Policing and Metrics
Just like in education, medicine, and the military statistics can be gamed to make the outcome support the desired conclusion. In recent decades metrics has transformed policing into a numbers game that can also be easily manipulated. It is common knowledge that for data tabulations the results are only as good as the data collected—garbage in, garbage out.
The failure or success of a city may rest on its perception of being a safe place to live and raise a family. Elected officials who can show a continued reduction in crime can usually get re-elected for a second term. For any city when you think of law enforcement the first line of defense is the local police department. In reality it not only involves the police department but the entire judicial system including the district attorney’s office, incarceration facility an available space, code compliance, parole system, economics, city vitality, and obviously the people that make up the community. The community has to do its part in controlling criminal activity by participating in and supporting crime deterrence and prevention programs such as crime watches, volunteers in patrol, neighborhood associations, and town hall meetings. In essence, the residents have to reduce the ease with which crimes are committed. They, in theory, are a vital part of community policing.
Violent crimes in the USA have declined since the 1990 due to changes in policing practices and that major change has involved the use of metrics. Metrics is defined as a standards of measurement by which efficiency, performance, progress, or quality of a plan, process, or product can be assessed. Policing to be effective has to evolve from being a reactionary force to being deterrence for criminal activity.
The use of metrics by Police Departments for internal use has been very beneficial for the deployment of resources in specific areas. On the other hand, the use of metrics released to the general public to bolster the reputation of politicians and police chiefs has created incentives for gaming and fudging the numbers. This is especially true when pay for performance or advancement relies heavily on the success of metrics depicting an overall decrease in crime.
Compstats is a crime analysis system using GIS to track crime incidents. It is used to discover crime patterns and for discussion in weekly meeting where police managers are held accountable for the results in their assigned areas. The data can be used to identify hot spots and timelines for activity and to deploy police resources accordingly. All of the sound logical and hypothetically should produce positive results.
The main concern with data collection is the accuracy and reliability of the data collected. Compstats is a useful system for gathering information and indicative metrics, but when individuals pressure police administrators to show improvements in the overall numbers, and that pressure is passed down the chain of command the message heard by the officers on the streets is that they will be penalized for an increase in reported crimes. When career advancement and pay for performance are dependent on a steady decrease in crime, officers and administration will find ways to fudge the number so that statistically crime is decreasing. Cities with low crime rates attract economic development and increase in population. When the crime index goes down politicians tout their success. When crime increases politicians are criticized by watch groups and by their opponents. The chances of not getting re-elected increase.
This scenario fosters tremendous pressure and temptations to insure that crime is constantly declining. Fudging the number is easy. Some tactics include:
- Intentionally misclassifying a case to a lower offense
- Altering the narrative to a lesser charge
- A house break-in becomes trespassing
- A garage break-in becomes criminal damage to property
- A theft becomes lost property
- A shooting becomes accidental discharge of a weapon
- A burglary, if nothing is stolen, becomes damage to property
- Auto theft if not an offence if the vehicle owner did not take proper precaution to secure the vehicle.
- Rape is consensual if the victim did not put up a fight.
- Family violence is a disagreement.
This way a major offence is transformed to a minor offense and a decrease in major crimes is reported and statistically crime is decreasing. The politicians are happy, the citizens are happy, and life goes on. But eventually the realization sets in that all is not right. In community meeting citizens continue to talk about the increase in crime in their neighborhoods. There is a dis-connect between the crime stats and reality as viewed by the citizens.
The other key element of Compstats is arrests. The effectiveness of the police department can be measured by the number of arrests made. A high crime rate with low number of arrest is not good. A low crime rate with high number of arrests is good sign of an effective and efficient police force. An arrest is an arrest and it doesn’t matter if it is a murder suspect or a jay-walker. It takes time to investigate and make a solid case for a drug distributor but in the same amount of time an undercover officer can probably arrest numerous teenagers a day selling drugs. Arresting one drug kingpin would significantly reduce drugs in an area as opposed to arresting the 5 dope heads selling marihuana on the street corner or at the car wash. But 5 is greater than 1. Since every arrest is of equal weight why not go for the easier arrests with higher numbers and get recognized and possibly a promotion.
Compstats is a useful tool for effective and efficient policing if properly used. Using statistics as measurements for rewards and punishment can lead to metrics that are subject to manipulation, reliability issues, and in some cases the Compstats can be totally counterproductive. Some agencies have gone as far as implementing a data verification system and auditing team to review data entry and to punish officers that habitually misreport offenses. Let’s not fall victims of common sense.
 Jerry Z. Muller, “The Tyranny of Metrics”, 2018, Princeton University Press, Chapter 10
 Mac Donald, “Compstats and its Enemies.”