Journalists and Researchers: How FBI Crime Data Can Bolster Your Work

Read this first

The FBI’s crime data collection system contains information on millions of crimes, victims and offenders, and covers more than 15,000 U.S. law enforcement agencies. The program, called the National Incident-Based Reporting System (NIBRS), includes invaluable information that can power your journalism or research.

For example, you can use NIBRS data to compare crime rates across cities, see how often crimes occur between family members, examine the race and sex of victims, or hold a local agency accountable for incorrectly reporting data to the federal government. For more story ideas, check out our full list of example story topics.

Even so, NIBRS is massive and can be difficult to interpret. The data has been known to contain errors, and information is frequently missing. Plus, the meaning of the data changes over time, as definitions are rewritten and the FBI updates its procedures.

This toolkit will explain what the FBI’s crime data collection program is and how it compares to other national sources of crime statistics. It will also provide tips for working with the data and some ideas about the types of stories it can help you tell. As you read this toolkit, you can follow along with the FBI’s Crime Data Explorer (CDE), which incorporates data from NIBRS. You can learn more about the CDE in our Accessing NIBRS section.

To learn more about responsibly analyzing and writing about police crime data coming from any source, check out our companion toolkit. If you’re working on a news story or research project about crime data and would like to talk to The Marshall Project, you can request a consultation.

Join us on Tuesday, Oct. 20, 2026, at 1 p.m. ET for our webinar about how to report on crime data this election season. Learn more and register here.

NIBRS: The FBI’s national crime data resource

Understanding NIBRS

NIBRS has been the FBI’s official system for collecting crime data from U.S. law enforcement agencies since 2021. Local law enforcement agencies voluntarily submit crime data into this system, either directly or through state agencies that process the information first. As of 2026, around 90% of the U.S. population lives in a police jurisdiction that reports crime data to NIBRS.

The data collected in NIBRS contains specific details about each crime “incident,” such as the number of victims and offenders, along with their demographic characteristics. Other details include the dates and times of the crimes and the types of offenses that police believe took place. This allows for rich analysis and is why NIBRS data is so voluminous, containing millions upon millions of records for one year of data.

Each summer or fall, the FBI releases a comprehensive dataset covering the prior calendar year. The FBI also began publishing monthly crime data in 2025, but it is being adopted slowly by law enforcement agencies, many of which are reporting no or partial monthly data, so it should be used with caution.

NIBRS data structure

Each incident in NIBRS represents a full event and can involve one or more crimes, also known as offenses. For example, a robbery with a gun could involve the crimes of bank robbery and aggravated assault. They can be logged in NIBRS as separate offenses, but can also be tracked as part of the same incident.

Each offense is categorized into Group A or Group B, which contain different levels of detail. The Group A offenses include what officials tend to consider more serious crimes, including murder, aggravated assault, motor vehicle theft, drug offenses and sex offenses. Group B offenses include crimes such as DUIs, disorderly conduct and trespassing. You can see a list of all NIBRS offenses here.

Most offenses fall into Group A, and police can submit detailed information available about them. This information includes:

• Type of offense, such as robbery, aggravated assault or identity theft
• Offense date
• Characteristics of victims, offenders and arrestees, including age, sex, race and ethnicity
• Details of the incident, including location, weapons, property stolen or seized, and victim injuries
• Residential status, specifically if the victim or arrestee resides in the district where the crime occurred
• Relationship between victim and offender, such as if they are married, strangers or have another type of relation

Group B offenses are only submitted to NIBRS if an arrest takes place. The information available about these offenses is limited to details about the arrest and arrestee. For example, if someone is suspected but not arrested for distributing obscene material — a Group B offense — that incident will not show up in NIBRS. Meanwhile, if someone is suspected but not arrested for illegal sports betting — a Group A offense — the incident will appear in NIBRS.

Available details about Group B offenses include:

• Type of offense
• Arrest date
• Characteristics of the arrestee, including age, sex, race, and ethnicity
• Weapons used in the crime
• Residential status, specifically if the arrestee resides in the district where the crime occurred

The underlying NIBRS data is organized into “segments,” which are interconnected tables containing different information about a given crime incident. For example, one segment includes information about offenders, another about victims and a third about the offenses themselves. Arrests are logged in a separate segment.

You can learn more about the segments in Jacob Kaplan’s handbook, Decoding FBI Crime Data, which is free to read online. Part III of the book breaks down each NIBRS segment in detail. For a visual representation of the data segments, check out page 5 of this FBI guide.

