Skip to Content

MNPD's Response to Gideon's Army Report

3/7/2017

Regarding RESOLUTION NO. RS2016-459, requesting response from the MNPD regarding findings in a recent report regarding MNPD traffic stop statistics in Nashville, I provide the following information, comment and context.

Previous correspondence copied to you and the presentation by Commander Terrence Graves at the Council hearing on this matter has made it clear that we take exception to the conclusions drawn in this report. Most egregious is the accusation of racial profiling. I categorically deny racial profiling is an element of any MNPD policing strategies and would undertake appropriate action to remove any officer that is determined to have engaged in such conduct.

Although the term ‘racial profiling’ used more than twenty (20) times throughout the report, there is no documented account of any such incident or incidents. It seems apparent that none of these persons described in the report have contacted the Office of Professional Responsibility about their allegations so that a thorough investigation can be conducted. If true, I would encourage you to facilitate that; for if racial profiling is occurring, I assure you that we want to investigate it.

The assumptions made in the report rely solely on the disparity in the African American census data compared to the vehicle stop data. As to this disparity, a competent and responsible researcher and statistician can reveal, simply by the numbers, that there is in fact a disparity. Those same people will tell you that, alone, the numbers cannot tell you the reasons, good or bad, that the disparity exists. Those reasons can be revealed, or at least hypothesized, only with additional, exhaustive research and analysis. That was not done in this situation.

Conveniently, the report fails to offer any accepted definition of racial profiling. As in any discussion, it is most helpful to know the definition of a term before making a determination as to whether it exists.

  • A commonly accepted definition of racial profiling, or bias based policing, the selection of individuals for enforcement intervention based solely on a common trait of a group, such as race, ethnic origin, gender, socioeconomic status, sexual orientation, or age.
  • Or, the use of race or ethnicity as grounds for suspecting someone of having committed an offense.
  • The American Civil Liberties Union has defined racial profiling as the discriminatory practice by law enforcement officials of targeting individuals for suspicion of crime based on the individual’s race, ethnicity, religion or national origin.
  • The National Institute of Justice has stated that racial profiling by law enforcement is commonly defined as a practice that targets people for suspicion of crime based on their race, ethnicity, religion or national origin.

Again, this report contained no documented account of any such incident or incidents. The assumption relies solely on the disparity in the African American census data compared to the vehicle stop data. Any such disparity does not constitute racial profiling and responsible researchers would also advise that disparity alone is not evidence of bias.

The fallacy of the entire report is the utilization of census data to support the allegations of racial profiling.

It has been long settled in the academic world that census data is not an appropriate or meaningful benchmark to support or refute any allegation of racial profiling. To repeat--Census data is not an appropriate benchmark. While census data has its purpose, it is only a snapshot, as best as can be determined, of where people live. Census data does not take into consideration where people drive their vehicles. As you know, Nashville is a vibrant community, surrounded by not only counties who provide Nashville with a workforce, but also is home to over 20 colleges, universities, technological schools and institutes of higher learning-of these, we are proud to be home to at least 4 colleges that would be considered historically black colleges and universities. The US Census Bureau struggles to answer the question where these college students are counted-with somewhere between 40-80,000 college students in Nashville-what effect does that have on Census accuracy; while clearly having an effect on who drives in Nashville. Similarly, an often unknown element of the census is prison populations. Clearly, with Nashville home to several large state prison facilities and private prison facilities, the population and demographics of the prison population clearly have an effect on Census data yet demonstrate an inherent flaw in using Census data to analyze who drives in the same city.

Although the report attempts to refine its analysis by utilizing census block and census tract comparisons, as opposed to overall city or county census comparisons, the same fallacy exists. People do not drive round and round limiting their driving only to their census block or census tract. In fact most people have no idea where their census block or tract ends or begins. People travel broadly across the city and the county, to school, run errands, to work, to shop, to visit family and friends. With the convergence of three US interstates, even more, possibly hundreds of thousands-if not more, pass through Nashville in route to Atlanta, Birmingham, Memphis, Knoxville and even further. Often they exit to get gas, sight see, shop and then move on. Clearly these people make up the driving population but are never considered in Census data.

Any attempt to understand why the existing and established research by credentialed statisticians and scholars was ignored would only be speculation as this information is readily available via the internet. For example, see:

Grogger, J., & Ridgeway, G. (2006). Testing for racial profiling in traffic stops from behind a veil of darkness. Journal of the American Statistical Association, 101(475), 878-887. “The key problem in testing for racial profiling in traffic stops is estimating the risk set, or “benchmark,” against which to compare the race distribution of stopped drivers. To date, the two most common approaches have been to use residential population data or to conduct traffic surveys in which observers tally the race distribution of drivers at a certain location. It is widely recognized that residential population data provide poor estimates of the population at risk of a traffic stop; at the same time, traffic surveys have limitations and are more costly to carry out than the alternative that we propose herein.” [Emphasis added.]

