Saturday, November 13, 2004

The police are doing a Cambridge High

The Herald reports that police in some parts of the country are being ordered to ignore P-labs and gangs because their commanders do not want to make crime statistics look bad. While National will no doubt claim this is a government conspiracy to hide the true impact of crime, the real explanation is buried deeper in the story. According to the Police Association, district commander's performance (and hence bonuses and promotions) is measured by their district's crime statistics.

The problems with this form of performance measurement are well-known. While using solid and clear performance metrics makes managers pay attention to them, the danger is that they will refuse to pay attention to anything else. The solution to this is to ensure that your performance metrics are in fact a complete description of the task, so that managers have to pay attention to everything. But this still creates problems where some measures are easier to influence than others; managers will pay disproportionate attention to these, in the hope of boosting their measured performance. And that is exactly what has happened here: cleaning up P-labs is hard, and expensive, and even saying that you busted one makes you look bad because it was there in the first place. Far easier to ignore them and focus on offending that is cheaper and easier to reduce. There's also an obvious incentive to focus solely on crime reported by the public, rather than "making trouble" by actively seeking it out.

If this all seems dreadfully familiar, it's because its exactly the same problem which led to the mess at Cambridge High. And the solution is the same: use some "fuzzier" metrics to ensure that police managers act like policemen, and principals like educationalists. Until then, they'll continue to manage by statistics, and focus on gaming the numbers to their advantage to the detriment of the wider goals of their profession.