Austin PD's "early warning system" is a failed PR stunt, like pretty much all of them

"Early warning systems" for police are one of those ideas that's touted by institutional players in the wake of bad police misconduct episodes - doubling down on the idea that we just need to identify and oust a "few bad apples" - but IRL, your correspondent has never seen one that worked well.

When I was Police Accountability Project Director at ACLU of Texas back in the day, I spent a fair amount of time trying to understand what information might be probative for managers to monitor in an "early warning system," and concluded that 1) there's no consensus about how these programs should operate and 2) in practice, they were touted by officials as a solution but never had real-world impact. As a result, it's not a reform Grits generally recommends.

It's been many years since I've been that deep in the weeds on the topic, but a new Austin city auditor report on their police department's "early intervention system" - known within the department as the "Guidance Advisory Program" (GAP) - confirms my sense that they're essentially worthless. Austin's, the auditor found, "does not effectively identify officers who may need assistance."

As is typical, there has been no local MSM coverage of the audit. (I know, gentle readers, you're shocked at the omission!)

APD's police early-warning system suffers both from over-identification and under-identification. It gathers only three, not-very-probative data points and ignores data used by systems in other cities. The thresholds to trigger review are set too low, so too many officers are identified for intervention and the system has little predictive value. At the same time, many officers meeting thresholds are not identified at all. On use of force (at APD, called "response to resistance), the department failed to identify about a third of officers who should have met the threshold for review. Moreover:

When officers are identified for assistance, the GAP does not connect these officers to existing APD support or wellness services. Also, APD does not track or analyze program trends to evaluate officer or program performance to ensure the GAP is fulfilling its mission. In addition, APD management has not generated true program buy-in and the GAP is not working as intended.

The auditor sampled 60 activations and found supervisors identified no issues 93% of the time, resolved the issue with a conversation 7% of the time, and NEVER created an action plan to correct officer behaviors, even though that's theoretically supposed to be triggered by the system. As a practical matter, they're just not doing anything with the information: 

APD staff said there are no performance metrics reported in relation to the GAP and they have no way to measure the program’s success. In addition, the department is not analyzing results to identify trends or determine if certain officers, assignments, or supervisors need additional support services.

Even an officer triggering the system three times in three quarters based on 45 total use of force incidents was found to have displayed no "pattern" that caused concern. Intervention after 45 incidents wouldn't seem particularly "early" to this writer, but if they're not going to review outliers, anyway, IRL it hardly matters.

The reality is, as the auditors wrote, "APD is not creating an environment of trust and transparency" regarding its responses to officer misconduct, either with officers or the public, and failures of the early warning system are a symptom of that broader problem.

That said, none of the other early warning systems in Texas work well, either. There are no real best practices and as a result, their structures are all over the map. Here's a summary from the report of the information gathered in each one, which varies quite widely.

Dallas' last chief Renee Hall proposed spending nearly a million dollars to revamp their system, with no results so far. The one in Houston tracks 10 different metrics, compared to 3 in Austin, but the Mayor's task force on police reform last year found it ineffective and recommended an upgrade (without specifying details).

I suppose it's possible an "early warning" system could be devised that would fulfill the goal of reducing misconduct, but academic reviews have found little evidence for their effectiveness (if plenty of enthusiasm for giving it the ol' college try). Grits believes their popularity stems largely from their PR value: It's something police chiefs can say they're implementing, improving, etc., that will take the heat off them in the near term because they ostensibly need time to launch a new program. The program never seems to work, though, whether they monitor three data points or 10. Then another scandal happens and suddenly we're revamping the early-warning system again.

Austin doesn't need APD to waste time on this pointless paper shuffling and IMO they should scrap it. If managers want a list of officers who need retraining or intervention, they should ask Farah Muscadin, the head of the Office of Police Oversight, for a list. She knows perfectly well who the problem officers are at this point, even if APD brass isn't paying attention.