2 Measuring Crime
Compare official statistics, victimization surveys, and self-report studies to understand what each can reveal about crime and where its evidence is incomplete.
Why crime is difficult to count
Crime is not measured by a single, complete count. Many offenses are never reported, detected, or recorded, and measurement methods define and capture crime differently. Researchers therefore compare multiple sources rather than treating any one set of statistics as a perfect measure.
are compiled from incidents known to law enforcement. In the United States, the FBI’s Uniform Crime Reporting (UCR) Program collects information from law enforcement agencies. These records are useful for examining crimes reported to police, geographic patterns, and cases that come to the attention of authorities.
The records details about individual incidents, including offenses, victims, and known offenders. Compared with older summary reporting, NIBRS can record multiple offenses within an incident and provide more detail.
do not represent all crime. An offense may be missing because no one reported it, police did not identify or classify it as a crime, or an agency did not submit complete data. Differences in reporting and recording practices can make comparisons across places or years difficult. A change in recorded crime may reflect changes in reporting or data collection as well as a change in crime itself.
ask people about their experiences with selected crimes, whether or not they reported the incidents to police. The , conducted by the Bureau of Justice Statistics, asks a representative sample of people in U.S. households about these experiences. It provides estimates of nonfatal personal and household-property victimization and can also gather information about incident circumstances and whether incidents were reported. For example, it can reveal a reported assault that never appeared in police records.
NCVS estimates are subject to because the survey covers a sample. Results also rely on respondents’ memories and willingness to disclose, and the survey cannot measure every type of crime. Its coverage excludes some populations and settings, and it generally does not measure homicide through interviews with victims. Results can also be affected by how questions define and screen for incidents.
ask people whether, and how often, they have committed specified offenses. Researchers often use them to study behaviors that may not come to police attention, including delinquency. Anonymous or confidential surveys can encourage candid answers and help researchers examine the circumstances and frequency of offending.
Self-reports have limitations: respondents may forget, misunderstand a question, exaggerate, or conceal behavior. Results depend on who is surveyed, which offenses are included, and how questions are worded. Some groups or serious offenses may be underrepresented, and there is no single complete record against which every answer can be checked.
Using measures together
Each measure captures a different part of crime. Official statistics describe incidents known to authorities; include selected offenses that were not reported; and offender self-reports can reveal behavior that neither police records nor victim interviews capture.
For example, an increase in police-recorded theft could reflect more theft, greater willingness to report it, or changes in agency recording. Comparing several sources helps researchers assess such possibilities, but it does not eliminate uncertainty. Careful analysis considers what each method includes and misses, along with the effects of definitions, reporting, recording, and research methods.