The numbers provided by the Cicilline administration did not involve the crime RATE, which is what the mayor cites. The crime rate is the number of crimes per 1,000 residents. Cicilline's staff produced the total number of crime reports, a cruder measure.

Yet if you look at population trends, the distinction may not be important. During the '60s and '70s, when the number of crimes was rising, Providence was losing one quarter of its population, which means the crime rate was going up faster than the number of incidents would indicate. The reverse was true during the 1990s, when the number of reported crimes fell, yet the city's population rose by 8 percent.

Thus, the mayor's assertion is still correct.

Second, when we cross-checked the numbers provided by Cicilline's people against the FBI uniform crime reports for Providence -- which are available online -- the numbers stopped matching up, beginning with the 2002 statistics.

Three members of the Providence Police Department and two representatives of The Providence Plan, which does data collection for the department, met with us to explain the discrepancy.

The FBI numbers originate from the Providence Police Department and they reflect the count as of Dec. 31 of each year. After they are reported to the FBI, those numbers change as late reports of crimes, particularly property crimes such as burglary and larceny, come in to the department, often from insurance companies. In addition, a crime may be reclassified as more information from an investigation comes in, said Chief Dean Esserman.

The numbers also sometimes differ because the FBI has been requiring departments to change their crime reporting systems to count the victims of crimes. Thus, a single robbery that had three victims is now reported as three robberies.

Providence incorporated the new method in 2007, but it also compiles data using the old standard, so year-to-year comparisons remain valid. All the statistics provided to us were based on the old system, said Jim Lucht, director of the Information Group of The Providence Plan.

But here's the important point. We found that in every year since the city's and the FBI's numbers diverged, the Providence police have reported MORE total crimes than the FBI. If Cicilline wanted to make the decline in the crime rate look even more favorable, he could have cited the FBI numbers. Instead, his administration is using the older system for compiling such statistics.