If you’re new to data journalism, the FBI and researchers have made NIBRS data more accessible by combining the data segments into formats that are easier to use. Read more about it in the Accessing NIBRS section below.

The rocky transition to NIBRS

Before NIBRS became the FBI’s standard system in 2021, agencies often submitted crime data through the old Summary Reporting System (SRS). While a few agencies may still submit information this way instead of through NIBRS, it has been almost entirely phased out.

Under SRS, law enforcement agencies submitted counts for how many crimes occurred in a given year, and they were told to apply the “hierarchy rule” to collect information on only the most serious crime in an incident. For example, if an incident included an aggravated assault and a murder, law enforcement was instructed to count the incident only as a murder in the data. This is no longer the case under NIBRS, where details of each offense are available.

NIBRS also allows agencies to differentiate between crimes cleared by arrest and crimes cleared by “exceptional means,” which include instances when an offender has died, a victim refused to cooperate or the prosecution decided not to pursue the case. Under SRS, all clearances were reported collectively as a single number.

NIBRS was first introduced as a data collection system in the 1980s, and some agencies have been using it for decades. However, as The Marshall Project has reported, many police departments were not prepared for the official transition in 2021 and failed to submit accurate or complete data that year. Researchers generally consider 2021 FBI crime data to be unreliable, and any apparent reductions in crime that year are likely due to reporting issues rather than real trends. Starting in 2022, the data quality improved dramatically.

The New York and Los Angeles police departments are among the agencies that were late to transition to NIBRS.

The FBI and researchers have ways to aggregate NIBRS data to match what was submitted historically under SRS. This work has allowed crime data from before and after the transition to be accurately compared to each other. The FBI Crime Data Explorer provides long-term national, state and local crime trends using data submitted through both systems.

Both NIBRS and SRS are part of the FBI’s Uniform Crime Reporting (UCR) program, which is a longstanding FBI initiative to collect, aggregate and report crime statistics from local law enforcement agencies. While researchers, data analysts and officials sometimes use the terms UCR and SRS interchangeably, they are not the same.

Demographic data

The FBI collects data on the race and ethnicity of offenders, victims and arrestees. However, this data often reflects an officer’s perception rather than how the individual identifies, leading to inaccuracies. Ethnicity information is often missing from NIBRS entirely — the most common ethnicity listed is “unknown.”

Agencies are supposed to enter the sex of victims, offenders and arrestees as either male or female. There is no identifier for individuals who are transgender or nonbinary.

For a victim’s age, police can input a range if they do not know the exact number (e.g., age 20-29), and the FBI will produce the average. There isn’t a simple way to tell if the age was originally input as a range. Keep in mind that police may use 1 or 99 if the age is unknown or use multiples of five when making estimates (e.g., age 20 or 25).

What’s not captured in NIBRS

Police do not submit victim or suspect names, incident addresses or narrative descriptions of events into NIBRS. The data also does not capture details about cases after an arrest is made, such as what occurs during prosecution or incarceration, or if the charges were dropped. Federal law enforcement agencies rarely submit records to NIBRS, and civil enforcement actions, such as civil immigration arrests made by ICE, are not tracked in the data.

NIBRS only captures data on crimes that are reported to police. If a crime victim chooses not to share information with law enforcement, it will not appear in NIBRS. Another resource, the National Crime Victimization Survey, collects information directly from crime victims and allows us to estimate how often different crimes are reported. Some offenses, like sexual assaults and hate crimes, have especially low reporting rates. Learn more about reported crimes and general tips for working with police-reported crime data in our companion toolkit.

NIBRS does not allow police to update a reported crime if it is ultimately determined to be unfounded, so you can’t calculate how many submitted crime reports turn out not to be credible.

Since NIBRS establishes a standardized data collection system across all police jurisdictions, it forces agencies to categorize their data to fit the FBI’s rules, which can leave out important information. For example, the criminal offense of harassment doesn’t exist in NIBRS data, which only has 62 crime categories — 52 Group A offenses and 10 Group B offenses. Police must decide whether to classify their harassment cases as simple assaults, stalking, intimidation or some other crime category, which could mean two agencies classify the same incident in different ways.