Note: Grogger and Ridgeway were awarded the Outstanding Statistical Application Award by the American Statistical Association for their work in formulating an alternative to the use of statistical data.

Johnson, Richard R. Ph.d. Biased-Based Policing Reports Are Failing the Police and the Community, Why Agencies Need to Stop Using Census Data. Many have used Census data as their benchmark for police activity because of its ease of access. The problem, however, is that the demographic characteristics of the people living at any one location have nothing to do with the driving population there, nor who is breaking the law in any specific area. We use or vehicle to travel to places away from our homes, as people generally do not work, shop, or recreate in their homes.” [Emphasis in original.]

Ridgeway, Greg and John MacDonald. Methods for Assessing Racially Biased Policing. Santa Monica, CA: RAND Corporation, 2010. http://www.rand.org/pubs/reprints/RP1427.html. “The primary reason for using US Census data to form the benchmark is that it is inexpensive, quick, and readily available. A number of studies attempting to assess racial bias in police behavior use population data from the census, some rely on estimates at local area levels like neighborhood census tracts (see Parker and Stults in this volume). However, for the reasons previously listed, benchmarking with census data does not help us isolate the effect of racial bias from differential exposure and differential offending. Even refinements to the residential census, such as focusing on subpopulations likeliest to be involved in crime (e.g., men or driving age young adults) are not likely to eliminate differences in the exposure of officers to criminal suspects or provide a good approximation of the population at risk for official police action. Fridell summarized the problem with using the census as a benchmark with regard to offender exposure by noting that, “this method does not address the alternative hypothesis that racial/ethnic groups are not equivalent in the nature and extent of their . . . law-violating behavior. Census estimates provide only the racial distribution of residents and not how these numbers vary by time of day, business attractors such as shopping centers, daily traffic patterns involving commuters, etc.”

Ridgeway, Greg and John MacDonald. Methods for Assessing Racially Biased Policing. Santa Monica, CA: RAND Corporation, 2010. http://www.rand.org/pubs/reprints/RP1427.html. “There is a compulsion in media reports on racial disparities in police stops to compare the racial distribution of the stops to the racial distribution for the community’s population as estimated by the US Census. For example, in 2006 in New York City, 53% of stops police made of pedestrians involved black pedestrians while according to the US Census they comprise only 24% of the city’s residential population. When the two racial distributions do not align, and they seem to do so rarely, such statistics promote the conclusion that there is evidence of racial bias in police decision making. Racial bias could be a factor in generating such disparities, but a basic introductory research methods course in the social sciences would argue that other explanations may be contributing factors. For example, differences by race in the exposure to the police and/or the rates of committing offenses may also contribute to racial disparities in police stop decisions. It is well documented, for example, that due to historical differences in racial segregation, housing tenure, poverty, and other sociopolitical factors minorities in the US are more likely to live in neighborhoods with higher rates of crime and disorder.”

Ridgeway, Greg and John MacDonald. Methods for Assessing Racially Biased Policing. Santa Monica, CA: RAND Corporation, 2010. http://www.rand.org/pubs/reprints/RP1427.html. “Police deployment in many cities also corresponds to differences in the demand for police services. Neighborhoods with higher volumes of calls to the police service typically have a higher presence of police. Additionally, research indicates that racial minorities, and in particular blacks, are disproportionately involved in serious personal offenses as both victims and offenders.”

Ridgeway, Greg and John MacDonald. Methods for Assessing Racially Biased Policing. Santa Monica, CA: RAND Corporation, 2010. http://www.rand.org/pubs/reprints/RP1427.html. “The crux of the external benchmarking analysis is to develop a benchmark that estimates the racial distribution of the individuals who would be stopped if the police were racially unbiased and then comparing that benchmark to the observed racial distribution of stopped citizens. The external benchmark can be thought of as the population at risk for official police contact. As we will see, estimating the appropriate population at risk is complicated. Crude approximations of the population at risk for police contact are poor substitutes and can hide evidence of racial bias or lead to exaggerated estimates of racial bias. The racial composition of the stops made by the police involves some combination of police exposure to offending/suspicious activity, the racial distribution of the population involved in those activities, and the potential for racial bias. To provide some context, we use some hypothetical numbers and consider an unbiased officer on a foot post who makes stops only when a pedestrian matches a known suspect description. This officer works in a precinct with 40 blacks matching suspect descriptions and 40 whites matching suspect descriptions. If we could somehow measure such numbers we would be inclined to propose a suspect-description benchmark of 50% black and 50% white. However, if the routine daily activities of whites and blacks differ then the officer will encounter different proportions of suspects by race. Say, for example, that the majority of the 40 white suspects stay inside most of the day, travel only by car, or avoid the specific areas with high police presence, then this officer will stop only a small number of white suspects, deviating substantially from the 50 percent benchmark. Even the less extreme situation, in which half of the white suspects are exposed to the officer, results in the officer stopping blacks in 67 percent of all of their stops decisions. The suspect benchmark in this context is only valid if the police are equally exposed to suspects from the various racial groups. Therefore, even with unbiased officers, we cannot necessarily expect what seems like a reasonable external benchmark to match the racial distribution of stops. This example effectively demonstrates that any of the external benchmarks described in this section must be viewed with caution.”

