For the last 300 years, policing has continued to evolve. We have come a long way from the volunteer constables of colonial times to the widespread reactive and proactive policing methods known today. The invention of call boxes, two-way radios, the 911-system, license plate readers, and body-worn cameras are examples of some of the technological changes along the way.1 Despite our ongoing evolution, policing has always been a human science. Cops with keen instincts have always been held in high esteem – as they should. However, instincts, hunches, traditions, and anecdotal experiences only take us so far. As a police practitioner and academic, I’ve noticed how they even contribute to conflicting beliefs about a data-driven research model called evidence-based policing (EBP).
To further compound the issue, each of the United States’ approximately 18,000 police departments have varied organizational cultures, deployment styles, budgetary constraints, social demographics, relationships with their communities, and educational and training requirements.2 At its core, staffing concerns and day-to-day stresses have a tendency to limit a profession still bound by social science – not hard science.3 In the next stage of our evolution, police have an opportunity to use evidence-based research to improve technology implementation, data collection, and department policy.
The Aversion to Evidence-Based Policing
Police typically do not have the time, or the motivation to sift through data. Front line police need to locate information quickly, and police administrators require quick turnarounds on reports. Data can and should be able to do this for us. However, herein lies the concern. Because of advancements in technology or lack thereof, the incoming data can be overwhelming and limiting all at the same time. Data not properly researched and policies hastily implemented create ineffective policing. The adage “garbage in, garbage out” rings truer today than it ever has.
…Instincts, hunches, traditions, and anecdotal experiences only take us so far.
I think the challenge for all of us who believe in forward-thinking law enforcement methods like evidence-based policing is to resist hastily implemented and carelessly thought out programs just to say we did them.3 These hasty plans cause frustration due in part because of the new perceived fad or new loaded term. As a result, proposed research-based policing and efficient data collection takes the brunt and is occasionally resisted by our front line level officers and sergeants. Frankly, this hasty implementation is in part why many of our “old and new dinosaurs” see research-driven policing as a cult, new fad, or worse “academic BS” to be ignored in a few years. They commonly see it as the “same notes, just different music” from academia or an administrator’s perceived self-serving master’s thesis or capstone project.7 Further, terms such as “procedural justice,” “police legitimacy,” “implicit bias,” “focused deterrence,” “risk terrain modeling,” “transparency,” and “de-escalation,” to name a few, are typically unclear to the average cop.
Some of these same “dinosaurs” or traditionalists long for the days of paper reports and where a policeman’s word was bond. It just seemed to make life more efficient. And, let’s face it – policing a bar fight is still the same today as it was 100 years ago.
In the end, we still must evolve and grow while paying attention to research and efficient data collection. Empowering our members to be adaptive in the often complex environment of policing is critical. It seems long gone are the days where citizens trust what we say just because we said it. They desire video recordings and evidence. Popular television shows such as CSI, Law and Order, and NCIS have created environments where cases can be solved in 30 minutes with forensics. There are countless videos in the YouTube era that seemingly discredit the profession. Both have affected courtrooms and American policing.
Challenges of Applying Evidence-Based Research
Due to our focus on evolving and enhancing “change,” we have a real opportunity to influence our millennial generation that are coming into this profession with positive ideals and are usually more driven in the early stages of their career. The 3-5 year cop is more apt to embrace technology considerably more than that 15-20 year cop who may see the increased technology as an increased workload stressor. Again, the challenge is that not all agencies are created the same. There are variations in culture, training, and hiring standards throughout the country.2 However, if we have research that shows something works and is vetted through sound experimentation such as randomized controlled trials, then why wouldn’t we pay attention to the data?3 The problem in my view is that the data is not easily interpreted, let alone retrieved, and when it is, some academics or others tend to proclaim blanket fixes to problems not easily applied to every jurisdiction. That is the problem: policing in New Orleans, Louisiana is much different than policing in Vallejo, California, as is policing in Sausalito, California.
We must start small and evaluate data efficiently while embracing change incrementally.
