BetaGov Spotlight: Renée Mitchell
Date: March 6, 2019
Renée J. Mitchell has worked for the Sacramento Police Department (SPD) since 1998 and is currently a sergeant in the Court Liaison Unit. She was the principal investigator on a department-led, 90-day randomized control trial on hot spots policing that won the 2012 IACP/Sprint Excellence in Law Enforcement Research Silver Award. Renée is also a member of the California Bar Association, the American Society of Evidence-Based Policing, the International Association of Chiefs of Police, and the International Association of Crime Analysts.
Please briefly describe your previous experience with field experiments
I have run or participated in several RCTs now. The Sacramento Hot Spots study in 2011 was my first RCT. If I am not mistaken, this was the first RCT run entirely by a police practitioner in the world :) This study examined the effects of 15 minutes of high-visibility policing in 21 crime hot spots. We ran the experiment for 90 days. We found a 25% reduction in Part I crimes and an 8% reduction in calls for service (CFS). The results were published in Justice Quarterly, Journal of Experimental Criminology, and Cambridge Journal of Evidence-Based Policing. I also participated in an RCT in Portland to examine the effects of community-engagement patrols in hot spots over a 9-month period. This experiment showed no overall effects on crime, CFS, or community perceptions of the police.
I am currently running an RCT with the Sacramento Regional Transit Police Services (which is part of Sacramento Police Department contract services) with BetaGov. We are examining the effects of communication training for transit agents on the number of calls for service the transit agents make for the officers as an assist on their calls. We are using the CFS as a proxy measure for the agents' inability to communicate with the customer. I'm also partnering with BetaGov on a study in Grand Prairie, TX. This study was borne out of a training I did for Grand Prairie on evidence-based policing. The assistant chief of the department, Ronnie Morris, and I were discussing the idea of a beat-less patrol versus a standard patrol and thought no one had really done much patrol testing (other than hot spots) since the Kansas City Patrol Experiment and we wanted to do something reminiscent of that study. I wanted to see if having a beat-less patrol focused on hot spots improved policing on several measures—crime, response times, CFS, employee outlook, etc. I think random patrol is really outdated, but policing continues to respond in the same way because we don’t have time in policing to really examine other ways to patrol. This study allows us the time to examine if this patrol strategy works or not. We were not able to randomize the city because to do a beat-less patrol you need multiple beats so we divided the city into North and South and are analyzing outcomes using a difference-in-differences, quasi-experimental design.
How has running field experiments influenced the way you think about innovation in the public sector?
Being involved with field experiments has made me realize many things about innovation. First, you don't need a lot of money to run experiments. You can run simple RCTs without a lot of overhead. You just need the passion, the determination, and someone with the skill set and you can do it. Second, RCTs give you a lot of information about your organization and how your technology works, where the bottlenecks are, who creates those bottlenecks, etc. Experiments help you learn whether your innovations actually are effective or efficient. I have had many, many innovations over the years, but looking back I never knew if they were effective or not. Field experiments allow you to understand this.
Do you feel police leadership appreciate the role of data in making policy and procedure decisions (and is that changing over time)?
I am not sure that police leadership quite appreciate the role of data in policy and procedure decisions the way they should. I do not think they understand data well, meaning they don’t understand regression to the mean or why you need a comparison group to understand whether an intervention works or not. I think police leadership still rely on heuristics more than anything.
What advice would you give to someone who wants to run their own field experiment?
Often leadership will support a project as long as they believe it will achieve a positive outcome and if they don't they will remove their support. Social capital is everyone's lifeline in an organization. When a person doesn't have social capital they easily become a pariah in their organization. If an experiment fails or is associated with negative outcomes and social capital is removed then this will set the standard for the rest of the organization and no one will want to attempt another field experiment again. The advice I would give to someone running their own experiment is to make sure that your leadership supports your experiment, meaning that they support the experiment no matter the outcome.