Over the weekend Mayor Bloomberg defended his comment that the NYPD's stop, question and frisk program "disproportionately stop[s] whites too much and minorities too little." He argued on Sunday that, "The numbers are the numbers, and the numbers clearly show that the stops are generally proportionate with suspects' descriptions."

Specifically, as the Wall Street Journal reported, City Hall noted Friday that whites made up 7 percent of murder suspects in 2012 but were stopped 9 percent of the time.

That was a slightly more specific twist on a long-time argument of the Bloomberg administration: If you look at the race of people whom witnesses identify as suspects in violent crimes, about 90 percent are black or Hispanic. So, their logic goes, that's who the cops stop.

But how relevant are murder statistics—or even more comprehensive violent crime numbers—in evaluating stop and frisk? Race aside, is violent crime really what the program is about?

Last year, City Limits crunched the numbers on police stops in Sector E, a section of East New York. East New York's 75th precinct topped the city in stops in 2011, and Sector E was the patrol sector within the 7-5 that had the most encounters.

We used a detailed database that the NYPD provides on its website, which lists important facts about every stop—including the crime that police suspected before initiating each encounter.

In Sector E in 2011, some 23 percent of the stops were for violent crimes—specifically, robbery (22 percent) and assault (1 percent). Criminal possession of a weapon, which is a nonviolent crime but obviously involves the threat of violence, represented 24 percent of stops.

But 34 percent were for non-violent crimes, like car theft, larceny or criminal trespass. Another 16 percent of stops were identified only as "felony" or "misdemeanor," which could mean violent or non-violent crime. So anywhere from a third to a half of stops were on suspicion of something other than violent offenses.

When it came to arrests in Sector E, nonviolent crime dominated. Weapons possession and assault made up roughly a quarter of the busts. Sixty-one percent of the collars involved nonviolent offenses like criminal possession of marijuana, trespass or loitering.

Is the racial breakdown of nonviolent crime different from that of violent crime? That's unclear because there are a lot of different kinds of nonviolent crime. But if you take marijuana possession, statistics on usage suggest whites deserve as much police attention as blacks.