Application of Multi-Disciplinary Analytics in the Selection of Police Officers

Share:

California law enforcement hiring seems to cyclically follow a hiring trend that is akin to its weather, “drought” or “torrential flooding rains.” Currently, the police are experiencing the “rain” mode, desperately trying to fill the empty basins created from the drought of the recessionary years of 2006 and 2009 when severe economic downturns caused departments to shrink their staffing. In the heady 1990s, COPS hiring initiatives put more than 100,000 cops out onto the streets (Kessler, 2014) to fight the challenges of the “crack cocaine” epidemic and rising crime throughout the nation. More cops, stricter sentencing, and increasing prison populations were all contributing factors to cause crime to recede (Levitt, 2004). Those cops hired twenty years ago have greyed, and are preparing to retire as current generations are poised to take their turn.

Dear New Police Officer
(Flicker)

A priority for policing has now become to hire the numbers to fill necessary staffing needs. The pressure to fill vacancies presents the temptation to engage in “muni-shuffling,” or accepting problematic officers who left their former agency after facing allegations or sustained behaviors of misconduct. Lateraling officers can bring a boon of experience and knowledge, but also undesirable challenges that an agency might be willing to overlook (Cops, 2017). For both experienced officers and recruits, though, the environment within which they will work is changing dramatically from the past.

Newly hired officers today will be working in an advanced, high-tech environment of autonomous vehicles, DNA and facial recognition, robotic assistance, artificial intelligence, and cybercrime. Given the nature of high-tech policing in the future, perhaps it is time throughout our decades-old, multi-pronged approach of oral boards, written tests, and background investigations. Instead, there is an emerging analytics process that can successfully assess candidates for their character and ethical standards, a process that can be tailored to each department’s needs. Perhaps it is time for policing to look at analytics as a way to address their hiring woes and begin to build for the future.

A Different Way

California law enforcement professionals know that an incident in one jurisdiction has a rapidly resonating effect throughout the state. Certainly, enduring a “Rampart” or “Ferguson” type incident would evaporate any hard-earned community capital and trust. Today, policing is judged at the speed of social media. An adverse police incident is costly, and often results in the loss of human life or decisions which result in enormous sums of money paid to litigants and their attorneys. Who pays? The taxpayer of course. Instead of continuing to pay out settlements, creating a hiring system that weeds out those who might be likely to engage in behaviors that would result in litigation is the key to ensure the fiscal survival of policing. Every police agency in the law enforcement community wants to hire individuals who are ethically sound, dedicated in service, courageous, and practice rational decision making. To make better decisions to hire those types of candidates, two fundamental questions emerge: are we picking well? Could we pick better?

The Costs & Alternatives

The current California police officer’s minimum hiring standards can be traced back over 50 years ago to the 1967 President’s Commission on Law Enforcement and Administration of Justice (President’s Commission Report), and the 1973 National Advisory Commission on Criminal Justice reports which led to the creation of California GC 1031 and POST regulations 11 CCR 1950 (Standards, Spilberg, & Corey, 2018). Once met, 11 CCR 1950 allows, “the adoption of more rigorous requirements, higher standards, additional assessments and more in-depth evaluations than those stated in these regulations is at the discretion of the employing department (California Code of Regulations).” At their discretion, police departments can surpass the POST minimum hiring standards. Even though there is great latitude in the ways police officers can be hired, most use a process that is quite similar from jurisdiction to jurisdiction.

Todays’ process to hire police officers is laborious and costly. It usually involves a multi-step battery of testing, background investigation, a polygraph, a psychological examination, and then a medical exam. Oral boards could be susceptible to appearance biases, marginalizing other such desirable competencies as reasoning and writing skills, or deselecting candidates who were not visually appealing (Mahajan, 2007). Instead of utilizing the “hitting” and sometimes “missing” style of selecting and hiring police officers, could traditional biases be eliminated from the equation?

Law enforcement leaders need to consider what their future officer corps composition will be, and the landscape they will face. The traditional crimes of old will not go away; they may, though, appear in new guises and frontiers as technology and social media allow people to connect in new ways. It is time to use some of the advances in technology to select those who want to enter the police profession. By employing multi-disciplinary analytics, we can efficiently and effectively improve the selection of future officers.

Multi-Disciplinary Analytics
(Flicker)

Analytics Instead of Instinct

There is value in the use of analytics, since it can be designed to look at the attributes of the candidate, forgoing the “thin-sliced” perceptions of candidates with seemingly stellar credentials and great loquacity (Gladwell, 2005). The current hiring process is unable to effectively “peel back” the veneer of candidate’s exterior to identify strengths and deficiencies they might possess. As a diagnostic tool, though, analytics gathers the individual’s data to develop patterns of behavior.

