2022-01-01 to 2025-12-31 · 56,092 victims · Martin Ngoh
Black women in DC are assaulted at 2 to 3 times the rate of Hispanic women and about 10 times the rate of White women. The gap is widest for the most serious assaults, and neighborhood accounts for part of it.
Per 100,000 residents a year, Black women are assaulted 4,205 times: 3.2 times Hispanic women, 10.1 times Asian women, 10.2 times White women.
Women only (26,341 Black women, 4,455 other women). Each cut asks whether a plain explanation accounts for the gap.
If the gap came from Black women reporting more, it would narrow for the crimes reported most often, and it widens instead (Severity, above). Nationally, 40% of simple assaults reach police against 75% of crimes involving a gun, and reporting differs little by race: 51% of violent crimes against Black victims and 48% against White victims in 2024 (56% and 42% in 2023). For reporting alone to make the 10.2x gap, White women would have to report under 10% of assaults even if Black women reported all of them. Unreported assaults stay invisible here.
NIBRS has no location. MPD's public incident data has locations but no victim race. This section places 4,741 assaults with a dangerous weapon from 2022 to 2025 in census tracts and averages each tract's rate per resident over where each group's women live. It describes places, not victims.
If every woman faced her own tract's rate, Black women's aggravated assault rate would be 2.8x White women's and 2x Hispanic women's. It is 14.4x and 4.7x. Neighborhood accounts for a factor of about 2.8 against White women; a factor of about 5.2 remains that this check cannot place. Other definitions give a smaller neighborhood factor: 1.8x counting all violent crime, 2x per square kilometer.
These are tract averages, not victims. Neighbors on different blocks can face different risks, and assaults away from home are not placed. Neighborhood poverty is shaped by segregation and discrimination, so the part of the gap found in neighborhoods is located, not explained away.
The national crime survey, which counts crimes whether or not they were reported, finds similar rates of violence for Black and White people: 23.4 and 22.1 per 1,000 in 2024, both sexes. Poor urban Black and White residents had similar rates too (51.3 and 56.4 per 1,000, 2008 to 2012).
Written down before it was run (docs/clearance-test-plan.md). An arrest within 90 days of the incident, from MPD's arrest records in the same NIBRS files, for women victims, compared inside each relationship and assault type, because DC law requires an arrest in domestic and family cases when there is probable cause.
Both comparisons are mixed by the rule written before the test. Compared like with like, Black women's cases end in an arrest 0.9 times as often as White women's (95% CI 0.85 to 0.96) and 0.86 times as often as Hispanic women's (95% CI 0.81 to 0.9). The gap sits in partner and other family assaults: in partner assaults, 0.74 times as often as White women's (95% CI 0.69 to 0.81). There is no clear difference for stranger or unknown relationships. It holds in both halves of the window. Against White women it sits in same-day arrests (0.88x), not later ones (1.06x).
Before comparing like with like, Black women's cases look more likely to end in an arrest (38.4% against 33.4% for White women). More of them are partner assaults (32.7% against 22.0%), where DC Code 16-1031 requires an arrest when there is probable cause, and fewer have an unknown relationship (18.5% against 41.9%), where arrests are rare.
Cases with Black women victims are closed as "victim refused to cooperate" 2.1 times as often as White women's (95% CI 1.54 to 2.87), like with like (6.4% against 1.9% before adjusting). That is the reason police recorded. It cannot separate a victim's choice, or her safety, from how the case was handled.
Clearance is not a measure of police effort. It depends on whether the person who did it was still there, whether there was probable cause, the victim's cooperation and the evidence. Nothing here measures offenders.
DC has one records system for the whole window, so the check compares the first two years with the last two, with identical definitions (MPD, offenses 13A and 13B, individual victims, the same race mapping and population). It tests whether the ratios hold over time, not whether a different coding system would find them.
The data says Black women are assaulted at a higher rate. It does not say why. Nothing here measures causes, offenders or circumstances.
Every number is a report that reached the police. Willingness to report, and recording practice, differ by group, area and time.
Rates count reports. Someone assaulted twice counts twice, so a rate is not the share of people assaulted.
Rates divide by where people live, not where they spend time.
