Mass killings, 4+ killed (AP) in the United States

America's crime record

incidents, 2024 · Down 9.5% from 2023

Timeline

Mass killings at AP's threshold, higher than the federal one: four or more people killed within 24 hours, not counting the offender, with any weapon, in any place, for any motive: family killings, killings during other crimes, and public attacks, and the people killed in them.

Source and method

AP, USA Today, and Northeastern University compile them from news, court records, and the FBI's homicide reports and check each case by hand. Its offenders and victims carry the fullest traits of any count here. A different definition from the 3+ count and from active shooters; they overlap and are never added.

No rate. An incident is not something that happens to a person, so incidents take no rate per resident; the people killed have one, per 1 million residents.

AP, USA Today, and Northeastern University mass killing database · Calendar year · 2006 to 2024

Every yearThe Supplementary Homicide Reports (SHR) master file and the NIBRS national master file are joined into one record. Mass killings, three or more killed: SHR records and NIBRS incidents that describe the same killing are linked into one incident, so none is counted twice.
2006 onThe AP, USA Today, and Northeastern mass killing database and the Supplementary Homicide Reports (SHR) master file count the same kind of event and are kept apart: never merged, added, or used to fill each other's years. Four or more killed is a measure of its own: the FBI's records are the government's count and AP's database the compiler's, and counts under the two thresholds are never added. Records of one killing are linked case by case, never merged.
2006 onThe AP, USA Today, and Northeastern mass killing database and the NIBRS national master file count the same kind of event and are kept apart: never merged, added, or used to fill each other's years. Four or more killed is a measure of its own: the FBI's records are the government's count and AP's database the compiler's, and counts under the two thresholds are never added. Records of one killing are linked case by case, never merged.
2025–2026Left out: the copy loaded runs to 2026-04-19, and its 2025 (18 incidents) looks incomplete against 24 to 46 a year before; both years are left out of comparisons until a newer copy loads.

Breakdown

People killed by age, 2006 to 2024

People killed by age, 2006 to 2024
Under 12	19%604	604
12 to 17	9%271	271
18 to 24	15%464	464
25 to 34	16%497	497
35 to 49	19%608	608
50 to 64	14%438	438
65 and older	8%252	252
Age unknown	1%19	19

All 3,153 people killed, 2006 to 2024, every year added.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

Every figure as a table

Mass killings, 4+ killed (AP), incidents, by year

Year	Count	Note
2024	38
2023	42
2022	42
2021	35
2020	24
2019	46
2018	25
2017	33
2016	34
2015	32
2014	25
2013	25
2012	22
2011	30
2010	26
2009	32
2008	37
2007	26
2006	38

No rate. An incident is not something that happens to a person, so incidents take no rate per resident; the people killed have one, per 1 million residents.

2025 to 2026 left out: Left out: the copy loaded runs to 2026-04-19, and its 2025 (18 incidents) looks incomplete against 24 to 46 a year before; both years are left out of comparisons until a newer copy loads.

AP, USA Today, and Northeastern University mass killing database · Calendar year

Mass killings, 4+ killed (AP), people killed, by year

Year	Count	Rate
2024	165	0.5
2023	216	0.6
2022	210	0.6
2021	172	0.5
2020	110	0.3
2019	234	0.7
2018	141	0.4
2017	230	0.7
2016	200	0.6
2015	167	0.5
2014	113	0.4
2013	116	0.4
2012	133	0.4
2011	141	0.5
2010	122	0.4
2009	172	0.6
2008	178	0.6
2007	150	0.5
2006	183	0.6

Rate: People per 1 million residents of the United States that year (Census Bureau estimates). Per million, not per 100,000 as elsewhere on this page, since these are rare.