Fridell, Lorie A. By The Numbers, A Guide for Analyzing Race Data From Vehicle Stops, A report prepared for the Police Executive Research Forum, funded by the Department of Justice Office of Community Oriented Policing Services. “In census benchmarking agencies compare the demographic profile of the drivers stopped by police to the demographic profile of the residents of the jurisdiction as determined by the U.S. Census. For a variety of reasons, such a comparison is of no scientific value for purposes of trying to measure racial bias in policing and, in fact has in misleading and often resulted unsupported findings.” [Emphasis added.]

Dr. Lorie Fridell is a national expert on biased policing, the author of numerous writings on this subject and developed the “Fair and Impartial Policing” model for law enforcement instruction.

It should also be noted that the use of census data does not in any way take into account the deployment of resources by the MNPD. While the report does note, using census data comparisons, an approximate 10 percent disparity, the report does not in any way take into account other disparities that guide the deployment of these resources. For example, and using the same census comparisons:

  • African American gunshot victims—a 54 percent disparity.
  • African American homicide victims—a 48 percent disparity.
  • African American aggravated assault victims—a 29 percent disparity.
  • African American carjacking victims—a 23 percent disparity.

Ignoring these real disparities of victimization in our African American communities by seeking to divide or drive a wedge into police community relations by attempting to draw attention to a false narrative of racial profiling, without clear evidence, is morally disingenuous. It ignores the practical realities of modern community based policing strategies and tactics-particularly the deployment of police resources, to these often disadvantaged neighborhoods, based upon victimization and calls for service from those actually living in these neighborhoods.

As Commander Graves stated in his presentation to the Council, these are the facts and figures used by each precinct commander to deploy resources to particular areas of Nashville where persons are most likely to be the victims of violent criminal activity and to respond to the requests for increased police presence by those living in these communities which are punctuated with violence daily. In my time as Chief, in hundreds of community meetings, I can never recall the residents of a neighborhood requesting LESS police presence. I am certain that the eight precinct commanders, who meet regularly with residents and neighborhood watch groups, would share the same experience.

Obviously, as the report seemingly chooses to disregard, the greater concentration of MNPD resources in a particular area will necessarily result in more vehicle stops than in areas wherein MNPD resources are deployed in a more limited manner. Thus, if more police resources are dedicated to areas with a disparity in the demographics of resident victims, logic dictates that there will be increased police contacts with those in that demographic. This perceived disparity is not bias. It is community based policing designed to reduce victimization and concentrate resources to the area of greatest need.

Previously, we supplied density maps to the Council to illustrate where the concentration of victims is the greatest and where our resources are deployed. These maps are attached, again, for your review to aid you in understanding this concept. Also, as previously supplied, attached is a chart showing the disparity African Americans experience as victims of violent crimes. This chart, in graphic detail, describes the disparity we all should be discussing-how to reduce the victimization of our African American communities.

We will continue to engage in deliberate, thoughtful, well-planned, and lawful policing practices. On occasion, we may fall short, but more often than not, the men and women of the MNPD-in responding to over 220,000 calls for service and over 640,000 additional contacts do get it right.

It remains my commitment, as Chief of Police, that the MNPD will continue to allow precinct commanders the discretion to assign police resources to those areas afflicted by violence and to provide police services to areas where crime and violence are the highest, where people are at greater risk of becoming a victim, and closest to the location of where calls for police services are located. They will continue to attend community meetings, host neighborhood events, sponsor activities to engage at risk youth and participate in other activities to enhance our ability to better serve and connect with the community.

I understand and respect your personal urgency to pass this resolution. It does appear, however, that over these months all has been said that can be said. Certainly, while I see no need for passage of the resolution, if passed we will create the requested response, however, what has been stated above and in other correspondence and presentations, would constitute the response.