Justice data poorly calibrated to researchers' needs

At a Austin City Council's Public Safety Committee meeting yesterday discussing 911 call data, the Police Department's Chief Data Officer, Jonathan Kringen, reiterated an observation that Grits has made several times over the years: "this data system is built to try to manage an activity. It's not actually built to support researcher analysis." 

That's the biggest problem with most criminal-justice number crunching: Data is collected at points when the bureaucracy either interacts with other government systems or makes decisions about individuals being processed. But those data serve the needs of government tracking, they can't explain the causes of crime or explore dynamics that occurred prior to government intervention. In general, we don't measure the things needed to answer the most critical questions about public safety because the data is collected for different purposes.

As Kringen explained, "there are many things in the data that the way they are set up are functional and they serve the operational need, but they don't actually serve the analytic function that should come thereafter to help figure out what best practices should be."

Amen, brother.

Here are a couple of related academic discussions about the paucity of useful criminal-justice data. The latter article argues that, "Blind trust in the police, fear of a powerful police lobby, and a law enforcement culture of secrecy and insulation results in a serious information failure."

Here's the power point presentation Austin city staff gave to the committee, video of their Zoom call, and here's a data dashboard providing various analyses of APD 911 data. Grits may dig deeper into these dashboards later.

Related: Three quarters of Austin PD 911 responses went to non-crime situations: National experts analyze Austin PD data for clues on how officers spend their time