AP, USA Today, and Northeastern University mass killing database · Calendar year

People killed by age, 2006 to 2024

Age	Count	Share
Under 12	604	19%
12 to 17	271	9%
18 to 24	464	15%
25 to 34	497	16%
35 to 49	608	19%
50 to 64	438	14%
65 and older	252	8%
Age unknown	19	1%

All 3,153 people killed, 2006 to 2024, every year added.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

People killed by sex, 2006 to 2024

Sex	Count	Share
Male	1,654	53%
Female	1,487	47%
Sex unknown	12	<1%

All 3,153 people killed, 2006 to 2024, every year added.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

People killed by race, 2010 to 2024

Race	Count	Share	Rate
White+MENA	1,033	42%	0.4
Black	499	20%	0.8
Hispanic	389	16%	0.4
Asian+NHPI	153	6%	0.5
AIAN	29	1%	0.8
Race unknown	367	15%	–

Hispanic is of any race; the other groups are not Hispanic. White+MENA includes Middle Eastern and North African people.

All 2,470 people killed, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name. Rate: people killed per 1 million people in the group; – means no rate for that group.

Of these people killed, 15% have no recorded race. They are in no group's rate, so each rate is a floor.

AP, USA Today, and Northeastern University mass killing database

People killed by relationship, 2006 to 2024

Relationship	Count	Share
Family	889	28%
Stranger or bystander	764	24%
Someone the offender knew	543	17%
Tied to someone the offender knew	324	10%
Intimate partner	183	6%
First responder	46	2%
Relationship unknown	326	10%
Other	78	3%

All 3,153 people killed, 2006 to 2024, every year added. Each victim's relationship to the offender as AP recorded it, grouped by CrimeDB.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

Offenders by sex, 2006 to 2024

Sex	Count	Share
Male	700	89%
Female	52	7%
Sex unknown	34	4%

All 786 offenders, 2006 to 2024, every year added. Each offender as AP recorded it.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

Offenders by age, 2006 to 2024

Age	Count	Share
Under 12	1	<1%
12 to 17	35	5%
18 to 24	175	22%
25 to 34	265	34%
35 to 49	211	27%
50 to 64	53	7%
65 and older	7	1%
Age unknown	39	5%

All 786 offenders, 2006 to 2024, every year added. Each offender as AP recorded it.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

Offenders by race, 2010 to 2024

Race	Count	Of which placed by CrimeDB	Share	Rate
White+MENA	249	16	41%	0.1
Black	195	12	32%	0.3
Hispanic	91	6	15%	0.1
Asian+NHPI	33	0	5%	0.1
AIAN	9	0	2%	0.3 (fewer than 20 cases; the rate may swing a lot.)
Race unknown	30	0	5%	–

Hispanic is of any race; the other groups are not Hispanic. White+MENA includes Middle Eastern and North African people.

All 607 offenders, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name. Rate: offenders per 1 million people in the group; – means no rate for that group.

0 of the 0 incidents from 2010 to 2024 (NaN%) have no known offender, so no one from them is in the offender figures. Those with none, by kind: .

AP, USA Today, and Northeastern University mass killing database; CrimeDB's name-based race model (crimedb-bifsg-v2), 2010 to 2024

Grey: fewer than 20 cases.

Offenders by race, family killings, 2010 to 2024

Family	Count	Of which placed by CrimeDB	Share	Rate
White+MENA	133	8	53%	0
Black	65	6	26%	0.1
Hispanic	32	2	13%	0
Asian+NHPI	15	0	6%	0.1 (fewer than 20 cases; the rate may swing a lot.)
AIAN	4	0	2%	0.1 (fewer than 20 cases; the rate may swing a lot.)

All 249 offenders, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. AP's incident type: family (most victims are the offender's family), felony (tied to a crime such as robbery, drugs, or gangs), and public (victims in a public place, with no family tie and no felony). Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name. Rate: offenders per 1 million people in the group; – means no rate for that group.

AP, USA Today, and Northeastern University mass killing database; CrimeDB's name-based race model (crimedb-bifsg-v2), 2010 to 2024

Grey: fewer than 20 cases.

