Interview with Steven K. Aurand
![](https://assets.losspreventionmedia.com/2022/09/steven-aurand-150x200-1.jpg)
A researcher and statistician, Aurand cofounded CAP Index in 1988 and is the company’s president and CEO. Today, the company he leads is the most trusted and recognized source of crime risk information for businesses and government agencies throughout the US, UK, Canada, and soon Mexico. He holds a BA degree in psychology and music from Haverford College and has completed graduate studies in criminology and statistics at the Wharton School of the University of Pennsylvania. Aurand can be reached at saurand@capindex.com.
Asset protection and security teams often seek statistics to help assess the risk of crime when developing programs to secure local store locations. But what data are best suited and most trusted? We asked Aurand for his perspective.
What tools should one use to assess crime risks?
Ideally an organization should use a multiprong, data-driven approach to assessing its crime risks involving what I like to call the “three legs of the stool.”
- Site-specific crime risk data such as CAP scores to objectively quantify the crime potential at a given location,
- Incident history data to track what occurred at the location previously, and
- Survey information such as distances to highway exits to identify mitigating factors that might draw crime toward or away from a specific location.
Isn’t data from police reports or the FBI’s Uniform Crime Reporting program enough?
No, using police reports and UCR data alone falls short of the mark. Police reports can help explain the historic crime environment at a specific location. However, UCR data has numerous shortcomings— their use is retrospective, the data often inconsistent from one jurisdiction to the next, they can be subject to political influences and financial constraints, they are not industry-specific, and police-gathered crime data is difficult or impossible to obtain in many jurisdictions.
UCR data have their place when looking at broad geographic and temporal crime trends. However, the FBI itself strongly discourages ranking and comparing cities and regions using UCR data alone. Furthermore, FBI data are not at all appropriate for conducting site-specific assessments. It is an ecological fallacy to assume that relationships observed at the city level also operate down at the neighborhood level.
Like with UCR data, the media tends to examine crime using broad geographic areas. There have been a lot of recent news reports characterizing certain cities as “dangerous” based on aggregate violence and shootings data. However, these reports often dumb down complex situations and fail to point out that these crime spikes are often concentrated in specific neighborhoods and are not citywide.
Nearly 40 percent of law enforcement agencies nationwide failed to report their 2021 crime data to the FBI. What impact will this have?
This gap in data will be a problem for those who are dependent upon UCR data for their intended purpose. However, it won’t impact companies like ours that build crime risk models using site-specific crime data gathered directly from police departments. So, while the NYC and LA crime data won’t be in the FBI-collated data, we have collected all their crime data for 2021. More importantly, we carefully review and clean all the police data we gather to make sure that it is appropriate for our crime risk modeling.
How can security professionals cut through media hype to make more objective assessments?
A one-size-fits-all security approach leads to underspending at some locations and overspending at others. Address-specific crime data illustrate the tremendous variability that can exist between neighborhoods. Armed with site-specific crime data, LP professionals can better allocate their resources to focus security measures where they are needed most to mitigate actual crime risks rather than misperceived ones.
How does CAP Index keep improving its assessments?
Every year, we focus on increasing the precision and utility of our crime risk data by:
- Collecting data from additional sources such as more suburban and rural police jurisdictions.
- Including more predictors, such as how the proximity of certain types of businesses can affect crime risks.
- Adding new crime types to the mix, such as simple assault and vandalism.
- Exploring ways to leverage new techniques such as AI and machine learning.
- Providing new scoring methodologies to account for perpetrators who travel varying distances to commit crimes.