By the Numbers: Measuring Bad Behavior; September 2001; Scientific American Magazine; by Rodger Doyle; 1 Page(s)

The Uniform Crime Reports (UCR), produced by the Federal Bureau of Investigation, have figured prominently in discussions of crime since at least the Nixon era, but their reliability has long been suspect. A major reason is substantial underreporting. For a variety of reasons, including distrust of law-enforcement officials, many crimes are not reported to local police departments, the source of the FBI data. Furthermore, the number of crimes that police departments report can vary from year to year depending on budgets. The FBI cannot legally enforce the cooperation of local police departments and state agencies, and so it is not surprising that for several years in the 1990s, six states (the largest in terms of population was Illinois) supplied no data, forcing the FBI to estimate the number of crimes in those states.

Local police sometimes cook the books, either underreporting to make crime in their area appear to be under control or overreporting to support requests for more funding. Fabrication of this kind has presumably declined as police departments have become more publicly accountable in the past few decades, but it still persists, as recent reports of data manipulation in New York City, Philadelphia and Boca Raton, Fla., testify.