In the world of criminology, there is a statistical relationship between the volume of crime and population. But these numbers alone give us an incomplete picture. When assessing crime risk across geographies, it is important to consider crime rates as opposed to crime counts.
Measures of crime rates vs. crime counts
Crime counts reflect the raw number of recorded offenses. These counts can be misleading when used for comparison. For instance, suppose place A and place B both have had 100 assaults over the last month. Additionally, suppose place A has a population of 500 people and place B has a population of 1,000 people. Using crime counts alone provides a misleading understanding of both areas’ crime risk. The two areas have had an identical number of assaults, but the second area has twice as many people. Thus, the probability of being victimized in place A is twice as high as place B.
Crime rates, instead, account for the differences in populations. A crime rate is defined as the count of offenses divided by the number of people in the area and can be interpreted as crimes per person. This results in a more meaningful metric for comparing crime risk across space.
The numerator and the denominator
An accurate characterization of a crime rate requires both a correct numerator (the count of crime) and an accurate denominator (area population). The Pinkerton Crime Index utilizes many different sources and procedures for constructing the most accurate counts of crime in space (for instance, see how our team approaches the problem of crime underreporting in Mexico.)
In addition to making corrections and adjustments to the numerator (crime counts), determining a realistic denominator (area population), similarly requires additional work. Measures of population typically reflect residential population. However, this fails to capture movements of persons across space (both people coming into areas and travelling out of areas).
Many downtown or commercial spaces have a high number of businesses or attractions but a relatively low number of residents. Tourist areas, museums, government buildings, and university campuses are all examples of places where official census population numbers fail to capture the full share of people utilizing the space. Failure to account for this additional “floating population” yields a crime rate that is an inaccurate depiction of crime risk, because the denominator reflects only the residential population and not the actual number of people potentially being impacted by crime in the area.
Importantly, the number of crimes that occur is strongly correlated to the number of people utilizing the space. If population is undercounted, you will observe higher than expected crime counts being divided by lower than expected population estimates, resulting in an extremely high crime rate. Getting the denominator correct is extremely important to produce accurate crime rates and make meaningful comparisons.
To correct for this, Pinkerton data scientists quantify factors that proxy for additional people in an area. This information includes commuting behavior, tourism, as well as the number of hotels, restaurants, and entertainment venues. This additional information is used to either inflate or deflate estimates of population, yielding a crime rate that more meaningfully reflects the amount of people who live, work, and utilize a space.
Reliable crime risk scores
This work is another instrumental step in producing reliable, meaningful risk insights accessible at your fingertips via the Pinkerton Crime Index. Possessing a realistic understanding of risk on highly local levels facilitates important business and guardianship decisions. Click here for more information on how the Pinkerton Crime Index can help your business.
