When the Trump administration set up a federal task force to address crime in Memphis, Tennessee, last month, it cited crime data that didn’t match statistics from the city’s police department.
While a White House press release from September says violent crime rose in Memphis in 2024, Memphis Police Department data shows violent crime went down.
“Murder is at a six-year low, aggravated assault is at a five-year low and sexual assault at a twenty-year low,” the MPD said in a press release last month.
Can both be true? Technically, yes.
Same city, different statistics
The White House drew on FBI “Crime in the U.S.” data that shows over 15,000 violent crimes in Memphis in 2024, thousands more than what the city reports in the same time period.
Two factors help explain the difference: geography and how the crimes are counted.
The FBI counts crimes across the nine-county metro area, spanning parts of Tennessee, Mississippi and Arkansas, where the population is more than twice that of Memphis. The Memphis Police Department only counts crimes within city limits.
The second crucial factor is that the FBI counts each victim of violent crime, while Memphis police count each incident. One shooting that injures three people is recorded as three assaults in FBI data but only one in the city police data, which can make totals appear higher if the distinction isn’t clear.
“Our political leaders need to be really careful on how they talk about crime. Do they talk about the definitions?” said Dr. Thaddeus Johnson, a senior fellow at the Council on Criminal Justice. “Some of that nuance was lost in that conversation and they put people on edge.”
How violent crime is defined
Crime statistics also differ in part because there is no one definition of “violent crime.” Different law enforcement agencies will use subtly different definitions.
As the FBI defines it, there are four core crime categories within violent crime: murder/nonnegligent manslaughter, rape, robbery and aggravated assault.
To count and publish crime data, until recently the FBI used a longstanding system called the Summary Reporting System, or SRS. The program counts the four violent crime categories and tallies each incident once, even if there are multiple victims.
In 2021, the FBI switched to a new system called NIBRS, or the National Incident-Based Reporting System. It uses the same violent crime categories as SRS but counts each victim rather than each incident.
“NIBRS data is a massive improvement,” Jacob Kaplan, professional specialist at the School of Public and International Affairs, said in his book “Decoding FBI Crime Data,” published earlier this year. “It has led to an explosion of research that allows a far more detailed analysis of crime.”
Local police departments like Memphis generally use the FBI’s definitions of violent crime when publishing stats online. However, Memphis includes all types of homicides and lists subcategories like non-fatal shootings.
The federal Bureau of Justice Statistics adds “simple assault” to its violent crime statistics, which includes attacks where no weapon was used and did not result in serious injury. They also include crimes never reported to police, but reported in annual surveys conducted by the agency.
With all these different definitions, one agency’s crime statistics can show crime rising while another shows it falling.
“All of them tell pieces of the truth,” said Johnson. “Read the fine print. Most local websites have the methodology. And it tells you, are they counting by incidents or the number of victimizations?”
Johnson recommends a few key checks when looking at crime statistics, whether on your local police dashboard or while scrolling social media.
-
Definitions: What crimes are included?
-
What’s being counted: Are totals based on incidents or the number of victims?
-
Time frame: Are figures year-to-date, a rolling 12-month average or a single month?
-
Reporting gaps: Many crimes are never reported to the police.
Ultimately, Johnson says, being careful with crime data is part of showing respect to the victims and their stories.
“People are not data points, right? The data points, you see, those are people, those are souls. Those are families,” he said.
