Traffic Deaths by State, 2021-2024: NHTSA FARS Data

Name: U.S. Traffic Fatalities by State, 2021-2024 (NHTSA FARS)
Creator: National Highway Traffic Safety Administration
Published: 2026-10-07
License: http://www.usa.gov/publicdomain/label/1.0/

This page counts fatal traffic crashes and the people killed in them, state by state, for calendar years 2021 through 2024. Every number comes from the National Highway Traffic Safety Administration's Fatality Analysis Reporting System, usually called FARS. We downloaded NHTSA's national FARS file for each of the four years, totaled it to one row per state per year, and then checked our totals against the figures NHTSA publishes itself. The full dataset is 204 rows and you can download it below as a CSV.

We did not model, estimate, adjust or smooth anything. These are counts.

Read this before you use the numbers

FARS counts deaths. It does not count injuries.

That one sentence rules out most of the questions people bring to a page like this. If you want to know how many people were hurt in car crashes in Ohio last year, this dataset cannot tell you, and neither can any other FARS table. A crash only enters FARS if someone died. Injury totals come from a different NHTSA system built on a sample of police reports, and the two are not interchangeable.

The full list of limits is further down. Please read it.

What each column means

A crash counts once. A death counts once per person. So a single crash that killed three people adds 1 to fatal_crashes and 3 to fatalities.

Column definitions for the downloadable CSV
Column	What it counts
`state_name`	State or the District of Columbia
`state_fips`	The state's FIPS code
`year`	Calendar year of the crash
`fatal_crashes`	Crashes in which at least one person died
`fatalities`	People who died
`alcohol_involved_fatal_crashes`	Crashes where at least one driver was recorded as drinking — see the note below, this is not the same as NHTSA’s drunk-driving figure
`speeding_related_fatal_crashes`	Crashes where an officer recorded racing, exceeding the speed limit, or driving too fast for conditions for at least one driver
`pedestrian_fatalities`	Pedestrians who died
`motorcyclist_fatalities`	Motorcycle riders and their passengers who died
`large_truck_involved_fatal_crashes`	Crashes involving at least one truck with a gross vehicle weight rating (GVWR) over 10,000 lb

About the alcohol column. This counts crashes where a driver had police-reported alcohol involvement or tested positive for any amount of alcohol. NHTSA’s published “alcohol-impaired driving” statistic is a different thing: it uses a blood alcohol concentration of .08 or higher, and it fills in missing test results statistically. Our column uses only what was actually recorded, so it runs lower. For 2024 we count 8,990 crashes; NHTSA’s impaired-driving figure for the same year is 10,832 fatal crashes. NHTSA’s own manual says this field “may under-count the actual number of drinking drivers” because alcohol test data is often missing. Treat the column as a floor, not a total.

The data: 2024 by state

The table below shows 2024 only, which is 51 rows. Four years at once would be 204 rows, which is more than anyone reads on a screen, so the earlier years live in the CSV. If you want 2021, 2022 or 2023, download the file — it has all four years in the same columns.

States are listed alphabetically rather than ranked, for a reason covered in the limits below.

Fatal motor vehicle crashes and deaths by state, 2024 (NHTSA FARS, Annual Report File)
State	Fatal crashes	Deaths	Alcohol-involved crashes	Speeding-related crashes	Pedestrian deaths	Motorcyclist deaths	Large-truck crashes
Alabama	895	962	238	263	128	127	128
Alaska	63	70	14	30	15	10	9
Arizona	1,118	1,229	250	394	250	219	118
Arkansas	547	603	147	125	69	76	85
California	3,583	3,876	954	1,031	1,090	522	355
Colorado	642	689	205	215	112	165	90
Connecticut	285	310	91	110	61	75	29
Delaware	121	126	25	28	28	22	11
District of Columbia	46	47	7	17	18	9	1
Florida	2,931	3,138	475	313	668	642	322
Georgia	1,312	1,403	215	307	286	182	181
Hawaii	97	102	29	40	35	26	2
Idaho	219	238	74	55	14	47	43
Illinois	1,085	1,177	255	376	213	149	130
Indiana	785	832	171	206	104	137	135
Iowa	324	356	75	63	30	63	52
Kansas	314	339	111	63	34	56	58
Kentucky	663	707	171	86	95	118	79
Louisiana	705	752	206	182	159	70	86
Maine	167	177	56	38	13	35	28
Maryland	552	578	152	108	152	95	60
Massachusetts	349	363	95	96	76	70	29
Michigan	1,011	1,098	265	255	150	184	100
Minnesota	431	477	123	125	57	76	70
Mississippi	678	753	123	117	104	55	92
Missouri	882	955	234	288	137	145	125
Montana	193	206	84	83	10	38	30
Nebraska	223	251	75	55	19	32	53
Nevada	378	417	97	120	108	89	35
New Hampshire	120	133	36	36	10	35	12
New Jersey	638	670	131	144	210	112	62
New Mexico	378	409	65	135	90	50	75
New York	1,036	1,101	138	332	289	209	113
North Carolina	1,509	1,619	314	616	264	212	184
North Dakota	84	90	26	18	5	17	23
Ohio	1,077	1,157	313	181	122	222	153
Oklahoma	594	645	150	214	77	106	99
Oregon	491	538	142	142	96	85	67
Pennsylvania	1,060	1,127	256	396	173	217	144
Rhode Island	48	52	16	14	13	8	4
South Carolina	948	1,038	305	349	156	154	124
South Dakota	134	146	49	30	9	33	30
Tennessee	1,093	1,197	274	190	158	188	157
Texas	3,774	4,160	938	1,346	763	583	647
Utah	251	277	84	77	42	53	41
Vermont	53	59	24	27	6	7	4
Virginia	867	917	256	267	123	122	110
Washington	674	730	183	222	152	113	67
West Virginia	238	256	72	72	25	38	36
Wisconsin	529	595	172	164	58	106	69
Wyoming	102	107	29	42	4	24	20

