The following are remarks by James K. “Chips” Stewart, director of public safety and senior fellow for law enforcement at CNA Corporation and retired director of the National Institute of Justice. These remarks are also featured in the Fall 2017 issue of Translational Criminology Magazine.
During the turbulent 1960s when police were at the epicenter of the civil rights movement, antiwar protests, and increasing violent crime, I made a decision to leave graduate school and join the Oakland Police Department as an act of social conscience. Policing needed good persons to be police officers in order to give our institutions time to catch up with the massive social changes underway that were stressing our communities.
What did the education of a police officer look like during these challenging times? The basic police training included an experiential and traditional method that, in the opinion of police leaders, “works best.” But in this education there was little of what August Vollmer had called for in the 1920s regarding “scientific policing.” However, at the same time, there was new focus on Supreme Court rules for admissible evidence (Mapp vs. Ohio, Escobedo, and Miranda) and the exclusionary rule.
Also during that era was the release of the report from President Johnson’s Commission on Law Enforcement and Administration of Justice, titled The Challenge of Crime in a Free Society (1967). This nine-volume commission report was a major contribution to our knowledge about the criminal justice system and also created the National Institute of Justice (NIJ). The NIJ was charged with using science to help inform state and local police and policy officials about what works in responding to crime. While this groundbreaking Commission Report was a catalyst for change in law enforcement, there was no basic police curriculum developed to implement the findings and recommendations.
When my nomination to be the director of NIJ was announced by the president in October 1982, there was a collective gasp from the academic community. The concern was that my experience was limited to policing operations rather than the methodologies of social science and the theories of social control and deterrence. My appointment also occurred shortly after Robert Martinson and his colleagues had reached their notorious conclusion that “nothing works” in social science research reporting on police and corrections. An additional review of much of the prior NIJ research revealed mostly correlational analysis, which was regarded by leading scholars (e.g., James Q. Wilson, Al Reiss, and others) as insufficient to suggest that something worked or didn’t work. Police leaders were skeptical of the reported scholarly findings and became very cautious about implementing policies and practices based exclusively on correlational and statistical comparison.
Further, in 1978 the National Academy of Sciences published another influential report, Deterrence and Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates. It provided some informed recommendations to me as the newly appointed director. This report called for more rigorous assessment by NIJ of policies and practices based on social control theories and use of deterrence for crime control. In light of these conclusions the attorney general, William French Smith, charged me to focus NIJ research on advising the police about “what works.”
One of my first tasks was to try and ascertain what constituted good evidence that a specific police intervention actually had an impact on deterring and reducing crime and recidivism. At the time I consulted with the criminological giants—James Q. Wilson, Al Blumstein, Novell Morris, Al Reiss, and others. Wilson and Reiss were adamant that NIJ had invested too many resources in correlational research that could not speak to causation. Only experimental research could reach the level of confidence that would produce scientifically verifiable research and reliable impact evaluations.
While I had several social scientists who were interested in doing experimental research, a bigger challenge was finding police departments that were willing to participate in a rigorous experiment where the treatment and control would be randomly assigned. This required police to fundamentally change their thinking about public safety and the consequences of random assignment. The police were skeptical of any research because of the highly negative studies produced in the late 70s.
I attended and met with the executive boards at national meetings of police associations, especially ones interested in research like the Police Executive Research Forum and the International Association of Chiefs of Police. At these meetings I recruited a couple of courageous police chiefs who were willing to subject their agencies to random assignment to advance our knowledge of what works.
The most serious concerns facing these chiefs at the time were increasing homicide rates and low clearances. Homicides were generally thought to be random events largely out of the influence of police discretion and prevention efforts. A 1976 Police Foundation study had reported that police had been called at least once (and usually more times) to intervene in a domestic dispute (family fights) in the previous two years in 85 percent of spousal homicides. As a police officer I knew that domestic disputes were some of the most difficult police calls and that violence was always a risk. Since spousal homicides represented about 25 percent of all homicides, better interventions were seen as a way to make a significant reduction in homicides. Thus, in 1981 NIJ supported Larry Sherman’s Minneapolis Domestic Violence Experiment to evaluate the effectiveness of three types of police response to domestic violence—arrest, counsel, or separation. These three responses were randomly allocated (with some exceptions) to each domestic violence situation police were called to. The results of the study showed a strong deterrent effect for arrest. These findings changed police practice; in many states police departments enacted policies for mandatory arrests without a warrant for domestic violence where the responding officer had probable cause to believe a crime had occurred.
This experiment was historic for the police and for social science, showing that interventions could work and were not always negative toward the police. The only regret I have was that many social scientists insisted that I replicate this study in several jurisdictions. The results were confusing and varied across every location. Policy makers and the police then became concerned that sociological research may not be as reliable as science.
Another example of my work in NIJ was focused on fear of crime. Fear of crime was causing communities to retreat and not help the police solve predatory crimes and open-air drug dealing. Gangs and drug dealers occupied public parks and mothers and babies refused to go into places where violence was visible daily. In 1982 NIJ awarded the Police Foundation $1,830,534 (a significant amount both then and now) to conduct an 18-month experiment in Newark, New Jersey, and Houston, Texas, to implement a randomized controlled trial examining how police may be able to reduce citizens’ fear of crime. Interestingly, in their March 10 and 14 hearings in 1983, Congress objected to this experiment as a waste of taxpayer dollars. The Washington Post ran a major story quoting congressmen criticizing the director’s decisions to invest in experiments “when everybody knows what to do to reduce the fear of crime is arrest the suspects.” I recall one congressman asserting that for a dime he would tell me how to stop fear of crime. Fortunately, the experiment was funded, and also achieved highly confident results. These results lead to the development of NIJ to develop a community policing strategy that several chiefs then implemented.
Finally, under my tenure the NIJ supported the Minneapolis Hot Spots Patrol experiment by Sherman and Weisburd. This hot spots experiment challenged the Kansas City preventative patrol experiment by showing that when police target crime concentrations, they can significantly reduce and deter crime at those locations. The Kansas City preventative patrol study did not use rigorous experimental methods, and instead tested police presence generally, and not where crime was concentrated. This study has become a landmark study for police deployment, showing that when police concentrate their resources where crime occurs, they can make a difference.
Through my experiences, I know that social scientists and police officers can work together to make a difference in the quality of life in our communities. The development of evidence-based policing was really implemented in new ways at the National Institute of Justice. We were inspired by the lack of evidence (Martinson) and the need for better policing methods (President’s Commission on Law Enforcement) that could be justified by science.
I share the Distinguished Achievement Award presented to me by George Mason University’s Center for Evidence-Based Crime Policy with so many police chiefs and scholars who helped me and accepted the risks of pushing experimental research into new areas where more informed policies were and are desperately needed. Today I continue this work with the Bureau of Justice Assistance’s Smart Policing Initiative, which teams police agencies with researchers and other justice-supported programs to make an impact. I applaud the work the CEBCP is doing to promote evidence-based policing and also celebrate the work of one of its professors, Laurie Robinson, for developing Crimesolutions.gov during her tenure at the Department of Justice. Collectively we have made a real difference, although the fight is far from over in pushing for more of the best evidence on “what works” to prevent crime and recidivism.