Offenders by race, felony killings, 2010 to 2024

Felony	Count	Of which placed by CrimeDB	Share	Rate
Black	77	2	47%	0.1
Hispanic	33	3	20%	0
White+MENA	33	1	20%	0
Asian+NHPI	8	0	5%	0 (fewer than 20 cases; the rate may swing a lot.)
AIAN	3	0	2%	0.1 (fewer than 20 cases; the rate may swing a lot.)
Race unknown	11	0	7%	–

All 165 offenders, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. AP's incident type: family (most victims are the offender's family), felony (tied to a crime such as robbery, drugs, or gangs), and public (victims in a public place, with no family tie and no felony). Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name. Rate: offenders per 1 million people in the group; – means no rate for that group.

Of these offenders, 7% have no recorded race. They are in no group's rate, so each rate is a floor.

AP, USA Today, and Northeastern University mass killing database; CrimeDB's name-based race model (crimedb-bifsg-v2), 2010 to 2024

Grey: fewer than 20 cases.

Offenders by race, suspected felony killings, 2010 to 2024

Suspected felony	Count	Of which placed by CrimeDB	Share	Rate
Black	14	2	41%	0 (fewer than 20 cases; the rate may swing a lot.)
Hispanic	12	0	35%	0 (fewer than 20 cases; the rate may swing a lot.)
White+MENA	4	1	12%	0 (fewer than 20 cases; the rate may swing a lot.)
Race unknown	4	0	12%	–

All 34 offenders, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. AP's incident type: family (most victims are the offender's family), felony (tied to a crime such as robbery, drugs, or gangs), and public (victims in a public place, with no family tie and no felony). Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name. Rate: offenders per 1 million people in the group; – means no rate for that group.

Of these offenders, 12% have no recorded race. They are in no group's rate, so each rate is a floor.

AP, USA Today, and Northeastern University mass killing database; CrimeDB's name-based race model (crimedb-bifsg-v2), 2010 to 2024

Grey: fewer than 20 cases.

Offenders by race, public killings, 2010 to 2024

Public	Count	Of which placed by CrimeDB	Share	Rate
White+MENA	52	3	52%	0
Black	28	2	28%	0
Hispanic	10	0	10%	0 (fewer than 20 cases; the rate may swing a lot.)
Asian+NHPI	8	0	8%	0 (fewer than 20 cases; the rate may swing a lot.)
AIAN	1	0	1%	0 (fewer than 20 cases; the rate may swing a lot.)
Race unknown	2	0	2%	–

All 101 offenders, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. AP's incident type: family (most victims are the offender's family), felony (tied to a crime such as robbery, drugs, or gangs), and public (victims in a public place, with no family tie and no felony). Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name. Rate: offenders per 1 million people in the group; – means no rate for that group.

AP, USA Today, and Northeastern University mass killing database; CrimeDB's name-based race model (crimedb-bifsg-v2), 2010 to 2024

Grey: fewer than 20 cases.

Offenders by race, other killings, 2010 to 2024

Other	Count	Of which placed by CrimeDB	Share	Rate
White+MENA	26	2	62%	0
Black	10	1	24%	0 (fewer than 20 cases; the rate may swing a lot.)
Hispanic	3	1	7%	0 (fewer than 20 cases; the rate may swing a lot.)
Asian+NHPI	2	0	5%	0 (fewer than 20 cases; the rate may swing a lot.)
AIAN	1	0	2%	0 (fewer than 20 cases; the rate may swing a lot.)

All 42 offenders, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. AP's incident type: family (most victims are the offender's family), felony (tied to a crime such as robbery, drugs, or gangs), and public (victims in a public place, with no family tie and no felony). Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name. Rate: offenders per 1 million people in the group; – means no rate for that group.

AP, USA Today, and Northeastern University mass killing database; CrimeDB's name-based race model (crimedb-bifsg-v2), 2010 to 2024

Grey: fewer than 20 cases.

Offenders by race, unsolved killings, 2010 to 2024

Unsolved	Count	Share
Race unknown	11	100%

All 11 offenders, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. AP's incident type: family (most victims are the offender's family), felony (tied to a crime such as robbery, drugs, or gangs), and public (victims in a public place, with no family tie and no felony). Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name.

Of these offenders, 100% have no recorded race. They are in no group's rate, so each rate is a floor.

