So, Chip Johnson’s column today is sort-of about general problems with the police department and sort-of about CompStat. I’ve noticed that the word CompStat is getting thrown around more and more often in discussions about OPD, and while I’m thrilled about that because I think we desperately need to do it here, I also think that it sometimes seems people aren’t exactly clear on what it is.
So, CompStat, basically, is a program developed in New York in the mid-1990s by then NYPD Chief William Bratton . He has since brought the program to his new post in Los Angeles, and many of his former deputies have implemented it in the departments they now head. CompStat-like systems, or at the very least the program’s driving principles, have been adopted by tons of agencies large and small throughout the country.
CompStat, at its core, is about using technology to assess real-time crime data and continually redeploying resources in response. At the October 14th Public Safety Committee meeting, the Police Department presented an update of their strategic plan, and much of the discussion centered around implementation of CompStat in Oakland, which the department identified as their top strategic priority, which an implementation goal of July 2009.
So what do you need to do CompStat? First, you need software. CompStat itself is not software – it’s a process that uses real-time data and pattern identification provided by software to shift resources to where they can be most effective, help departments address crime in a somewhat pre-emptive fashion, measure progress, and create greater accountability. Right now, our first challenge is to get the software we need to do that. Every department uses something different – LAPD, for example, built their crime data system themselves. We’re not going to be doing that, because we really don’t have anywhere near the resources in the City that such an effort would require. So instead, we’re buying (or, hoping to buy) a system from Bishop Rock.
Oakland’s IT department has been, and as far as I know, still is (all this information was presented at the October 14th meeting, it’s possible things have changed in the meantime), working with Bishop Rock to develop a program for Oakland. Because data collection and entry, as well as information needs, are different in every city, Bishop Rock has to customize their software for any agency interested in purchasing it. So right now we’re working with them on development and testing (at no cost to us), and specifically trying to work out some issues with data integration and data validation. The IT department has created a timeline for Bishop Rock with defined deadlines for meeting our needs, and hopefully, that’s going to work out. If it doesn’t, then we’re kind of back to square one. That would mean looking for other companies to work with. They have, as a back-up in case Bishop Rock doesn’t work out, begun looking into whether its possible to purchase and use LA’s system.
Anyway, getting the software is just the first hurdle. CompStat isn’t magic. It’s technology, and just like with all other technology, its utility is dependent on how you use it. It doesn’t matter how much expensive photo-editing software my Mom has on her computer if all she’s going to do is crop and resize her images for printing. Similarly, it doesn’t matter what expensive crime data software OPD has if they don’t have anyone to analyze that data. Implementing CompStat will require more crime analysis staff, and when and if the discussion ever reaches this point, it’s really important that we find a way to fund these positions. One investigation into the use of CompStat in three different cities (PDF) found that all three failed to fully realize the potential benefits of the CompStat system because they didn’t properly fund the accompanying analysts it needs to work:
Given the ostensible importance of strategic crime analysis to these Compstat programs, it is noteworthy that departments failed to give their CAUs enough staff, training, and support. As a result, the CAUs could only provide police decision makers with limited, and fairly rudimentary crime analysis. This is not to demean the CAUs contributions, which are quite impressive, especially given the many challenges involved merely to process data to a state that the software could analyze. It does, however, indicate that the accouterments of new information technology and data analysis were introduced into the organizations without preparing their members to make the most effective use of them.
It’s great that we’re doing CompStat. And having decent crime tracking software is in and of itself a huge step forward. At a Public Safety Committee meeting a couple of months ago, a Police Department representative told the Committee that he had no idea what had been happening with respect to certain types of crime in the month or so since the data in the report was compiled, because their crime data software had been down for weeks. Obviously unacceptable. But it’s really important, if we want to program to have a real impact on crime in Oakland, that we make sure not to just adopt a skeletal version of the program and pat ourselves on the back for copying a best practice. We need to make sure its fully implemented, properly implemented, and – this is the part I’m most worried about in a time of budget crisis – fully staffed. The Council has to fund the staff, and we hope the Police Department can properly handle the rest. But the Council also needs to start exercising more vigorously their oversight role, and make sure, when we finally do it, that we do it right.