Multi-disciplinary analytics incorporates personal data from current and previous background investigations, the results of polygraph exams, psychological fitness batteries, and from financial and social networking applications. With quick computer processing, large amounts of the candidate’s data can be reviewed and collated to create a suitability profile and an analysis of their likely future performance. The agency would already have created a profile of successful officers to use as a template from which to compare. This template would include parameters to assess ethical characteristics, a willingness to assume responsibility, and a demonstrated capacity to consistently perform over a long duration.

Due to these demanding requirements, just one type of analytics might not achieve a full candidate discovery. Rather, the necessity to amalgamate other analytics such as behavior, learning, and diagnostic dimensions would be included in any profile. For instance, the banking and financial firm Well Fargo uses analytics to hire employees, not only to fulfill their performance needs but also to fit within their corporate culture (Kuehner-Hebert, 2013). They utilize biometric data to determine accurately how long a person has been in a job, how positions are available, and the highest level of education. They then use an online 65-question assessment instrument designed by banking industry experts. This instrument studied a cross-section of company culture and interviewed protected classes.

early warning program
(U.S. Army National Guard photo by Spc. Jason Dorsey, Illinois National Guard Public Affairs)

The online questionnaire has been extremely useful to sift through a large number of job seekers and in a short amount of time, finding those good company “fits.” Over six months, Wells Fargo experienced a 15% teller and 12% personal banker retention rate of their newly hired personnel. This process also identified “success indicators” such as the combination of tellers who have experiences in financial services, retail telecommunications, hospitality, and demonstrated strong academic performance in high school and other education (Kuehner-Hebert, 2013)

“Predictive analytics identifies, in detail, the behavioral talents that fundamentally drive performance,” said Hugh Massie, founder and President of Atlanta-based DNA Behavior International (Carloza, 2018). “This enables the recruiter to consider fit for the role, fit to line management, fit under pressure, and gaps that need to be addressed with training” (Carloza, 2018). For police agencies, there could be a shortened version of questions regarding a candidate’s integrity, decision making, technical, cultural diversity, interpersonal attributes, and communication skills. Since this would be a relatively new component to the hiring process, a department working in conjunction with their city’s human resources department and an analytics vendor would have to create a template of attributes sought by the police. This preliminary process could be completed online and submitted to an agency’s HR for evaluation and determination candidate suitability. If a candidate matches the sought profile, then HR would invite the candidate for a more comprehensive examination.

retention problems
Boston Police Department recruit class 55. (Photo courtesy bpdnews.com.)

The Analytic Process

A process using a battery of analytic assessments would be computer-based and interactive. The candidate would complete a pre-screening questionnaire process to determine their suitability. This would include a DOJ portion to check their criminal and driver histories. If the candidate’s personality patterns are acceptable, then the next phase would be onto the polygraph and psychological testing again putting all the developed data together. At this preliminary stage of recruiting and hiring, departments can also pass on those candidates who have predispositions to any moral or ethical challenges, absenteeism, inappropriate use of force applications, and low work productivity. Instead, candidates possessing desired attributes of high moral and ethical standards, strong deductive and reasoning competencies, creative traits, and community minded.

A customized approach to the specificity in candidate selection can be applied from the design by city, police, and community stakeholders. A department would have to determine what those necessary officer traits should encompass possibly incorporating employee evaluations, desired parameters of typical police officer behaviors, and community input. As such assessments come online, it is useful to consider how significantly it would change the process for a majority of police departments in the State. Let’s look at the transition to analytics in a small community in the near future.

A Case Study

In Sanger, CA, a community of 26,000, the Police Department struggles to fill its ranks, competing with other agencies in the region for quality candidates. The City is too small to have a dedicated team of HR specialists in either its Police Department. The City HR department often moves too slowly to find and hire the right people before larger agencies beat them to it. Rather than resorting the way it has always been (inefficiently) done, Sanger’s use of multi-disciplinary analytics would begin when the candidate submits their online application, and also complete a questionnaire for assessment (a panel comprised from the community, HR, and executive staff would have created the desired profiles). Due to the size of the department, it would require consulting with vendors that have done comparable work in the industry. If the candidate has the desired qualities, they would be invited to take a more in-depth and comprehensive questionnaire which would determine if the candidate has the “Sanger” fit. Analytics becomes the equalizer, finding those candidates that fit the Department’s and community’s needs.

Those police officer candidates that score high enough in the initial variegated examination process would then continue to the background and polygraph portion, depending on their viability they would proceed to the psychological and medical evaluations. Fortunately, the department’s investment is quite low if the candidate should fail at any level. Only if a candidate proceeds to hire, the costs (such as outfitting of equipment and field training) would begin on the journey to their becoming a full-fledged police officer. Instead of “barrier costs” at the outset, Sanger’s sunk costs are low for candidates who may or may not succeed in the field training process or their first formative “rookie” years as a neophyte officer.