How we interpret data is also going to vary widely. For example, we have a tendency to claim credit for crime drops, but we tend to make excuses for high crime when the data shows an increase. Some claimed the great crime drop of the 90’s was due to such factors as a robust economy, mass incarceration, and baby boomers aging out of crime.4 To others, these might be examples of improperly evaluated data based on variables that were not clearly connected. Crime may have just dropped because it had rose so high in the late 1980’s due to the crack cocaine epidemic that it then regressed to the mean.8
Further exacerbating the issues are those that dive too deeply in the minutia of a single field, such as predictive policing, procedural justice, or implicit bias research. They have a tendency to lock themselves into “make believe expertise” where real world application is neither applicable nor typically based on common sense.
In policing, many times we attack what we don’t know or understand. For example, try asking a line-level cop if he/she has ever heard of a randomized control trial , let alone some of our police administrators – but that thinking is okay. We must start small and evaluate data efficiently while embracing change incrementally. It took the medical association well over a thousand years to evolve from bloodletting to embracing clinical trials.6 In an era of 140 characters and instant “Google” knowledge, some academics might be surprised to learn that most police don’t have the time, or the interest to read lengthy academic journals occasionally full of statistical regression analysis. However, we might be surprised to find that when people embrace evidence-based policing that they might become more adaptable and empowered.
Big Data’s Unanswered Questions
The same goes for the incident data officers interact with daily. Police want to be able to put keywords in a box and instantly come up with data – people associated with residences, associated cars, girlfriends and other associates to suspects, mobile phones, GPS hits on license plates, contact locations, arrests, guns registered to and even facial recognition via video surveillance (I know, a futuristic idea) are just a few of the dataset expectations. This efficient data collection is important to cops, and they want it quickly and accurately without having to navigate various windows. Police officers want to be able to put a name of a suspect in a field and come up with the data – even cross-jurisdictionally – and to do it without navigating complicated windows or data boxes requiring “equal to” or “matching this.” Initial data retrievals should start big, be efficiently queried, and have the ability to be drilled down to further detail. Our police especially don’t want to deal with multiple names for the same person, which is typical in systems that require competent initial data entry or a keen resistance to garbage in/garbage out data entry.
Police typically do not have the time, or the motivation to sift through data.
But with big data and the affinity for visual charts and statistics, we run the risk of creating environments where the data is never enough. The risk is that we become enamored with statistics and interpret the data the way we want to see it. With increased data, however, we create more questions than are accurately answered, along with a continued and even bigger thirst for instant knowledge and statistical interpretation. The concern is that many of these numbers fluctuate so much based on the common variables.
While embracing data is critical to molding our police cultures toward research, data must be targeted, tested, tracked and have the ability to be replicated. 2 Data without context is just data, but with the research, we can test and contextualize it.5 Evidence-based policing challenges what we think we already know while empowering our police to be more effective. It should be our hope that one day police data collection and retrieval through research advances to the point where we are able to efficiently and correctly measure it. How can we advance and improve the policing profession if we are not accurately measuring outcomes such as data on national police use of force usage – among other things?
(originally published as a Mark43 blog at: https://www.mark43.com
1.Lawrence, M. (2015). Lights, Camera, Action: The Age of Body Cameras in Law Enforcement and the Effects of Implementing Body Camera Programs in Rural Communities. NDL Rev., 91, 611.
2 Potts, J. (2017, April). Translational Criminology. Police Research on the Front Lines. Retrieved from: http://cebcp.org/wp-content/TCmagazine/TC12-Spring2017
3 Constable, N. (2017, April 22). A Course of Leeches. Retrieved from: https://nathanconstable.wordpress.com/2017/04/22/a-course-of-leeches/
4 Ford, M. (2016, April 15). The Atlantic. What Caused the Crime Decline in the US. Retrieved from: https://www.theatlantic.com/politics/archive/2016/04/what-caused-the-crime-decline/477408/
5 Wilkinson, D. (2017, April). The Oxford Review. What’s the Difference Between Data and Evidence? Evidence-Based Practice. Retrieved from https://www.oxford-review.com/data-v-evidence/
6 Syed, M. (2015). Black Box Thinking: Why Most People Never Learn from Their Mistakes–But Some Do. Penguin.
7 Davis, Ron. (2015, October 24).Oral Presentation. Improving Community Relations: Evidence on Police Legitimacy and Procedural Justice. International Association of Chiefs of Police Conference – Chicago
8 Farrington, D. P., & Welsh, B. C. (2006). How Important is “Regression to the Mean” in Area-Based Crime Prevention Research? Crime Prevention & Community Safety, 8(1), 50-60.