According to this week's National Crime Victimization Survey - or NCVS - violent crime fell 15 percent in 2000 while property crime fell ten percent. These findings, based on a Census Bureau survey of 160,000 Americans, are out of sync with the FBI's annual compilation of police statistics - the Uniform Crime Reports - which showed that absolute crime reports remained flat between 1999 and 2000 while the crime rate declined slightly. Between 1992 and 2000, America saw the longest and steepest period of crime reduction since record keeping started in the 1930s.
Overall, the latest NVCS appears to provide a lot of good news. Rape fell by about a third; robbery, 11 percent; and motor-vehicle theft, 14 percent.
Only two subcategories, attempted motor-vehicle theft and attempted property theft, saw increases. Even this isn't bad news because it may well show that improved security measures have begun to foil criminals.
So, why the large discrepancy? Well, to begin with, the two studies look at different things: The NVCS is a telephone survey that asks respondents whether anyone in their household fell victim to a crime during the previous year, while the FBI surveys crimes reported to local police departments. As a result, the NCVS doesn't account for what happens when crime rates change dramatically for a particular city or demographic group. Indeed, it's possible for crime rates to go up as victimization goes down if fewer households fall victim to more crime.
The survey provides some evidence that this happened last year: Latinos reported just as many property victimizations as the year before even though every other sizeable demographic group saw declines. But this can't explain the discrepancy alone because groups that have borne the brunt of previous crime waves - poor people and African-Americans - saw continued reductions in victimization and even Latinos didn't fall victim to any more crime than the previous year. The NCVS, however, can't identify every demographic group. If a sufficiently delineated demographic group, such as white suburbanites with high-school-age children, began falling victim to dramatically more crimes, then the study would probably miss the shift.
The two studies also capture different types of crime. Many people won't bother to file a police report for minor property crimes like the theft of a garden hose but will remember them when a survey-taker calls. Victims of rapes and other particularly traumatic assaults, likewise, may decide that the difficulties of reliving their experience outweigh the benefits of bringing the attacker to justice. In addition, the UCR reports that the FBI released in May gave crime rates only for cities with more than 100,000 residents. If crime remained level in major cities while falling significantly in the suburbs, then that could explain the discrepancy. But suburban and urban crime rates typically move in sync with one another.
One factor certainly explains part of the discrepancy: People reported more crimes to the police in 2000 than they did in 1999. According to the NCVS, the percentage of crimes reported to the police hit historical highs during 2000. Reporting rose most dramatically amongst blacks and Latinos. Still, the reporting levels rose only modestly for the population overall.
No survey can give us a fully accurate reading of the exact number of crimes committed in a given year, but the sum of several surveys can show us the direction of America's crime trends. The most recent NCVS survey indicates that America's progress against crime hasn't ground to a sudden halt, but no obvious factor can explain dramatic decreases in victimization rates. It seems likely that some other factor, yet unknown, will befuddle criminologists for the next few years.
Eli Lehrer is visiting fellow in the Center for Legal and Judicial Studies at The Heritage Foundation.
Originally published on National Review Online (06/14/01)