For some data fields, the FBI limits collection to a smaller set of data points than all that could potentially apply. When reporting an incident in NIBRS, police can list up to 10 associated offenses, but when reporting an arrest, police are only able to log the most serious crime someone is arrested for in connection to that incident, even if the person was arrested on multiple charges. Similarly, police can only report up to the three most serious drugs seized at a crime scene even if more are present, and they are instructed to put the most serious drug first.

Accessing NIBRS

The simple way

The FBI Crime Data Explorer is the easiest way to access the FBI’s crime data, especially for users without a lot of data reporting experience, because it offers charts and drop-down menus on an online dashboard. The explorer can be a valuable resource when looking at a specific agency’s data if it submits information regularly, but it is less reliable when examining statewide or national trends because some agencies don’t provide detailed data on a consistent basis.

In the explorer, you can filter by state, police agency, crime and time frame to view charts showing how many offenses occurred each month. Below the charts, you can see how many of those crimes involved offenders and victims of different age groups, sexes, races and ethnicities, as well as breakdowns for other incident details. The dashboard incorporates data from NIBRS and the FBI’s old reporting system, SRS, to show longer-term trends.

Only aggregate counts are made available in the dashboard, so it may not provide the level of detail you want. The data explorer also doesn’t allow you to filter for multiple case details at once, such as murders where the victim is female and the offender is male.

For more information about what the dashboard can do, click on the dashboard’s “How these numbers are calculated” button and read the “About the data” section below the charts. If you’re hitting limitations with the CDE, try reaching out to one of the experts we’ve listed below to troubleshoot.

The FBI does not provide details on the accuracy or comprehensiveness of the information in the CDE, so apparent changes in crime trends in the charts should be interpreted with caution. Sudden dips or jumps are likely signals of data entry issues rather than actual shifts in trends. This vigilance is especially important when looking at recent months of data, as many agencies do not report complete monthly data on a regular basis.

The FBI regularly releases reports on crime trends, which can be downloaded through the CDE. By using statistical models to estimate how much data is missing, these reports often contain more accurate findings than can be obtained through the CDE’s dashboards.

The FBI sometimes adjusts topline findings from prior years, like the number of murders across the U.S., when they release their annual “Reported Crimes in the Nation” report, so be sure to cite the most recent available document. Take a look at the fine print on these reports to understand the sources they rely on and any caveats in their conclusions.

In addition, many states collect crime data from local law enforcement agencies, standardize it and then submit it to NIBRS on their behalf. You can ask your state’s law enforcement department for this data directly, which may offer more detail than what’s available through the data explorer while also not requiring complex data work.

The tech-savvy approach

For users with data reporting experience, there is a lot more you can explore with NIBRS.

The FBI Crime Data API allows users to pull a limited amount of data at a time from NIBRS. It returns data in JSON or CSV formats.

The FBI bulk data downloads contain summary tables by state and year with crime counts. Larger files may be difficult to download. The “Annual Master Files” contain full years of NIBRS data in ASCII text formats.

You can also download bulk NIBRS data that researchers have preprocessed and made available online.

Working with NIBRS

Changing definitions

The FBI’s definitions of specific crimes or data fields can change in subtle and significant ways over time. It’s important to check if a definition changing over the period of data you’re examining could be responsible for what appears to be changing crime trends.

In 2025, the FBI announced that the crime category of “fondling” would be replaced with “criminal sexual contact,” a broader category that covers more incidents. In 2013, the FBI adopted a broader definition of rape that includes sexual assault with an object and sodomy, which greatly increased the number of rapes reported nationwide. Tracking changes like this will help you understand why you might see numbers rising or falling without more or fewer crimes actually occurring.

The FBI sometimes adds completely new data fields or category options. In 2025, the FBI added a new option for agencies to report “gunshot wound” as an injury type. Previously, shootings were collected in the aggravated assaults category along with other non-gun crimes. While this should allow for a better examination of gun violence using NIBRS, keep in mind that it can take years for police departments to begin using new fields.

At the top of the NIBRS user manual and technical specifications documents, you can see a list of changes the FBI makes to its data collection processes each year.

Police agencies may also use different protocols for categorizing crimes because of ambiguous or overlapping definitions. For instance, sometimes the same crime could be considered an aggravated assault or a simple assault, and some places may appear to have higher rates of aggravated assault because of how officers categorize these cases. Since aggravated assaults help determine violent crime rates, it’s best not to rank states or cities based on these data points.

