December 2, 2023

The challenge of implementation


PERF members,

The policing profession spends a lot of time discussing and studying programs and policies, and less time thinking about how they’re implemented. Why can a program that’s effective in one city be a disaster in another city? How much of a program or policy’s effectiveness can be attributed to the way it was implemented?

At the American Society of Criminology conference two weeks ago in Philadelphia, PERF convened a panel discussion on implementation with practitioners and academics. I want to share a few highlights from the session.

Former Camden County, New Jersey Chief Scott Thomson discussed his experience implementing the ICAT training program. One lesson he learned was that it’s important that the change not be optional. He made it part of the agency’s policy, and there were consequences for not following agency policy.

He found that he could not have the training taught by the staff assigned to the training academy, because telling them to change their old ways of doing things would be seen as invalidating their previous work. Instead the training was taught by cops working in the field who were widely respected and open to new ideas.

Years ago Chuck Ramsey, who served as police commissioner in Philadelphia, police chief in Washington, DC, and deputy police superintendent in Chicago, told me he had a similar experience when implementing the Chicago Alternative Policing Strategy (CAPS) program. He bypassed the academy and selected cops from the field to teach community policing. Police academies are not always open to new ideas, so police leaders may need to find others who can serve as “change agents” in their agencies.

The conversation turned to research about implementation. Northeastern University Professor Eric Piza discussed the lack of research about implementation science. He said most research has focused on program and policy evaluation, and the field needs to start thinking about implementation as well.

Jeremy Barnum, PERF’s deputy research director, said he’s starting to see more agencies with their own research staffs to help plan for, track, and assess implementation. Going forward, these research teams may be able to share lessons learned about the best ways to implement new policies and programs.

Brandon del Pozo, who was the police chief in Burlington, Vermont and is now an assistant professor at Brown University, agreed there’s a lack of research and said he could only find a few studies about implementation. He also brought up the need for research into “de-implementation,” such as how to stop cops from shooting at moving vehicles. As a former police chief who is now a researcher, Brandon understands why implementation science should be a priority for researchers.

Peter Neyroud, who served as chief constable of the Thames Valley Police in England and led the UK National Policing Improvement Agency, said there has been considerable research about implementation in public health, and he believes many of those lessons would be transferable to policing. He said turnover among police chiefs is also a challenge. The typical police chief’s tenure is less than four years, so most chiefs won’t last long enough to see their changes fully implemented. And he said another challenge is that there’s a decent chance that the change being made will fail. He had a few recommendations for police leaders implementing change:

  • Never introduce a new idea to the entire department.
  • Never start at the police academy.
  • Track outcomes and share data.
  • Capture and spread stories, because cops hate figures and want stories with endings.
  • There has to be a point where you stop calling it “implementation” and it just becomes the way you do business.

Peter said that some agencies in the UK do a “pre-mortem” to identify everything that might fail and address these potential failures before implementing change. By anticipating potential failure points and adapting plans accordingly, agencies can increase the likelihood of successful implementation.

Meagan Cahill, PERF’s research director, said that when implementing a new strategy, agency leaders should understand that the work doesn't stop once all officers go through an initial training. They should think about what happens if implementation is successful. Chiefs should especially consider how the change could be institutionalized, to ensure it stays regardless of turnover in agency leadership.   

Blake Norton, chief strategy officer at the Philadelphia Police Department, said that when implementing a program that’s been successful in another city, it’s important that the program be championed by both the leader of the department and other internal stakeholders.

Voorhees, New Jersey Chief Bill Walsh said that when implementing a new program, he takes the time to explain the reason for the change. For example, when he began requiring officers to see a psychologist once a year, he had to take the time to explain that it was a wellness check, not an evaluation of whether they were fit for duty.

I have a few lessons about implementation that I’ve learned over the years:

  • There often is no sense of urgency to implement change until there’s some sort of crisis. An unexpected crisis can present an opportunity to implement a new policy or program.
  • Any complicated new policy or program should be piloted first. When agencies began using body-worn cameras, many piloted the cameras with a small portion of the agency and learned from that experience.
  • The message from the chief or sheriff is important, but the real change agents are the first-line supervisors. Without them no change will happen.
  • As Scott Thomson pointed out, any changes to training must also be put into policy, and officers must be held accountable for following that policy.
  • Agencies do not do a good job of eliminating obsolete policies, some of which may conflict with new policies.
  • Any changes must be measured, and success stories should be shared.
  • Finally, when a chief pilots a new program, the person they select to lead it will be someone they know is terrific and has a good chance of successfully implementing the pilot. The challenge is to find that same level of enthusiasm and capability throughout the organization.

Thanks to everyone who participated in this discussion!