FBI Crime Statistics Explained: How to Read UCR Data

What Is the UCR Program?

The Uniform Crime Reporting (UCR) program is the United States' oldest and most widely cited source of national crime statistics. Launched in 1930 and managed by the FBI, the program collects data from law enforcement agencies across the country covering offenses reported to police, arrests, and law enforcement staffing.

Participation is voluntary, but the program covers a substantial portion of the country. In recent reporting years, agencies participating in the UCR program represented approximately 98% of the U.S. population. However, "participation" does not guarantee complete data. Some agencies report for only part of the year or submit data for some categories but not others.

The FBI publishes UCR data through its annual Crime in the United States reports and through the Crime Data Explorer (CDE), an interactive tool for browsing the data. PlainCrime draws from these published tables to present crime data for every state, county, and city that reports.

A crime rate is only as meaningful as the population it is divided by and the agencies that chose to report it.

Part I vs Part II Offenses

The UCR divides crimes into two broad categories. Understanding this distinction is critical because most published crime statistics and "crime rate" rankings focus exclusively on Part I offenses.

Part I Offenses (Index Crimes)

Part I offenses, also called "index crimes," are the eight most serious offenses tracked by the UCR. They are divided into violent and property categories:

Part I data is collected based on reported incidents, meaning if someone reports a burglary to police, it counts in the statistics regardless of whether an arrest is made. This gives Part I data a broader picture of crime occurrence.

Part II Offenses

Part II covers 21 additional offense categories, including fraud, embezzlement, vandalism, weapons violations, drug abuse violations, driving under the influence, and disorderly conduct. Unlike Part I, these are tracked only through arrest data, not incident reports. This means Part II numbers significantly undercount actual offenses since many crimes in these categories never result in an arrest.

How Crime Rates Are Calculated

Raw crime counts are not useful for comparing places of different sizes. A city of 1 million people will naturally have more crimes than a town of 10,000. To enable fair comparison, crime statistics are expressed as rates per 100,000 population.

For example, if City A (population 250,000) reports 1,000 violent crimes, its violent crime rate is (1,000 / 250,000) × 100,000 = 400 per 100,000. If City B (population 50,000) reports 300 violent crimes, its rate is (300 / 50,000) × 100,000 = 600 per 100,000. Despite having fewer total crimes, City B has a higher rate.

You can explore crime rates for specific places on PlainCrime. Browse by state or search for a city to see violent and property crime rates side by side.

The Hierarchy Rule

Under the traditional UCR Summary Reporting System (SRS), agencies apply the "hierarchy rule." When multiple offenses occur during a single criminal incident, only the most serious offense is counted. For example, if a robbery also involves an assault, only the robbery is recorded in UCR data.

The exception is arson, which is always counted regardless of other offenses in the same incident. This hierarchy rule means the traditional UCR data systematically undercounts less serious offenses when they co-occur with more serious ones.

The NIBRS Transition

The National Incident-Based Reporting System (NIBRS) is a more detailed alternative to the traditional Summary Reporting System. Instead of submitting aggregate monthly counts, agencies using NIBRS report detailed information on each criminal incident: the specific circumstances, victim-offender relationships, weapon use, location type, and more.

NIBRS eliminates the hierarchy rule, capturing all offenses within an incident. It also tracks 52 offense categories (compared to the SRS's 29), providing a much richer picture of crime. The FBI mandated that all agencies transition to NIBRS by January 2021, but adoption has been uneven:

• By 2022, approximately 73% of agencies covering about 77% of the population were reporting via NIBRS
• Major holdouts initially included New York City and Los Angeles, though both have since begun NIBRS reporting
• The transition caused temporary data gaps in 2021-2022, when agencies that had stopped SRS reporting were not yet NIBRS-certified

Data on PlainCrime reflects what the FBI publishes in its annual tables. For years affected by the transition, some agencies may be missing from the data entirely. We note this where applicable on individual state and city pages.

How to Compare Cities Fairly

Crime rate rankings are popular, but they require careful interpretation. The FBI itself cautions against simplistic ranking comparisons, noting that many factors affect crime beyond what statistics capture. When comparing cities, keep these principles in mind:

1. Compare crime rates, not raw numbers. Always use per-capita rates to account for population differences.
2. Consider population thresholds. Very small cities can have extreme rates from just a few incidents. A town of 2,000 with one murder has a murder rate of 50 per 100,000, which appears alarming but reflects a single event.
3. Account for reporting completeness. A city that reports 12 months of data will show higher totals than one reporting only 9 months. Check whether agencies reported for the full year.
4. Look at multiple years. A single year can be an anomaly. Trends over 3-5 years are more meaningful than any single data point.
5. Understand what drives rates. Factors like poverty, population density, urbanization, age demographics, policing strategies, and seasonal tourism all influence crime rates.

The rankings section on PlainCrime applies population minimums to filter out statistical noise from very small jurisdictions.

Limitations of UCR Data

No crime dataset is perfect. The UCR program has several well-documented limitations that users should understand:

• Voluntary reporting. Agencies choose whether and how much to report. Some jurisdictions have historical gaps.
• Reported crimes only. UCR captures offenses reported to police, not all crimes that occur. The Bureau of Justice Statistics estimates that only about 40% of violent crimes and 30% of property crimes are reported.
• Definition changes. The FBI revised its definition of rape in 2013, broadening it from the 1929 definition. This makes pre-2013 rape data not directly comparable to later years.
• No demographic context. UCR tables show crime totals and population but do not include socioeconomic variables like poverty rates, education levels, or employment that provide essential context.
• Agency-level variation. Different agencies may classify borderline incidents differently, such as whether an assault is "aggravated" or "simple."

For a deeper exploration of these issues, see our guide on why crime statistics can be misleading.

Frequently Asked Questions

What is the FBI UCR program?

The Uniform Crime Reporting (UCR) program is a nationwide statistical effort managed by the FBI. Nearly 18,000 law enforcement agencies voluntarily submit crime data, which the FBI compiles into annual reports covering offenses, arrests, and law enforcement personnel. The program has been running since 1930.

What is the difference between Part I and Part II offenses?

Part I offenses are the eight most serious crimes tracked by the UCR: murder, rape, robbery, aggravated assault (violent crimes), plus burglary, larceny-theft, motor vehicle theft, and arson (property crimes). Part II offenses cover 21 additional categories including fraud, vandalism, drug offenses, and DUI, but are only tracked through arrest data rather than reported incidents.

How are crime rates calculated per 100,000?

Crime rates are calculated by dividing the number of reported offenses by the population, then multiplying by 100,000. For example, a city of 50,000 people with 200 burglaries has a burglary rate of (200 / 50,000) × 100,000 = 400 per 100,000. This normalization allows fair comparison between places of different sizes.

Why do some cities not appear in FBI crime data?

UCR participation is voluntary. Some agencies submit incomplete data, report for only part of the year, or do not participate at all. The FBI excludes agencies that report for fewer than 6 months. The ongoing transition from the Summary Reporting System to NIBRS has also caused temporary reporting gaps.

What is NIBRS and how does it differ from the traditional UCR?

The National Incident-Based Reporting System (NIBRS) captures detailed data on each criminal incident, including relationships between victims and offenders, circumstances, and whether force was used. The traditional Summary Reporting System only provided aggregate monthly counts. The FBI mandated NIBRS transition by January 2021, but adoption has been gradual.