Understand what’s missing in the data

Though law enforcement agencies may require officers to undertake detailed data entry, officers still may not fill out all the relevant information about a crime, which could ultimately leave the NIBRS crime data incomplete.

Police agencies also will not report information if they don’t track it, which is a particularly acute problem for new data fields. For example, a department may not report animal cruelty incidents — a newer crime category that was not available under the FBI’s old SRS system. Agencies are much more likely to submit comprehensive data on longstanding crime categories such as murder and aggravated assault. You can review this user manual to get a sense of when updates have been made to various crime categories.

Police agencies express missing data in different ways, even when there are specific instructions to mark certain fields as “unknown.” Sometimes officers leave fields blank, while other times they may submit values like “0,” “-99” or “N/A.” In NIBRS, police cannot leave the date field blank — they have to record the date an incident occurred — so they may enter approximate incident dates or mark the first day of the month or year. Looking closely for outliers and unexpected entries will help identify missing or incorrect data.

Here are some examples of what missing data for a police agency could look like in NIBRS:

• Data for partial years or partial months
• Data for some offenses but not other offenses (e.g., reporting aggravated assaults but not simple assaults)
• Missing data for detailed description fields (e.g., types of drugs collected at a crime scene)

Before reporting on any numbers, check what percentage of your data has missing or incomplete values and try to estimate how much the results of your analysis could change if the missing details were filled in. If the missing data would significantly change the findings of an analysis, it’s best not to draw sweeping conclusions.

Depending on what aspect of NIBRS you’re examining, the data might not be comprehensive enough for a national or regional analysis, but it could support a local one. Conversely, a particular agency may not have robust enough data for a local analysis, but looking at trends across an entire region would be appropriate because issues with one agency’s data won’t dramatically impact the findings.

Your local police and NIBRS

Some police agencies don’t make crime data easily accessible, but reporters have had luck requesting data by specifying that they are seeking the same fields the agency submits to NIBRS. The agency is likely already tracking data for NIBRS submissions, so this could streamline your requests.

Transitioning to NIBRS has been a time-consuming and expensive undertaking for many police agencies. Some changed their database management systems, updated their manuals and educated officers on new data entry protocols.

These are some questions to ask your local law enforcement when analyzing their NIBRS data, especially if you find anything unexpected:

• Do you report to NIBRS? If so, do you submit data monthly or annually?
• When did you transition to NIBRS? Is the transition still ongoing?
• Did you change the way any crimes are recorded or defined during the NIBRS transition? If so, when?
• Do you submit all crime categories to NIBRS or only some?
• Do you vet the crime data before submitting it for accuracy and completeness?
• What do you do if there is a correction to data you’ve already submitted to NIBRS?
• Can you provide the data that you submit to NIBRS to me directly on a more regular basis than what you submit for the FBI program?

Sometimes, the data police agencies report on their own website or keep internally differs from what they report to NIBRS. Comparing these datasets could allow you to catch errors or identify situations where police have withheld or manipulated statistics. If the numbers don’t match, ask your police department to explain the discrepancy.

Locally kept data is often more up-to-date than what is available through NIBRS and frequently contains more details, such as addresses and descriptions of incidents. It may make sense to rely on local data for your analysis if you’re interested in a single police jurisdiction. On the other hand, NIBRS may allow you to compare your agency to other jurisdictions, or their data may include crime information that your police agency does not publish.

Tips for interpreting NIBRS data

Know what you’re measuring. Be careful about conflating incidents, offenders, victims and offenses. An incident can contain multiple offenses, as well as multiple offenders and victims. A growth in the number of victims of robbery does not necessarily mean there is a growth in the number of incidents of robbery — it could indicate that there are more victims per incident. Likewise, there is a distinction between a growing number of murders and a growing number of people committing murders, since one person can commit multiple murders.

Beware of recent data. The FBI’s monthly crime data is not generally considered reliable for the most recent three months due to a lack of reporting and the tendency for errors, which may be updated at a later time.

Monthly data can be useful for specific law enforcement agencies that regularly submit it, but should be avoided for state or national analyses. Be careful when relying on this data and speak with experts to discuss your analysis. If you’re looking for recent crime statistics that you can compare across jurisdictions, a good source is the Crime Index, which operates on a two-month lag and produces data that closely mirrors what ultimately appears in NIBRS.

