Jason Potts is a Lieutenant with the Vallejo Police Department where he has served for 17 years. He is a National Institute of Justice (NIJ) LEADS Scholar, an ASEBP board member, and a Police Foundation Fellow. He is also a Reserve Special Agent with the Coast Guard Investigative Service and has a background in drug, gang, and general investigations. He is a former member of the Vallejo Police Department SWAT Team and possesses a Master of Advanced Studies (MAS) in Criminology, Law and Society from UC Irvine.
Recently two National Institute of Justice LEADS Scholars, Captain Cory Nelson of the Madison Police Department and Sergeant Jeremiah Johnson 1 of the Darien Police Department contributed articles to Translational Criminology (Spring 2018 edition) – a magazine of the center for evidence-based crime policy at George Mason University – see attached (2, 3). Board members at ASEBP recommend reading these compelling articles. The following are short summaries:
Captain Cory Nelson described in a persuasive piece, his initial reluctance for evidence-based policing before being selected as a LEADS Scholar, citing concerns that it might not translate to the real world (2). However, he was fortunate to observe a master class on evidence-based policing by esteemed Professors Cynthia Lum and Chris Koper and because of that course, his enthusiasm and shift in thinking increased exponentially – especially when hearing of the potential benefits of implementing the Koper Curve at his police department. As we know, the Koper Curve Principle suggests that police can effectively create a residual deterrent effect by spending 15 minutes every two hours at hotspots engaged in various activities such as citizen engagement. He and the crime analysts employed by Madison PD targeted tested and tracked shooting incidents in multiple hotspots. Armed with research examples, they then formulated a plan where officers would visit these hotspots four times a day during an eight-hour shift. While implementing their interventions, something remarkable happened to Cory and his colleagues. He discovered that officers thoroughly enjoyed being given clear direction and purpose toward fighting crime in specific problem areas. During the three months of their interventions, Madison PD observed a 72 percent decline in violent crimes and a 66 percent reduction in shots fired incidents and hotspot patrols/engagements during the testing and tracking of those hotspot areas. Cory and his colleagues, seemingly debunked many who believe cops dislike deterring crime as it isn’t perceived as being very exciting– me being one of them. I genuinely believe that most cops enjoy the hunt, the benefits of autonomy and discretion, and really just dislike the boredom of deterrence, but when you seemingly provide clear task where they are empowered to see and own the outcomes, then we see the results of evidence-based policing concepts such as what was impressively done in Madison to combat shooting incidents. Most veteran cops realize the perils of boredom for cops, which occasionally creates the antithesis of effective policing, but at the same time, most would agree with the assertion that “crime prevention coupled with the absence of crime and disorder is preferable to the visible evidence of police action.”(4)
Sergeant Jeremiah Johnson smartly disentangled the confusing meaning of a P value. In his piece, he explained the importance of starting with the notion of a hypothesis and described an example that most cops understand. For example, if a person is weaving in and out lanes at 0130 AM on the weekend, officers will hypothesize that the driver might be under the influence. He also shared another symbolic similarity that criminal courts and scientists both begin with similar presumptions – in both realms a threshold must be overcome in order to prove something. However, when thinking about this from a research perspective and the importance of evaluating the hypothesis with statistical significance, then the idea is enthusiastically relevant to policing. For example, one might suggest that automatic license plate readers (ALPR) lead to more arrests for auto theft. However, how do we know that unless we evaluate and test the data? This concept is essential, as Jeremiah stated in his paper, “Knowing how to interpret a statistical test of this hypothesis is essential to knowing whether organizations should continue spending time, money, and risking officers’ lives for that operation.”(3) Or in the previous example about ALPR, continue to spend money on police technology which may or may not have significant policy implications but in policing, we would never know unless a particular project or intervention was targeted, tested, and tracked.
The next concept that Jeremiah discussed, of course, is the p-value and the numeric threshold of .05. Jeremiah concisely informs the reader that in simple terms, a p-value is a measure of the probability that would not have occurred merely by chance. He used an excellent example of the game of craps and the odds of rolling a seven as compared to a 2 or 12. However, with the risk of confusing the matter and as many jurisdictions set different levels of blood alcohol content, Jeremiah stated in his paper that there is no bright line of significance. Even though most researchers set their threshold at p =.05, this is not set in stone. Just as states changed their DWI laws from .10 to .08, scientific consensus may eventually move away from p = .05
Finally, not all research is created equal, and the final point from Jeremiah’s paper is on sample sizes and research methodology. For example, Jeremiah described a concern one would have if there was a thought that someone had a loaded dice and had correctly hypothesized that the odds of continuing to roll a 7 would not occur in a series of rolls. However, this would be flawed– especially if there was a limited sample size of rolling the dice 20 times versus 150 times and evaluating the data of obtaining many more 2 or 12’s over rolling a 7 and the resultant statistical significance of reaching that threshold of 5 percent or .05 or an even higher threshold of .01. Lastly, Jeremiah discussed the importance of not relying heavily on just statistical significance but looking at such conclusions from an evidence-based policing matrix – the one created by Cynthia Lum and Chris Koper (5). Or running a randomized controlled trial (RCT) as opposed to a quasi-experimental design, or the limitation of completing an anecdotal policing project at best without any rigorous evaluation of empirical data, or worst no evaluation of any data at all. The following was quoted from Jeremiah, “As the evidencebased policymaking and evidence-based policing movements grow, there is a danger that public organizations could latch on to research that is methodologically weak or statistically flawed, but that show statistical significance.” Translational efforts play an important role by not only pushing out robust research to practitioners (e.g., the Evidence-Based Policing Matrix) but also training officers how to recognize good research and use it to their advantage.”(3)
In conclusion, it is my opinion that most police practitioners typically do not have the time, nor
interest in sifting through lengthy academic articles full of such things as regression analysis.
Ultimately, they want to know how the data makes them more effective and how it can be interpreted in context. However, we still must reframe our measurements and incentivize our outcomes. The adage what gets measured gets done rings even truer today because ultimately many police officers enjoy the autonomy of the profession, but it is my experience that they too are receptive to research with a strong desire to be effective in their crime-fighting efforts as evidence in Madison, Wisconsin.
(1) NIJ (2018). The Law Enforcement Advancing Data and Science Program – A partnership between NIJ and the International Association of Chiefs of Police. Retrieved from https://nij.gov/topics/law-enforcement/Pages/nij-iacp-leads-program.aspx (accessed 28 Mar 2018).
(2) Nelson, C. (2018). From Research to Practice, Two Decades Later: Evidence-Based Policing in Madison, Wisconsin” Translational Criminology, 7-9.
(3) Johnson, J. (2018). P-What? A Practitioner Guide to Understanding P-Values Translational Criminology, 14-15
(4) Johnson, J. (2018). A Hippocratic Oath for policing. Police Foundation https://www.policefoundation.org/a-hippocratic-oath-for-policing/ (accessed 29 Mar 2018)
(5) Lum, C., Koper, C. S., & Telep, C. W. (2011). The evidence-based policing matrix. Journal of Experimental Criminology, 7(1), 3-26 – http://cebcp.org/evidence-based-policing/the-matrix/