The headline numbers
107.6 severe injuries per 1,000 construction establishments (2015–2025).
OSHA-reported construction hospitalizations, amputations, and eye losses across 100,514 aggregated establishment records.
In the construction subset (NAICS 236/237/238), plus 2,842 amputations.
Read this before the ranking: two honesty rules
- This is aggregate only — never a contractor scorecard. Every figure is a state total or a trade-sector total. OSHA reports injuries at the establishment level, but attaching a severe-injury event to a named business in a public directory is a defamation class we refuse to touch. There are no business names, no EINs, and no per-contractor rows anywhere in this study or its open data.
- Only federal-OSHA states are ranked. Federal OSHA is the system of record for private-sector construction injuries in 30 states. The other 21 are OSHA State-Plan states that run their own private-sector reporting — so their severe injuries are largely absent from this federal dataset. Ranking them here would publish a near-zero rate that reflects reporting jurisdiction, not safety. We name them below and withhold a rate.
Federal-OSHA states, ranked by severe-injury rate
OSHA Severe Injury Report events in construction (NAICS 236/237/238) per 1,000 BLS QCEW construction establishments, 2015–2025. The median is 32 per 1,000 — read the rate as an incidence signal across each state's establishment base, not a verdict on any one contractor.
| Rank | State | Per 1,000 estabs | Severe injuries | Amputations | Fatalities |
|---|---|---|---|---|---|
| 1 | District of Columbia (DC) | 107.6 | 146 | 29 | 0 |
| 2 | Texas (TX) | 57.1 | 3,666 | 594 | 186 |
| 3 | North Dakota (ND) | 48.5 | 191 | 42 | 4 |
| 4 | Ohio (OH) | 39 | 1,016 | 183 | 64 |
| 5 | Alabama (AL) | 38.7 | 490 | 86 | 48 |
| 6 | Pennsylvania (PA) | 38.6 | 1,181 | 153 | 56 |
| 7 | West Virginia (WV) | 38.5 | 181 | 22 | 4 |
| 8 | Florida (FL) | 38.5 | 3,266 | 361 | 132 |
| 9 | Colorado (CO) | 37.2 | 808 | 123 | 30 |
| 10 | Kansas (KS) | 36.7 | 282 | 46 | 24 |
| 11 | Louisiana (LA) | 36.1 | 459 | 76 | 38 |
| 12 | Nebraska (NE) | 35.8 | 272 | 39 | 20 |
| 13 | Oklahoma (OK) | 33.4 | 373 | 64 | 14 |
| 14 | Wisconsin (WI) | 33.2 | 527 | 111 | 32 |
| 15 | Arkansas (AR) | 32.4 | 268 | 48 | 16 |
| 16 | Georgia (GA) | 31.6 | 882 | 152 | 46 |
| 17 | Mississippi (MS) | 31.6 | 217 | 34 | 18 |
| 18 | South Dakota (SD) | 31.3 | 138 | 20 | 6 |
| 19 | Delaware (DE) | 27.4 | 103 | 11 | 0 |
| 20 | Missouri (MO) | 27.2 | 456 | 65 | 40 |
| 21 | Illinois (IL) | 22 | 720 | 100 | 54 |
| 22 | New York (NY) | 21.4 | 1,127 | 174 | 50 |
| 23 | Massachusetts (MA) | 20.6 | 450 | 70 | 24 |
| 24 | Maine (ME) | 18.5 | 120 | 31 | 8 |
| 25 | New Hampshire (NH) | 17.8 | 99 | 21 | 2 |
| 26 | New Jersey (NJ) | 16.4 | 364 | 42 | 16 |
| 27 | Rhode Island (RI) | 16.4 | 67 | 12 | 0 |
| 28 | Connecticut (CT) | 15.1 | 147 | 28 | 10 |
| 29 | Idaho (ID) | 14.6 | 200 | 29 | 12 |
| 30 | Montana (MT) | 14.3 | 114 | 21 | 0 |
OSHA State-Plan states — named, not ranked
These 21states run their own private-sector severe-injury reporting, so the federal dataset undercounts them. We list the federal-only count for transparency, but we do not compute a per-1,000 rate from it, because that number would measure reporting jurisdiction rather than safety. For these states, consult the state plan's own published data.
By construction sector
The aggregate severe-injury and fatality counts across the three construction NAICS sectors (all states, all years in the pull). Specialty trade contractors (NAICS 238) — the plumbers, electricians, roofers, and concrete crews most homeowners hire — carry the largest share.
| Construction sector | Severe injuries | Amputations | Fatalities |
|---|---|---|---|
| Specialty trade contractors (NAICS 238) | 11,249 | 1,478 | 746 |
| Heavy & civil engineering construction (NAICS 237) | 3,911 | 755 | 730 |
| Building construction (NAICS 236) | 3,408 | 609 | 168 |
| Construction (other / unspecified) | 2 | 0 | 0 |
How construction workers get hurt
The most common event mechanisms behind the severe-injury reports, largest first. Falls to a lower level dominate construction severe injuries — which is why fall protection is OSHA's most-cited construction standard year after year.