Starting from Today

In the future, as it is today, the Sanger Police Department’s goal is to improve the selection process through an in-depth assessment. In the future, though, it would utilize the full potential of analytics. Using applied multi-disciplinary analytics offers desired profiles, and more accurate assessments of candidates, by shifting and digesting great amounts of information related to their character and effectively locating a candidate with the desired profile who will be the right “fit” for the department’s requirements. With the proper selection of performance metrics for both the department and the candidate, an algorithmic model can be established (Danieli, 2016).

In the initial recruiting process, human resource analytics start to shine. Analytics can offer a department, quicker “time to hire” features. If a candidate is not a good fit, they can pass and focus on the ones that will fit the need of the department, or speed up the process after one has been identified so that the vacancy can be secured. Second, if departments are locating those particular candidates, the cost will be reduced in the recruiting and hiring process. Candidate expenditures can be determined earlier in the process since the randomness has been eliminated or greatly diminished (Van Vulpen, 2016).

Conclusion

Every community has different demands and expectations for their law enforcement agencies; this will be a customized application that can be administered in the primary analytics metrics. There are tremendous benefits and dividends to be had by having a department that is genuinely synched with the community or county that it serves, bolstering trust and reducing those embarrassing and costly situations that are caused by misaligned officers. Analytics as a tool can provide for more exact character profiles, which will allow for specific searching of particular behavior models that will make for a competent and diverse department, ever evolving to meet the requirements of not only the department but also the community.

Multi-disciplinary analytics will broadly assess the candidate in a variety of areas such as their past finances, traffic and criminal histories, social media interactions, results from their background, polygraph, and psychological testing. Analytics can minimize bias and select candidates through algorithmic flow, the parameters having been formulated to achieve the required expectations. The use of analytics has a proven application across many disciplines and could be an applicable tool in selection of cops, as computer technology improves, departments will have the ability to rapidly process and analyze a candidate’s biographic information, coupled with integrating various other selection facets can seek and eventually select candidates that will meet the suitability of a community’s future law enforcement challenges.

———

REFERENCES

Carloza, L. (2018). Putting predictive analytics to work in hiring. Journal of Accountancy.

Cops, P. /. (2017). Hiring for the 21st Century, Law Enforcement Officer. Perf / Cops.

Danieli, O. (2016, October 17). How to Hire with Algorithms. Retrieved from To all recruiters — use machine learning to hire better candidates. 18 June 2016, medium.com/@deadlocked_d/to-all-recruiters-use-machine-learning-to-hire-better-candidates-c5aad22f3319.

Gladwell, M. (2005). Blink. New York: Hachette Book Group.

Kessler, G. (2014, September 26). The Washington Post. Retrieved from The Washington Post: https://www.washingtonpost.com/news/fact-checker/wp/2014/09/26/bill-clintons-claim-that-100000-cops-sent-the-crime-rate-way-down/

Kuehner-Hebert, K. (2013, September 6). Predictive Analytics for Hiring. Retrieved from www.bai.org: https://www.bai.org/banking-strategies/article-detail/predictive-analytics-for-hiring

Levitt, S. D. (2004). Understanding Why Crime Fell in the 1990s: Four Factors that Explain the Decline and Six that Do Not. Retrieved from https://pubs.aeaweb.org/doi/pdf/10.1257/089533004773563485

Mahajan, R. (2007). The Naked Truth: Appearance Discrimination, Employment, and the Law. Asian American Law Journal. Retrieved from https://scholarship.law.berkeley.edu/aalj?utm_source=scholarship.law.berkeley.edu%2Faalj%2Fvol14%2Fiss1%2F6&utm_medium=PDF&utm_campaign=PDFCoverPages.

Training, C. C., Spilberg, S. W., & Corey, D. M. (2018). Peace Officer Psychological Screen Manual. California Commission on Peace Officers Standards and Training.

Van Vulpen, E. (2016). Predictive Analytics in Human Resources: Tutorial and 7 case studies. Retrieved from Predictive Analytics in Human Resources: Tutorial and 7 case studies. 2016, www.analyticsinhr.com/blog/predictive-analytics-human-resources/.

———

Kent Matsuzaki is a captain with the Sanger Police Department (Central Valley, California). He is a member of California POST Command College, Class #64. He currently has 23 years of law enforcement experience and oversees Operations. He has worked patrol, investigations, and K-9. For the last 19 years he as taught as an instructor at the Fresno Police Academy. He graduated from CSUF Fresno with a bachelor’s degree in history.

Share:
Submit a Correction
Related Posts