Download: myinjurypros-us-traffic-fatalities-by-state.csv — 204 rows, all four years, plain CSV.

National totals

National fatal crashes and deaths by year (NHTSA FARS)
Year	Fatal crashes	Deaths	Release
2021	39,785	43,230	Final
2022	39,422	42,721	Final
2023	37,769	41,025	Final
2024	36,297	39,254	Annual Report File

Deaths fell in each of the three years after 2021. The 2024 figure is a first release and is likely to move — see the limits.

Limits of this data

These are not footnotes. They change what the numbers mean.

1. Fatal crashes only. No injuries at all.
FARS is, in NHTSA’s words, “a census of fatal motor vehicle crashes.” A crash qualifies only if it killed someone. There is no partial coverage of serious injuries, no minor-injury category, nothing. A state with 900 deaths may have had a hundred thousand injury crashes that year, and none of them appear here.

2. The death has to happen within 30 days.
If someone is injured in a crash and dies 31 days later, the crash is not a FARS case and does not appear in these counts. NHTSA’s rule is death “within 30 days (720 hours) of the crash.”

3. Public roads only. Private property is excluded.
A crash qualifies only if a motor vehicle was “traveling on a trafficway customarily open to the public.” Parking lots, private driveways, farm tracks and private land are out of scope. Someone killed in a parking-lot collision is not in this data.

4. 2024 is not final yet. 2023 is.
The 2021, 2022 and 2023 figures come from NHTSA’s Final File for each year. The 2024 figures come from the Annual Report File, which is the first release and which NHTSA says is “subject to change when it is finalized the following year.” This is not hypothetical. The 2021 count went from 42,939 at first release to 43,230 final. 2022 went from 42,514 to 42,721. 2023 went from 40,901 to 41,025. Expect 2024 to rise by something in that range. If you cite the 2024 number, date it.

5. These are counts, not rates.
Texas has the most deaths of any state in 2024, at 4,160. Texas also has about 31 million residents and a great deal of driving. The District of Columbia shows 47. Reading that as “Texas roads are 88 times more dangerous than DC’s” would be wrong. A table of raw counts mostly ranks states by how big they are. To compare risk you need deaths per head of population or per mile driven, and neither is in this file. That is why we list states alphabetically instead of ranking them.

6. Check our figures against NHTSA’s current reports, not superseded ones.
NHTSA revises FARS figures after first publication, so an older report can disagree with us even when we are right. The 2021 motorcyclist count is the example. Overview of Motor Vehicle Traffic Crashes in 2022 (DOT HS 813 560, June 2024) gives 6,143. Motorcycles: 2023 Data (DOT HS 813 732, July 2025) then states that “due to amendments in the 2021 FARS Final file, the 2021 motorcyclist fatalities changed from 6,143 to 6,144,” and the current Motorcycles: 2024 Data (DOT HS 813 824, July 2026) still reports 6,144. Ours is 6,144. All four of our motorcyclist totals match NHTSA’s current published figures.

7. Fifty states and DC. No territories.
FARS itself covers the 50 states, the District of Columbia and Puerto Rico, but NHTSA ships Puerto Rico in its own file rather than the national file we used, so Puerto Rico is not here. No territory appears in these counts.

Where the numbers come from, and how we checked them

Source: National Highway Traffic Safety Administration, Fatality Analysis Reporting System (FARS), national CSV files for 2021, 2022, 2023 and 2024. nhtsa.gov/crash-data-systems/fatality-analysis-reporting-system

Column definitions follow the FARS Analytical User’s Manual, 1975-2024 (DOT HS 813 794).

We checked our totals against NHTSA’s published figures rather than trusting our own arithmetic:

All 204 state-year death totals were compared against NHTSA’s FARS Encyclopedia table “States — Fatalities and Fatality Rates, 1994-2024.” All 204 matched.
National fatal-crash counts, death counts and pedestrian death counts matched NHTSA’s published figures for all four years.
Speeding-related and large-truck death totals matched NHTSA’s published figures for all four years.
Motorcyclist death totals matched NHTSA’s current published figures for all four years, including 2021, which NHTSA revised from 6,143 to 6,144 after first publishing it.

NHTSA publishes FARS as public domain. The DOT open-data record for FARS carries the US Government public-domain label (usa.gov/publicdomain/label/1.0/) and lists its rights as public. NHTSA asks for no attribution, though we credit it anyway.

This page was compiled by My Injury Pros. We did not collect the underlying data and we make no claim to it. Corrections are welcome — if you find a mistake in our aggregation, tell us.

Data downloaded and page last updated: October 7, 2026.