Today, the Bay Citizen reported that the San Francisco Police Department has been misclassifying arrest rates for Hispanic and Asians because of an outdated computer database -- installed in 1972 -- that lists only three racial categories: "blacks," "whites," and "other." (Which seems primitive even for 1972).
As noted in last week's feature, "Menace to Society," San Francisco's numbers in that database stand out: As of 2010, 59 percent of gang members were black and 31 percent were Hispanic. It's a notable chunk for a city that is just 6 percent black. By comparison, less than 20 percent of all gang members in California were black. No other county had a proportion higher than 29 percent.
Here's how the rest of the city's CalGang data looks: 4.5 percent white; 3 percent "designated race;" 2.3 percent Asian or Pacific Islander. The numbers entered into the database are not necessarily the numbers local authorities use. In CalGang, San Francisco had 79 gangs and 465 gang members. In a 2011 grant application for a state anti-gang initiative, though, San Francisco had 41 gangs and 2,500 gang members.
But the antiquated arrest input system does not necessarily suggest that the CalGang data might be off. While the racial disparities in the arrest statistics are clearly off the mark, San Francisco's CalGang proportions do reflect real-world numbers: Of the 93 people on the city's gang injunction list, about two-thirds of them are black and roughly one-third are Hispanic.
What the arrest data does show is how police stats can be wrong for years, without the police department noticing. Or, for that matter, caring.
As the Bay Citizen reported:
"You're making it sound like officers choose to do this. It's what the system has available to the officers to put in," [Deputy Chief Lyn] Tomioka said.
She said she did not know when the department began misclassifying arrestees but said it does not plan on "looking back at those statistics."
The police department has no idea if any of the statistics it reports to the state are accurate, according to Susan Giffin, its chief technology officer.
"Not only can we not tell you if the numbers are right, we really can't articulate what the problems are, or if there are problems," Giffin said.