1 in 3 gun homicide perpetrators drank heavily before murdering their victims, 30% of gun homicide victims drank heavily before being killed, and a quarter of gun suicide victims were heavily drinking before they died by suicide, researchers found .

Among legal gun owners, alcohol misuse (as measured through DUI and other alcohol-related convictions) increases the risk of interpersonal gun violence, including intimate partner violence, a series of studies from the Violence Prevention Research Program at the University of California, Davis, found.

The COVID-19 pandemic only exacerbated these issues. Excessive drinking increased by 21%, and alcohol-related deaths increased approximately 25%, one study found . Meanwhile, gun sales increased by 40% and gun homicides by 35%, and gun suicides had the largest one-year increase ever recorded, two other studies concluded .

Do U.S. laws regulate the risk of alcohol and guns?

Bringing a gun to a place where alcohol is present is legal in many states but prohibited in others. Some states exempt concealed carry permit holders from laws prohibiting firearm possession in places where alcohol is consumed, the report said.

Most states do not have laws prohibiting people who misuse alcohol from purchasing or possessing firearms, the report said. But even states that do often fail to provide clear definitions of who is disqualified, which makes the policies difficult to enforce and hard to evaluate, the researchers said.

Michelle Spencer, deputy director of equity and community partnerships at the Johns Hopkins Center for Gun Violence Solutions, said alcohol has been socialized into our everyday lives and is an "under-appreciated" risk factor for gun violence. "It's not seen necessarily as a trigger as other substances are viewed," Spencer said Thursday.

What researchers recommend on alcohol and guns

Drawing on available data, the report authors outlined a series of recommendations. Chief among them, the researchers say states should pass laws prohibiting the purchase and possession of firearms by people convicted of two or more instances within a five-year period of driving under the influence of alcohol or driving while intoxicated.

Female firearm homicide is 19% lower in the five states with DUI penalties that activate federal firearm prohibitions after one or two DUI convictions, compared to states without such laws, a study last year found.

The authors recommend states adopt concealed carry laws to temporarily prohibit anyone with a court record of alcohol misuse within the past five years from receiving a concealed carry license. They also say states and cities should prohibit the public carry or possession of firearms in locations where alcohol is consumed and prohibit people from carrying or possessing firearms while intoxicated.

Other recommendations focus on increased education, encourgaing courts to consider evidence of alcohol misuse when making decisions about firearm prohibitions and ensuring alcohol offenses that are firearm prohibitory under state or federal law are entered into the National Instant Criminal Background Check System in a timely manner.

Dig deeper on gun violence

Memorial Day Weekend: Gun violence rages with at least 20 mass shootings recorded

Poll: Is stopping gun violence more important than gun rights? Most Americans say yes.

Do gun buyback programs work? Thousands of firearms surrendered in New York in one day.

Human Trafficking Statistics and Facts In 2024

“Human Trafficking is the recruitment, transportation, transfer, harboring or receipt of people through force, fraud or deception, with the aim of exploiting them for profit. Men, women and children of all ages and from all backgrounds can become victims of this crime, which occurs in every region of the world. The traffickers often use violence or fraudulent employment agencies and fake promises of education and job opportunities to trick and coerce their victims.” ( United Nations ) 

An important term to know is child sexual abuse material (CSAM), which is any image or video of sexually explicit content involving a minor (someone under 18 years of age). 

Commonplace of Human Trafficking Worldwide

From high profile cities in the United States to remote locations around the world, human trafficking is a relatively silent epidemic that impacts communities everywhere. For this reason, Operation Underground Railroad (O.U.R.) operates internationally in seven regions .

Primary Causes of Human Trafficking

Traffickers, just like other types of predators, usually target vulnerable populations. According to the U.S. Department of Homeland Security , victims are likely to suffer from: 

• Psychological or emotional vulnerability
• Economic hardship
• Lack of a social safety net
• Natural disasters
• Political instability

Common Types of Human Trafficking

• Forced Labor – victims are forced to work against their will
• Sex – individuals are required to engage in sexual acts
• Forced Marriage – victims are forced to marry another person without giving consent
• Domestic Servitude – forced labor that occurs in a private household
• Child Soldiers – minors are used as fighters in acts of war

Statistics of Human Trafficking Worldwide

Human trafficking statistics reveal a very sad reality for many individuals. Behind every stat is a person – someone’s mother, father, brother or sister. These stats provide insight into the severity of the issue. 