- Other fall to lower level, unspecified2,022
- Compressed or pinched by shifting objects or equipment1,029
- Other fall to lower level 6 to 10 feet879
- Injured by slipping or swinging object held by injured worker796
- Other fall to lower level less than 6 feet756
- Exposure to environmental heat664
- Other fall to lower level 11 to 15 feet644
- Struck by falling object or equipment, n.e.c.612
- Caught in running equipment or machinery during regular operation530
- Struck by falling object or equipment, unspecified416
Severe-injury reports by year
Count of construction establishments whose most recent severe-injury report fell in each year, 2015–2025. This reflects report dates in the pull, not a complete annual time series — read it as coverage, not a trend line.
- 1,6542015
- 1,8012016
- 1,7862017
- 1,8952018
- 1,8712019
- 1,6472020
- 1,4642021
- 1,4482022
- 1,5612023
- 1,5522024
- 1,1902025
Methodology and honesty notes
- Aggregate only — no per-contractor attribution. The committed seed holds state and NAICS-sector rollups plus a national injury-type mix. No establishment name, EIN, or per-contractor row is included or reconstructable. Per-pro OSHA attribution is a defamation class; we publish the rollups and nothing finer.
- Numerator. OSHA Severe Injury Reports (mandated reports of work-related hospitalizations, amputations, and eye losses, 29 CFR 1904.39) plus Injury Tracking Application summaries, filtered to construction NAICS 236/237/238, summed by state.
- Denominator. 2024 BLS QCEW construction establishment counts, summed over the same NAICS scope (236+237+238) for the same state — so the rate is numerator and denominator from a consistent construction definition.
- Federal vs. State-Plan (the load-bearing caveat). Only the 30 federal-OSHA states are ranked. The 21 OSHA State-Plan private-sector states run their own reporting, so their federal counts are an undercount; we name them and withhold a rate rather than publish a misleading one.
- What the rate is not.A severe-injury report is a serious outcome, not a citation or a penalty (OSHA's inspection/violation tables require a separate API key we do not key here, so this study makes no citation or fine claims). The rate is an incidence signal across an establishment base, not a measure of any single business's safety.
- Establishment-name non-uniqueness + report lag. OSHA establishment names are not unique and reports can be revised or contested after filing, so even the aggregate totals carry the normal public-records lag. We aggregate to defuse both — small per-record noise washes out in a state or sector sum.
- Source + license. AGGREGATE ONLY — state and NAICS-sector rollups plus a national injury-type mix. No per-contractor row, name, or EIN is included or reconstructable. A 'severe injury report' is an employer's federally-mandated report of a work-related hospitalization, amputation, or loss of an eye (29 CFR 1904.39). The federal SIR program is the system of record for private-sector construction only in federal-OSHA states; the 21 listed OSHA State-Plan states run their own private-sector reporting, so their federal counts are an undercount and no per-1,000 rate is published for them. Public domain (17 U.S.C. 105). OSHA / U.S. Department of Labor + U.S. BLS.
- Computed live. Every number is read from the committed aggregate seed at render time. Rebuild the seed from a fresher OSHA pull and this page updates automatically — nothing here is hardcoded.
Why a contractor directory publishes this
Safety record is a real trust signal, but it is one that's easy to weaponize dishonestly. Plenty of sites would happily slap a red flag on a named contractor from an OSHA hit. We do the honest inverse: we publish the verifiable aggregate — which states and trades carry the most severe-injury risk — and we refuse to attribute a single event to a single business. Pair this with our license-lapse study and the establishment-density study for the full picture of any state's contractor landscape.
Frequently asked questions
Which state has the highest construction severe-injury rate?
Among the 30 federal-OSHA states, District of Columbia reports the highest rate — 107.6 OSHA-reported severe injuries per 1,000 construction establishments over 2015–2025 (146 severe-injury reports against 1,357 establishments). The median federal-state rate is 32 per 1,000. This is an incidence signal across each state's establishment base — not a per-contractor figure and not a ranking of any individual business.
Why are California, Michigan, and other big states missing from the ranking?
Because they are OSHA State-Plan states that run their OWN private-sector severe-injury reporting. Federal OSHA is not the system of record for private construction there, so their severe injuries are largely absent from this federal dataset — a near-zero federal count would be a reporting-jurisdiction artifact, not a safety record. We name all 21 of them and deliberately withhold a rate rather than publish a misleading one.
Is this a list of dangerous contractors?
No, and it deliberately cannot be. Every number on this page is an AGGREGATE — a state total or a trade-sector total. There is no per-contractor attribution anywhere, no business names, no license numbers. OSHA enforcement data is reported at the establishment level, but attributing safety events to a named contractor in a public directory is a defamation class we will not touch. We publish only the state and sector rollups.
What exactly is a 'severe injury report'?
Under 29 CFR 1904.39, employers must report every work-related in-patient hospitalization, amputation, or loss of an eye to OSHA within 24 hours (fatalities within 8 hours). Those reports are the numerator here, summed for construction NAICS 236/237/238 by state over 2015–2025. It is a count of severe outcomes, not minor injuries and not citations or penalties.
Where does the data come from, and can I reuse it?
OSHA Severe Injury Report events in construction per 1,000 BLS QCEW construction establishments. The numerator is OSHA's Severe Injury Reports + Injury Tracking Application (public records, public domain under 17 U.S.C. 105); the establishment denominator is 2024 BLS QCEW. We read both through ProFix's committed aggregate seed and recompute every figure live. The machine-readable data is at /api/construction-safety-by-state-2026.json, free to reuse with attribution.