This is a pretty interesting idea:
The annual self-evaluation JAG recipients are required to complete measures performance in a way, says the Brennan Center report, that is "roughly analogous to a hospital counting the number of emergency room admissions, instead of considering the number of lives saved." Agencies are asked how many arrests they made, and prosecutors are asked how many cases they won. Not only is that data rather useless in terms of assessing the effectiveness of a given policy, it also says to the person answering the questions that their numbers should be really big.Via The Atlantic Cities
Whenever I look a statistical performance measures for the police department where I work, I am always hesitant to include data for arrests and traffic citations. The reason being is that these can be pretty poor ways for an agency to measure how effective they are being at making the community safer.
These are numbers that are pretty easy to jack up. If your agency is being criticized, then you go out and make a bunch of low level arrests for trivial offences or you go out and write a bunch of tickets for people driving just a few miles an hour over the speed limits. You'll have an impressive chart showing an increase in these activities but you haven't demonstrated that crime has been decreased or that there are fewer traffic accidents.
It will be interesting to see what changes the Justice Assistance Grant program makes in measuring how effective these grants are.