Part II :The Validity of Criminal Justice data

The “Ferguson Effect” Series

By Joseph Roy, Graduate Student, and Patrick J. Solar, Ph. D, Assistant Professor in Criminal Justice at the University of Wisconsin-Platteville, Special for  USDR

The most common source of information quoted throughout this debate is the Uniform Crime Report, or UCR. UCR data is collected from police departments across the country and is designed to provide “a reliable set of set of crime statistics for use in law enforcement administration, operation, and management.” The UCR is broken down into two sections, part I crimes which are the violent crimes, and part II crimes which are property  crimes.

The Uniform Crime Report  (UCR)

On its face, the UCR is probably the most complete and accurate set of crime statistics that are available to anyone for free. However, the data contained in the UCR reports only represents those crimes that were reported to police. According to a 2012 report by the Bureau of Justice Statistics, from 2006 to 2010, only 48% of violent crimes were reported to police. The offense, modus operandi, and victims of the other 52% are mostly  unknown.

Instrumentation errors abound in the UCR data as it is voluntarily transmitted by over 18,000 police departments across the country every year. Because the system lacks uniformity in data entry there is a high probability that errors regularly occur such as miscoding crimes, categorization errors, missing crimes and, of course, “cooking the books” to downplay crime increases or enhance crime decreases.  So, the 48% of reported crime statistics the FBI is left to work with may actually be closer to only 30% or 40% of the actual violent crimes  committed.

Those who argue against the “Ferguson Effect” rely heavily on the report issued by the Brennen Center for Justice, which relies totally on incomplete, partial year UCR data. The researchers themselves cited some of these problems in the methodology section of their report. Yet, they still felt comfortable enough to draw a nationwide consensus from the data they were able to locate, which, by their own admission, doesn’t contain the major metropolitan cities of “Jacksonville, San Francisco, Indianapolis, Columbus, Detroit, El Paso, Washington DC, Memphis, Nashville, Las Vegas and  Louisville.”

The sanctity of any data collected at levels lower than the UCR should be questioned even further. Conducting scientific statistical research can be a very costly and time consuming proposition. Therefore, often statistical conclusions are drawn from very small sample sizes such as those seen in the report authored by Pyrooz, Decker, Wolfe, and Shjarback which claims that there is no such thing as the ‘Ferguson Effect.” The report only surveyed 81 cities with populations over 200,000 people, and at that they relied on voluntary questionnaires and data reported on police department websites as their means of data gathering. Of course, when bolstering their belief that such a method is statistically sound, they cite researchers from the other side of the argument such as Rosenfeld. It is not possible to determine whether or not there was a ‘Ferguson Effect’ on the crime rate in all major cities without looking at  them.


Both sides of this debate do seem to agree that there has been in an increase in the number of homicides recorded in major cities over the past year. Where they differ is in the cause for that  increase.

The Crime Rate is Rising

Heather Mac Donald, in her Wall Street Journal article supporting the existence of the Ferguson effect, declined to reference the UCR and instead relied on a report compiled by the Major Cities (Police) Chiefs Association. The Association reports that their data was collected via survey but leaves the details regarding the sample size, margin of error, and collection procedure out of the report. Without that information making any informed decision on the validity of the data is  difficult.

Mac Donald takes the data provided by the survey and correlates it with her belief that police officers have begun to “de-police” certain areas out of fear of reprisal for conducting proactive police work. Mac Donald also quotes Dr. Richard Rosenfeld, a one-time Ferguson effect denier turned semi-supporter, as support for her  theory.

According to Dr. Rosenfeld’s report titled “Documenting and Explaining the 2015 Homicide Rise: Research Directions,” published by the U.S. Department of Justice, the final data required to make the correlation between de-policing and the rise in violent crime has not yet been made available by the FBI. The complete UCR data set for 2015 isn’t expected until Fall  2016.

Crime Is Not Increasing,  Significantly

Those on the other side of the argument tend to cite reports on the crime rate published by organizations like the Brennan Center for Justice and Research. While these reports do concede that, at a minimum, homicides in most major cities did rise in 2015, they provide significantly different rationales as to  why.

The report published by the Brennan Center for Justice, which does cite UCR data as the primary information source, indicates that while homicides did rise in 14 major U.S. cities, it also fell in 11 other similarly sized cities. If one was to accept Mac Donald’s theory that de-policing is occurring, seemingly the Brennan Center’s homicide statistics would support that assumption. However, then there must be some explanation for the decrease in homicides in the other cities. Mac Donald doesn’t address this point. The report goes onto explain that some of those 14 cities which experienced an increase had murder rates so low that the addition of only a few murders would cause large statistical increases in the murder rate.

Beyond just murder, the overall crime rate in 2015 is predicted to be down 22 percent since 2010 and is down 66 percent since 1990. Pyrooz, Decker, Wolfe, and Shjarback take it one step further in their report and highlight the fact that a single year change in a crime rate does not, in and of itself, indicate a  “trend.”

Has crime increased? Statistically, it is hard to determine with precision and scientific certainty at this point.  However, there is abundant reason to believe that the police are not engaging in proactive policing as they once  did.


Dr. Patrick Solar is an Assistant Professor for the online criminal justice master’s degree program at the University of Wisconsin-Platteville with nearly 30 years of service as a police officer. He currently serves on the governing board for the Law Enforcement Accreditation Program (ILEAP), setting the standards for professional law enforcement in  Illinois.

All opinions expressed on USDR are those of the author and not necessarily those of US Daily Review.