Understand where NIBRS data ends. NIBRS data doesn’t extend beyond the crime incident or arrest. What happens after that — trials, pleas or sentencings — doesn’t get tracked. If someone is arrested for a crime and then the charges are dropped, or they go to trial and are found innocent, the details of their arrest will still show up in NIBRS.

Example story topics

NIBRS data can help you pursue many possible stories. Some questions will be more straightforward to answer, while others will be challenging and require data analysis experience.

Here are some examples of what’s possible:

• When a high-profile murder happens in your community, use NIBRS to calculate the per capita murder rate in your area to see if it’s increasing or decreasing over time. You can also compare your city’s murder rate to its neighbors’ by dividing the number of murders in each city by the city’s population.
• If a local politician has been talking about instituting a youth curfew to deal with the trouble they say is caused by teenagers, use NIBRS to see how many crimes are actually committed by people under 18 in the categories the politician is worried about. What is the most common age bracket for victims and offenders, and has that changed over time?
• If a community group is complaining about cops not taking car break-ins seriously, use NIBRS to look at the clearance rate for auto crimes and compare it to other cities or a state or national average. You can also analyze which types of crime tend to be solved through “exceptional means,” where an arrest is not made.
• If your community is experiencing a rash of stabbings, look for trends in the injuries associated with different types of crimes. (Injury data is often incomplete in NIBRS, but you can mention this limitation in the story.)
• If your local police agency publicizes a large crime scene seizure, examine the most common types of items that police say they’ve taken and the estimated value of the items. (Seizure data is often incomplete in NIBRS, but you can mention this limitation in the story.)

Other national crime data sources

Other federal agencies, research groups and advocacy organizations publish their own databases that can complement what is offered through NIBRS. Sometimes, referencing more than one data source in an article can help highlight different aspects of the topic you’re reporting on or allow you to fact-check whether similar trends are appearing across multiple sources of information.

Learn about other national crime data sources in this tipsheet.

Expert sources

Criminologists and other experts can help you think through the complexities when working with NIBRS and offer broader context on crime trends. These experts have agreed to speak with journalists for articles about NIBRS crime data.

• Jeff Asher, AH Datalytics, co-founder. Asher runs AH Datalytics, a criminal justice data analytics firm, and he is a nationally recognized crime data analyst. He previously served as a crime analyst for the City of New Orleans and publishes a newsletter, Jeff-alytics, that explores the nuances of working with crime data. Email: jasher@ahdatalytics.com
• Jill Castellano, The Marshall Project, data reporter. Castellano reports on the uses and misuses of criminal justice data in American society. Her stories cover crime statistics, policing, the perception of crime and the broader criminal justice system. She is the author of this toolkit, along with our toolkit on how to report on local crime data. Prior to journalism, Castellano studied criminology and worked in a crime research lab. Email: jcastellano@themarshallproject.org
• Thaddeus Johnson, Council on Criminal Justice (CCJ), senior fellow. Johnson is also an associate professor of criminal justice and criminology at Georgia State University, a former law enforcement official in Tennessee, and an expert on policing, crime control, and governance equity. His research focuses on police lethality and coercion, street violence, recidivism, algorithmic and predictive bias, and correctional control. Email: tjohnson230@gsu.edu
• Jacob Kang-Brown, Prison Policy Initiative, senior researcher. Kang-Brown is an expert on crime and criminalization, as well as the author of Understanding National Crime Data, published by the Vera Institute of Justice. His work has appeared in numerous publications, including The Lancet Public Health, The Atlantic, USA TODAY and American Jail Magazine. Contact form: https://www.prisonpolicy.org/contact.html
• Ernesto Lopez, Council on Criminal Justice (CCJ), senior research specialist. Lopez has authored more than a dozen reports for CCJ on crime trends in large U.S. cities and served as research staff for CCJ’s crime trends working group. He presents to policy groups regularly, including the White House Domestic Policy Council and the National Governors Association. Email: elopez@counciloncj.org
• Alex Piquero, University of Miami, Department of Sociology & Criminology, Arts & Sciences distinguished professor. Piquero previously served as former director of the Bureau of Justice Statistics, which collects and analyzes crime data as part of the U.S. Department of Justice. He has provided congressional testimony on evidence-based crime prevention practices and offered counsel to criminal justice agencies and elected leaders. Email: axp1954@miami.edu

News stories about FBI crime data

We’ve identified a handful of strong news stories as examples to help you understand and write about FBI crime data, and to potentially emulate in your own community.