Key facts
The FBI's UCR counts crimes reported to police. The victimization survey counts crimes people actually experienced. When their numbers disagree, the disagreement is the story.
Every crime debate in America is secretly two debates. One side is arguing about the FBI's numbers. The other side is arguing about the victimization survey's numbers. They're both right and neither is complete — because the United States measures crime with two instruments, built for two different jobs.
Instrument one: the UCR. The FBI's Uniform Crime Reporting Program, administered by the Bureau since September 1930 (the International Association of Chiefs of Police started it in 1929), counts what police know. Agencies report monthly, and the program tracks murder and nonnegligent manslaughter, forcible rape, robbery, aggravated assault, burglary, larceny-theft, motor vehicle theft, and arson — plus arrest data for 21 additional categories. It's an administrative record: big, official, and limited to crimes somebody actually told the police about.
Instrument two: the NCVS. The Bureau of Justice Statistics' National Crime Victimization Survey, launched in 1973 and redesigned in 1992, asks people directly. Each year, BJS interviews a nationally representative sample of roughly 169,000 people age 12 or older, and it captures crimes whether or not they were ever reported to law enforcement — including the reasons victims give for staying silent. What it can't do: measure homicide (victims can't be interviewed) or commercial crimes like store burglaries.
Why do the two sometimes disagree? Three reasons the FBI itself spells out. First, they measure an overlapping but nonidentical set of crimes: the NCVS includes crimes never reported and simple assaults, but excludes homicide, arson, and commercial crime; the UCR includes those, but collects only arrest data for simple assault. Second, their definitions differ by design — the UCR defines burglary as unlawful entry to commit a felony or theft (which requires knowing the offender's motive), while the NCVS, which doesn't ask victims to read motives, defines it simply as entry of a residence by someone with no right to be there. Third, they compute rates on different bases: the UCR reports property crime per 100,000 persons; the NCVS reports per 1,000 households.
None of this means the numbers are useless. The FBI notes that once analysts adjust for the known differences — for instance, comparing NCVS police-reported robbery rates against UCR noncommercial robbery rates — the long-term trend lines for robbery, burglary, and motor vehicle theft fall into close concordance. And each program has a strength the other lacks: the UCR's homicide data remain the most reliable and timely measure of killing in the nation, while the NCVS is the only window into the crimes that never reach a police blotter.
So the next time two politicians cite "the crime statistics" and sound like they're describing different countries: check which instrument each one is holding. The honest answer to "is crime up or down" starts with a better question — according to which count?
---
Reported with AI assistance from public sources. Align Newsroom's standard: every factual claim above carries its evidence link in the article metadata.
Sources & evidence
- The U.S. Department of Justice administers two statistical programs measuring crime: the FBI's Uniform Crime Reporting (UCR) Program and the Bureau of Justice Statistics' National Crime Victimization Survey (NCVS). corroborated
- The FBI has administered the UCR Program since September 1930 (the IACP began publishing in 1929); the NCVS began in 1973 and underwent a major redesign in 1992. corroborated
- The UCR compiles crimes reported to law enforcement (murder, rape, robbery, aggravated assault, burglary, larceny-theft, motor vehicle theft, arson, plus arrest data for 21 other categories). corroborated
- The NCVS interviews a nationally representative sample of about 169,000 people age 12+ each year and captures crimes whether or not they were reported to police; it does not measure homicide or commercial crimes. corroborated
- The programs' definitions differ (e.g., UCR burglary requires unlawful entry to commit a felony or theft; NCVS defines it as entry of a residence by someone with no right to be there) and they use different rate bases: UCR per 100,000 persons, NCVS per 1,000 households. corroborated
- The two programs' long-term trend lines can be brought into close concordance after adjusting for their definitional and procedural differences (e.g., comparing NCVS police-reported robbery rates with UCR noncommercial robbery rates). corroborated
Reported with AI assistance from public sources; reviewed before publication. — Powered by Creytix.