Race is recorded by officers; the population counts people who are Black alone. There are 9% more people who are Black alone or in combination. If multiracial victims are recorded as Black, the worst case is 2.96x Hispanic, 9.38x White, 9.31x Asian instead of 3.22x, 10.21x, 10.13x.
1.2% of Black residents are also Hispanic, so they sit in both denominators.
2,844 victims are left out of the rates: their sex is unknown, or their race is unknown or outside the compared groups. Race is unknown or outside the groups for 3.4% of women and 6.5% of men.
The FBI's NIBRS files record no address, ward or police district, so the victim-level location test and the tract-level model from the Los Angeles analysis cannot run. The neighborhood section uses MPD's public incident locations instead. Those carry no victim race, so that check describes places, not victims.
NIBRS records nothing about a victim's income, and DC's economic divide is extreme. 26.3% of Black women are below the poverty line, against 5.1% of White women. Median household income is $60,764 for Black households and $170,201 for White ones. 43.1% of Black women live in census tracts with 25% or more poverty, against 6.2% of White women. Income could account for a large part of the gap; the neighborhood section measures only the part that runs through where women live.
The DC file covers the Metropolitan Police and Metro Transit Police only. Assaults recorded only by federal or campus police are missing. If those skew toward White women, the gap is somewhat overstated.
Police data reflects where officers patrol and who calls them. This data cannot separate more policing or more reporting from more assaults.
NIBRS records ethnicity separately from race, and it is unknown or unspecified for 30.6% of women victims. 28.8% of women recorded as White have unknown ethnicity, and where it is known, 53.8% of them are Hispanic. They are counted as White here. If they were Hispanic, Black women's rate would be 2.22x Hispanic women's instead of 3.22x, and 19.11x White women's instead of 10.21x.
1,059 women victims have unknown race. NIBRS has no location, so where they cluster cannot be mapped. Spread over the four groups like known victims of the same assault type, premises and resident status, the ratios move to 3.18x Hispanic, 10.02x White, 9.88x Asian from 3.22x, 10.21x, 10.13x.
DC draws many commuters and visitors, and rates divide by residents only. Resident status is blank for 36.3% to 43.9% of women victims depending on group (36.3% Black, 40.9% Hispanic, 43.0% White, 43.9% Asian), so the non-resident test is partial.
Partner assaults are identified from the victim-offender relationship, which is unknown or blank for 18.2% of Black, 28.0% of Hispanic, 41.8% of White, 42.6% of Asian women victims. Where it is missing more often, the partner share is understated more, so partner comparisons are least certain against the groups with the most missing.
Aggravated (13A) and simple (13B) assault reported by the Metropolitan Police Department, individual victims of any age. Left out: 5,678 Metro Transit Police victims, whose system extends into Maryland and Virginia; 3,067 assaults on officers; 18,691 intimidation victims (threats with no attack). 57 American Indian or Alaska Native and 33 Native Hawaiian or Pacific Islander victims are counted in the totals but not compared, because counts this small give unstable rates. 138 victims with unknown age or under one year old fall out of the age test only.
Victims cover January 2022 to December 2025; the population is the ACS 2024 5-year estimate, an average over 2020 to 2024.
The Asian comparison moved +22% between 2022 to 2023 and 2024 to 2025 (-25% for aggravated assault), against 5% or less for the others. It rests on 160 Asian women victims in 2022 to 2023 and 129 Asian women victims in 2024 to 2025, so it is left out of the headline.
Rates are victims per 100,000 residents per year over 4.0 years; cells under 2,000 residents are not rated. Women only from the tests onward. Sources: FBI National Incident-Based Reporting System (NIBRS) state files for the District of Columbia, 2022 to 2025; US Census Bureau ACS 2020 to 2024 five-year estimates via Census Reporter; DC Open Data, MPD Crime Incidents 2022 to 2025 (locations, no victim race); BJS, Criminal Victimization, 2024 (NCJ 310547); BJS, Household Poverty and Nonfatal Violent Victimization, 2008-2012 (NCJ 248384); DC Code 16-1031, arrests in intrafamily offenses (code.dccouncil.gov).