Forget state crime stats
In this day and age of constant digital connectivity and real-time communication, it’s a little puzzling that the FBI’s timetable for tallying an annual crime report still takes nearly a whole year to prepare.
It’s been an autumnal ritual now for many years: each October, the FBI tells us about violent and property crimes for the year that ended 10 months earlier.
More accurately, the FBI issues data calculations based on law enforcement reporting of criminal incidents. Statistical analysis can and often does involve techniques such as numerical averages and ratios, etc.
Those can be very useful in some instances and very worthless in others. For crime, specifically, some of the most common misimpression and misinformation errors involving averages are outsized.
Outliers are the bane of many data-set analyses, because they skew the average. Crime concentration levels are so disparate in terms of geography and population density that statewide averages are worse than meaningless—they promote false meaning.
And yet, what’s the only location filter the
FBI’s Crime Data Explorer (CDE) allows?
State or national choices.
Maybe that was all the methodology and technology could allow back in 1929, when the Uniform Crime Reporting system was first implemented.
Today, data transmission flows at the speed of light on fiber-optic cables, and there are pocket-sized high-definition video screens for every man, woman and child in the U.S. We are a nation pervaded by electronic devices that inundate us with information, and yet some of the most basic questions and summaries about crime go unanswered or unaddressed by the FBI.
Every incident report lists a location where the crime occurred, so why does the FBI CDE only allow drill-down by law enforcement agency? I’m no full-fledged data sleuth, so it’s possible that a crime-by-neighborhood or even crime-by-city report exists somewhere among the FBI downloadable spreadsheets.
But it’s not available on the Data Discovery Tool or the Crime Data navigation tabs. There are aggregate categories for locations—residence home, highway/alley/street/sidewalk, parking lot/garage, hotel/motel and some 40 more—but they can’t be ungrouped by geography.
Categorical location data can be informative: It is interesting, for example, to see that there are more violent crimes committed in Arkansas jails/ prisons/corrections facilities than in any of roughly 30 other location categories. And more than half of all violent crimes occur at home.
It would be more beneficial, however, to also see actual geography incorporated into the data subsets for cross-referencing with other characteristics like victim’s relationship to offender, type of weapon used and other linked offenses.
The only filters involving cities appear to be groupings by population, not by individual municipalities. So it lumps all small towns and big cities together, as well as everything in between, based on population ranges. That’s even worse than the proverbial apples-to-oranges folly; it’s more like trying to compare fruits to meats.
And while crime data can be broken down by law enforcement agency—that gives a better glimpse into cities by police department—there isn’t a table containing all agencies for a broad comparative view.
The purpose of tabulating crime data in the first place is to help law enforcement target resources and become more effective. In 2020, the FBI changed its methodology from the original Summary Reporting System (SRS) to the more modern National Incident-Based Reporting System (NIBRS) in hopes of capturing more detail.
But that meant law enforcement agencies and their officers had to accurately track more detail through a different system. Consequently, the number of agencies reporting plummeted the next year.
Last year, the FBI allowed SRS submissions again as well as NIBRS, and reporting went back up, but there are still data-collection kinks to work out.
It’s understandable that the FBI is hindered in its reporting by agencies that don’t submit data. But if the goal is to improve crime-fighting and enable better solutions to crime challenges, why limit reporting data to only agency-generated incident information?
Again, in our interconnected digital age, one would think cross-referencing crime data with other sources would be not only easy, but also extremely valuable in identifying broader contextual factors.
Even ZIP code breakdowns would be helpful. Since one in six violent crimes is committed by a schoolage offender, an education district cross-tabulation might also be productive. There’s a wealth of collected, stored digital data that could quickly and efficiently give greater context to the FBI’s crime report. And the cost of such cross-referencing would be small potatoes in the FBI’s $10 billion budget.
Crime occurs locally, committed by individuals in specific locations. Reducing crime can only occur when those specific areas are properly targeted with policies and protocols and programs. That requires more specific data reporting.
Violent crime could go up in half of Arkansas cities, and down in the other half, and our statewide trend would be reported as “no crime increase.” That’s useless.
People want and need to know the risk of crime to them and their family, in their community, and whether it’s getting better or worse.