AP, USA Today, and Northeastern University mass killing database

Offenders by race, undetermined killings, 2010 to 2024

Undetermined	Count	Of which placed by CrimeDB	Share	Rate
Black	2	0	40%	0 (fewer than 20 cases; the rate may swing a lot.)
Hispanic	1	0	20%	0 (fewer than 20 cases; the rate may swing a lot.)
White+MENA	1	1	20%	0 (fewer than 20 cases; the rate may swing a lot.)
Race unknown	1	0	20%	–

All 5 offenders, 2010 to 2024, every year added. From 2010, the first year CrimeDB has population by race. Rates are per 1 million people in the group a year. AP's incident type: family (most victims are the offender's family), felony (tied to a crime such as robbery, drugs, or gangs), and public (victims in a public place, with no family tie and no felony). Asian and Pacific Islander are one code (Asian+NHPI); Other names no race, so it is race unknown. AP, USA Today, and Northeastern record race from news, court records, and the homicide reports, checking each case by hand; Hispanic is one of the race values. Offenders with no race recorded are placed by CrimeDB's name-based model (crimedb-bifsg-v2: first name, surname, and the county's residents), the lighter part of each bar; with AP's recorded race hidden, its sums land within 2% of AP's counts for White, Black, Hispanic, and Asian+NHPI offenders, though it puts about one person in five in the wrong group. AIAN and two or more races can't be checked and stay race unknown, as do offenders with no name. Rate: offenders per 1 million people in the group; – means no rate for that group.

Of these offenders, 20% have no recorded race. They are in no group's rate, so each rate is a floor.

AP, USA Today, and Northeastern University mass killing database; CrimeDB's name-based race model (crimedb-bifsg-v2), 2010 to 2024

Grey: fewer than 20 cases.

Incidents by kind of killing, 2006 to 2024

Kind of killing	Count	Share
Family	296	48%
Public	113	19%
Felony	111	18%
Suspected felony	19	3%
Unsolved	18	3%
Undetermined	5	1%
Other	50	8%

All 612 incidents, 2006 to 2024, every year added. AP's incident type: family (most victims are the offender's family), felony (tied to a crime such as robbery, drugs, or gangs), and public (victims in a public place, with no family tie and no felony).

18 of the 612 incidents from 2006 to 2024 (3%) are unsolved; AP lists their offenders, where it lists any, as race unknown.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

Incidents by weapon, 2006 to 2024

Weapon	Count	Share
Handgun	201	33%
More than one kind of weapon	90	15%
Gun and a weapon that is not a gun	64	11%
Rifle	38	6%
Fire	36	6%
Knife or cutting tool	30	5%
Blunt object	19	3%
Shotgun	18	3%
Vehicle	7	1%
Strangling or suffocation	3	1%
Gun, type unknown	103	17%
Other weapon	3	1%

All 612 incidents, 2006 to 2024, every year added.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

Incidents by place, 2006 to 2024

Place	Count	Share
Home	424	69%
Business	61	10%
Open space	41	7%
More than one place	35	6%
Vehicle	17	3%
School or college	14	2%
Government property	10	2%
House of worship	7	1%
Health care	2	<1%
Other place	1	<1%

All 612 incidents, 2006 to 2024, every year added.

AP, USA Today, and Northeastern University mass killing database, 2006 to 2024

How we count

Age: Age at the offense, in the same bands on every page.
Sex: Biological sex: female or male.
Race and ethnicity: Six groups, the same on every page. Hispanic is of any race; the other five are not Hispanic, so no one is in two.
Relationship: The victim's relationship to the offender, as police recorded it.
Counting once: Each person counts once in each row they fit, so rows can add up to more than the total.
Unknowns: Unknown is always its own row, and every share reads against everyone.
Small counts: Every count is shown, however small.
Marks on figures: When a figure is missing, the page says why.
Rates: Pages call it the rate: how many for every 100,000 people, or every million for rare events. It always sits beside its count, never in place of it.
Who reports: Police figures count only the agencies that sent records to the FBI.

Every rule in full