How antisemitic incidents get counted

What it is

There are three separate machines producing the numbers you see in antisemitism headlines, and they are not measuring the same thing.

The first is the FBI. It publishes a national hate crime count assembled from what police departments send in. For 2025 the FBI recorded 10,881 hate crime incidents nationally, down from 11,679 the year before. ¹ Jews were the target in 1,528 of them, which is 63% of the 2,408 religion-based hate crimes and 14% of all bias incidents that year. ¹ The next most-targeted religious group was Muslims, with 205 incidents. ¹ The anti-Jewish figure included 162 assaults, three murders in Washington, D.C., and Boulder, Colorado, and 920 property-destruction offenses, the single most common category. ¹ The data covered agencies serving roughly 95% of the U.S. population, though about 14% of agencies did not participate at all. ¹

The second is the ADL's annual Audit of Antisemitic Incidents. Its 2024 edition recorded 9,354 incidents, up 5% from 8,873 in 2023 and the highest total in the 46 years the organization has been tracking. ² The breakdown was 6,552 harassment, 2,606 vandalism, and 196 assaults affecting 250 victims. ² The ADL counts vandalism of property, or harassment or assault on individuals or groups, where either circumstances indicate anti-Jewish animus on the part of the perpetrator, or "a reasonable person or group of people could plausibly conclude that they have been victimized due to their Jewish identity." ² It explicitly includes both criminal and non-criminal incidents. ² The material comes from direct victim reports, law enforcement, media monitoring, and partner Jewish organizations, and the ADL says it examines the credibility of everything reported and obtains independent verification when possible. ²

The third is survey research, which does not count incidents at all. It asks a sample of people what happened to them and scales the answer up. The federal version is the National Crime Victimization Survey, which estimated roughly 246,900 hate crime victimizations per year between 2005 and 2019. ³ Over roughly the same window, police departments recorded 6,628 hate crime incidents in 2010 and 7,314 in 2019. ³⁵ The Bureau of Justice Statistics attributes the gap to victims often not reporting to police. ³ In 2019 the survey put the rate at 1.0 violent hate crime victimization per 1,000 people age 12 and older, and hate crimes at 1.6% of all nonfatal victimizations, up from 0.9% in 2005. ⁴

For Jews specifically, the best-known survey is Pew's, built on 4,718 interviews with Jewish adults drawn from 68,398 screening interviews, conducted by mail and online between Nov. 19, 2019, and June 3, 2020. ⁶ In the prior 12 months, 37% of respondents had seen anti-Jewish graffiti or vandalism in their community, 19% had been made to feel unwelcome because they are Jewish, 15% had been called offensive names, 8% had been harassed online over their Jewish identity, and 5% reported a physical attack. ⁶ Separately, 53% said they felt less safe as Jews than five years earlier, and 75% believed there was more antisemitism in the United States than five years before. ⁶

Why it still matters

Put those three next to each other and the arithmetic looks broken. The FBI counted 1,528 anti-Jewish hate crimes in 2025. ¹ The ADL counted 9,354 antisemitic incidents in 2024. ² Pew found 5% of Jewish adults reporting a physical attack within a single year, a share that describes a far larger group of people than either tally contains. ⁶ Nobody is lying. The three numbers answer three different questions.

The FBI number is a floor. It counts things police recorded, classified as bias-motivated, and forwarded to a federal collection that about 14% of agencies skipped entirely. ¹ A crime that was never reported, never coded as bias-motivated, or reported in a jurisdiction that sat out the year is not in there. The victimization survey exists to estimate what that floor misses, and its answer is roughly 246,900 victimizations a year against about 7,000 police-recorded incidents. ³⁵

The ADL number is wider by design, because it counts conduct that is not a crime. ² The case for it is that most of what Jews actually encounter is legal: a swastika chalked on a sidewalk, a coordinated pile-on, a professor's remark, a flyer. A criminal-only tally erases nearly all of that, and a Jewish family deciding whether to put a mezuzah on the doorpost is not consulting the penal code. The case against it is that a category defined partly by what "a reasonable person" could "plausibly conclude" is elastic, and an elastic category will register more growth than a fixed one during a period when people are primed to interpret ambiguous events as hostile. ² Both of those are true at once. Read the ADL's assault subtotal, 196 in 2024, as the part most comparable to a criminal count, and the harassment subtotal, 6,552, as measuring something the FBI series was never built to capture. ²

Direction and level are also different things. A trend inside one series tells you something real; the same comparison run across two series tells you nothing. The FBI's anti-Jewish count fell 21% from 2024 to 2025. ¹ The ADL's total rose 5% from 2023 to 2024. ² Those cover different years, different definitions, and different collection pipelines, and setting them against each other produces a conclusion neither dataset supports.

Campus numbers deserve their own caution. The ADL logged 1,694 incidents at colleges and universities in 2024, an 84% increase over 2023 and 18% of its national total. ² An 84% jump can mean more antisemitism, or better reporting infrastructure, or both, and the audit's own sourcing includes media monitoring and partner organizations, which expanded their attention to campuses over exactly that period. ² A student survey and a campus incident tally are also not interchangeable. One measures how many students experienced something. The other measures how many reports reached someone who was collecting.

So when a headline number lands, ask who counted it and using what definition, whether criminal conduct was required, what period it covers, and what it's being compared to, which should almost always be the same series in an earlier year, not a different organization's total.

In the FBI's own data, Jews were targeted in 14% of all U.S. hate crimes in 2025 while making up around 2% of the population. ¹

The words you'll hear

Hate crime. In the federal count, a criminal offense that police recorded and classified as motivated by bias against the victim's race, religion, ethnicity, or similar characteristic; bias against race, ethnicity, or ancestry accounted for more than half of recorded incidents from 2010 to 2019. ⁵

Incident (ADL usage). Vandalism, harassment, or assault where circumstances indicate anti-Jewish animus or where a reasonable person could plausibly conclude they were targeted for being Jewish. It covers criminal and non-criminal conduct alike. ²

NCVS. The National Crime Victimization Survey, the federal household survey that estimates crime including offenses never reported to police, and the source of the roughly 246,900 annual hate crime victimizations figure. ³

Agency participation. Police departments supply the FBI's hate crime data, and coverage is partial. The 2025 release drew on agencies covering about 95% of the population, with roughly 14% of agencies not participating. ¹

Underreporting. The gap between crimes that occur and crimes police record. The Bureau of Justice Statistics names victims not reporting to police as the reason its survey estimates and law enforcement counts diverge so sharply. ³

Self-report survey. A study that asks a sample about their own experiences rather than tallying reported events, as when Pew found 15% of Jewish adults had been called offensive names in the prior year. ⁶