• Today, there are 49.6 million people in modern slavery worldwide, and 12 million of them are children. ( ILO , United Nations )
• 54% of those trapped in modern slavery are women and girls. (ILO)
• Sex trafficking is the most common type of trafficking in the U.S. ( Polaris ) 
• There were 88 million child sexual abuse material (CSAM) files reported to the National Center for Missing and Exploited Children ( NCMEC ) tip line in 2022. 
• Child sex trafficking has been reported in all 50 U.S. states . (NCMEC) 
• Human trafficking is a $150 billion industry. ( UNICEF ) 
• Human trafficking is the second most profitable illegal industry in the U.S. ( UNICEF ) 

Role of Technology in Combating Human Trafficking

The growth of technology has provided numerous tools to human traffickers, increasing the challenges faced by law enforcement and advocate organizations. One such tool is social media. These platforms are often used by traffickers to recruit victims because they provide direct and relatively easy access. 

O.U.R. helps by offering resources to expand and amplify the anti-human trafficking and exploitation efforts of law enforcement.

Human Trafficking Statistics Focusing on Children

Multiple organizations estimate that 500,000 predators are online every day, leaving minors vulnerable each time they access a social media account. 

The following statistics are courtesy of ParentsTogether , a nonprofit organization providing independent reporting and commentary on issues that affect kids and families. 

• 1 in 3 children are first exposed to social media at age 5 or younger. 
• 1 in 3 children are expected to have an unwelcome sexual experience online before they turn 18. 
• Younger social media exposure correlates with more sexual harm online and peaks for kids who start using social media at 11-12 – the age around which most American children get their first smartphone. 
• 43% of kids exposed to inappropriate sexual content online were under 13. 
• Kids with disabilities, special needs, or who identify as LGBTQ+ are 2-4x more likely to send explicit images of themselves than their peers. 

The most chilling fact about these statistics is that they only reflect the reported numbers. Human trafficking lives in the shadows, meaning it is impossible to ever know how many cases are happening without being reported.  

Protect your children from predators online with O.U.R.’s Start Talking: A Guide to Keep Children Safe Online . 

Trafficking Statistics Specific to the United States

With an estimated population of 335 million people (U.S. Census Bureau), the United States plays an important role in the fight against human trafficking. The Trafficking Victims Protection Act of 2000 “equipped the U.S. Government with new tools and resources to mount a comprehensive and coordinated campaign to eliminate modern forms of slavery domestically and internationally,” according to the U.S. Department of Justice. It also established three pillars in the fight: protection, prevention, and prosecution. 

Additionally, the United States’ Office to Monitor and Combat Trafficking in Persons publishes a Trafficking in Persons (TIP) Report annually. The U.S. Department of State describes it as the “U.S. Government’s principal diplomatic tool to engage foreign governments on human trafficking. It is also the world’s most comprehensive resource of governmental anti-trafficking efforts and reflects the U.S. Government’s commitment to global leadership on this key human rights and law enforcement issue.” 

Agencies like the Department of Homeland Security work tirelessly to intercept traffickers and ensure justice for victims domestically and abroad. 

Global Initiatives Against Human Trafficking

On a global scale, the United Nations and International Labour Organization provide guidance in the fight against human trafficking. Both bring awareness to the issue and help provide standards and protocols that are recognized by most countries around the world. 

The international community recognizes the importance of collaboration to end all forms of trafficking. Why? Because the numbers show an urgent need for those around the world to unite against predators. The issue is too large to ignore – too large to leave to any one government or entity.  

So, What Can You Do?

Awareness is the first line of defense against any crime, including human trafficking. Once someone is made aware of the issue, they are better equipped to address it. 

Additional steps to take: 

• Education – learn the signs of human trafficking
• Take Action – join the fight with O.U.R. 
• Contact Authorities – report potential cases of trafficking  

About OUR Rescue

We lead the fight against child sexual exploitation and human trafficking worldwide. 

Our work spans the globe as we assist law enforcement in rescue efforts and help provide aftercare to all those affected. While we prioritize children, we work to empower the liberation of anyone suffering at the hands of those looking to sexually exploit. We offer vital resources to authorities around the world and work tirelessly to raise awareness and meet survivors on their healing journey. Our resolve never falters, and we will boldly persevere until those in need are safe. 

Get emails from OUR Rescue

Subscribe to stay up to date on rescues, operations, and exclusive content.

By clicking Subscribe, you agree to recieve emails from OUR Rescue. You can opt out at any time.