Deciphering The 2024 National Crime Victimization Survey

The Bureau of Justice Statistics released the National Crime Victimization Survey for 2024 recently and it finds that neither property nor violent crime were statistically different from 2022 and 2023 levels. NCVS can be tough to wrap one’s head around sometimes, but in my opinion the most useful takeaway is that NCVS confirms that both violent and property crime in the United States are down substantially from where they were in the 1990s and are largely at the same levels they have been for most of the last 20 years.

Let’s Hook‘em and dive in.

Under the hood, the NCVS was redesigned for the first time since 1992 leading to a sample that was roughly half as large as usual. Remember, NCVS is a survey, and the result from a smaller sample is a much larger standard error throughout the estimative process. As BJS writes:

“…weights in 2024 were greater than in a typical year, leading to more variation in estimates and higher standard errors. Despite the loss of estimate precision in 2024, BJS and the U.S. Census Bureau determined that the stability gained from maintaining the current sample size over the long term outweighed the loss of statistical precision.”

The impact of a smaller sample can be clearly seen in the graph of violent victimizations. The estimated number of offenses (red line) suggests a slight increase in 2024, but adding in the 95 percent confidence interval shows far less certainty about the direction of violent crime last year. All we can say for certain is that NCVS finds that violent crime in 2024 was not statistically different from 2023.

Graphing the size of the standard error each year drives home just how large the confidence intervals are this year compared to recent years.

Property crime fell per NCVS but the drop is also not statistically significant.

The property crime rate standard error also rose, but the gap wasn’t nearly as big compared to previous years as it was with violent crime because there are so many more property crimes.

Overall, the share of violent victimizations reported to police (excluding simple assaults) jumped from around half to over 60 percent in 2024. This could be a change or (most likely) it could be a function of the smaller sample. Property crime reporting to police was largely in line with previous years.

Motor vehicle theft stands out as an outlier where NCVS isn’t picking up the national trend seen in the FBI data. Motor vehicle thefts weren’t statistically different in 2024 from 2023 which isn’t what I’d expect to see based on the FBI data. The graph below shows the sharp decline in the FBI’s most recent UCR estimate compared to a very different trend suggested in NCVS.

The FBI estimates that motor vehicle theft fell nearly 19 percent in 2024, so NCVS saying that it was either up or down slightly doesn’t really match what UCR is telling us. This difference is weird because auto thefts tend to be the crime that victims say get reported to police the most frequently. This was true in 2024 with 75 percent of victims saying they’d reported it compared to just around a quarter of other thefts.

But it’s also possible there’s more going on.

I wrote last year about the strengths and weaknesses of NCVS. As I noted there, one of the major factors that has to be considered is that responses are coded to year of survey rather than year of offense. A respondent will be asked if they’ve been the victim of a crime in the last six calendar months, so a survey taken in January 2025 will ask about July through December 2024.

Because of this feature, NCVS 2024 covers crimes committed between July 2023 and November 2024.

This is important because motor vehicle thefts were rising throughout 2023 during the first part of NCVS 2024 survey collection. The RTCI clearly shows the peak coming in December 2024 followed by a sharp downward turn.

So while motor vehicle thefts were down nearly 20 percent in 2024 compared to 2023 per both the RTCI and FBI, the decline was far more muted if you compare the timeframe that the NCVS sample is pulling from. Motor vehicle theft was only down 7 percent comparing July 2023 to November 2024 with July 2022 to November 2023. I can’t say for certain, but it would make sense if the wonky NCVS survey timeframe explains some of why NCVS and FBI disagree on the motor vehicle theft trend.

NCVS remains a valuable tool that helps to define the scale of non-reporting in America and highlights the contours of our nation’s reported crime trends. The methodological change for the 2024 NCVS adds more uncertainty than usual into the survey’s findings, but uncertainty is the rule in crime data. It’s true of the FBI’s estimates, it’s absolutely true of the Real-Time Crime Index, and it’s true of the NCVS.

Ultimately, NCVS confirms that American violent crime remains substantially lower than it was 30 years ago and is largely steady with where it has been for most of the last 15 years.