59 large US cities · 2022 to 2025
In all 46 large cities with complete data, Black women's reported assault rate is higher than White women's
In the typical city it is about 3.8 times White women's and 2.4 times Hispanic women's. The pattern holds under cautious assumptions, and nothing measured explains it.
This report applies one method to every large US city that qualifies: at least 250,000 residents, at least 10,000 Black women residents, and complete data in the FBI's NIBRS files from the city's own police department. The rule was set before any results were seen. Forty-six cities have complete data for 2022 to 2025. Another 13 are complete only for 2024 to 2025 and are shown apart. Together they hold 2,552,921 assault victims.
In all 46 cities, Black women's reported assault rate is higher than White women's. In all 41 that record ethnicity, it is higher than Hispanic women's. In the typical city, it is about 3.8 times White women's rate and about 2.4 times Hispanic women's. The 13 cities shown apart look the same: Black women's rate is higher than both groups' in all 13.
All 46 cities with complete data for 2022 to 2025. The five that record no ethnicity are counted at their White-women ratio, which is a floor. The typical-city figures in the text leave out cities flagged for unstable data. Source: FBI NIBRS files; ACS 2024 5-year population.
How sure the finding is
It is not chance: every one of the 170 ratios in the full analysis has a 95% interval above 1. Victim race is recorded by officers, and the finding holds under cautious assumptions about how race and ethnicity are recorded: against White women in every city, and against Hispanic women in every city but Cincinnati, which has few Hispanic residents and unknown ethnicity for 57% of women victims. It holds at the same age mix in every city and outside intimate partner assault in every city. It is widest for aggravated assault in 45 of the 46.
Among cities that record ethnicity, the gap with White women runs from 1.8 times in Stockton and El Paso to 10.2 times in DC, with Minneapolis (7.7) and Chicago (7.3) next. Departments code offenses differently, so rates aren't comparable from one city to the next. The ratio within a city is the measure that travels.
Nine cities, closer up
Before this run, an earlier comparison set nine cities side by side, using each city's own police records for Los Angeles and Baltimore and FBI files for DC, Dallas, Houston, San Antonio, Austin, Fort Worth and El Paso. It found the same pattern: in all nine, Black women's reported assault rate is higher than Hispanic and White women's, from about 1.8 times White women's in El Paso to about 10 times in DC.
Sources differ by city (LAPD and Baltimore police records; FBI NIBRS files elsewhere) and windows run 2020 to 2023 for Los Angeles and 2022 to 2024 for Baltimore. Compare the pattern, not the levels. Baltimore's Hispanic comparison can't be settled because ethnicity is missing for 41.6% of women victims.
It also reached what the FBI files can't. Where neighborhood could be tested, in Los Angeles, Baltimore and Dallas, comparing women in similar census tracts still left 2.1 to 3.1 times against Hispanic and White women. And the gap was not concentrated in intimate partner assault: in six of the eight cities that record the relationship, it was smaller within partner assault than outside it. The full nine-city analysis has every table.
What this does not show
- Why. The data shows how often assaults on each group are reported, not why. Nothing here measures causes or offenders.
- Neighborhoods. The FBI's files have no victim location. Where neighborhoods were tested (Los Angeles, Baltimore, Dallas and New York's precincts), the gap narrowed but remained, at about 3 times within New York precincts of similar makeup.
- Every assault. Only assaults that reached the police are counted, and willingness to report differs by group, place and time. What the assault-victim gap measures compares these records with the national victimization survey and hospital records. Police data also reflects where officers patrol and who calls them, so it can't separate more policing or more reporting from more assaults.
- People. Rates count reports, not people: someone assaulted twice counts twice.
- Every city equally. Five cities record no ethnicity (Toledo, Cleveland, Tulsa, Detroit and Oklahoma City), so they have no Hispanic comparison. Some thresholds, the typical-city figures and keeping Columbus and Tucson were decided after seeing the data; the full analysis discloses each.
- The nation. These are large cities with complete data, not a national sample.
Part of the series Assault victims by race and sex, 6 reports.