Rail Safety by the Numbers: What FRA Data Tells Us About Human Factors
2 out of every 3 rail accidents are caused by human factors. Here’s what the data demands.
A track foreman arrives at the yard on a Tuesday morning. The crew is assembled. Equipment checks pass. The job briefing is complete. Everyone has signed the form.
Two hours later, an injury occurs during a switching operation — not because anyone was untrained, and not because anyone violated a rule they didn't know about. It happened because when the foreman laid out the work plan, one of the crew members had a question about communication sequencing between the hi-rail and the ground crew. He didn't ask it. The environment didn't feel like a place where slowing down to ask questions was the norm. So he defaulted to what he assumed, and the assumption was wrong.
This scenario — the unasked question, the assumption that turns out to matter — is among the most common threads running through rail incident reports. It shows up in the data with remarkable consistency.
The Numbers We Need to Confront
2018–2023
~565 per year
in 2023
Highest in recent years
in 2023
+7% from 2022
attributed to
human factors
Source: U.S. DOT Office of Inspector General Report ST2025029, "Federal Railroad Administration's Oversight of Roadway Worker Protection," May 14, 2025
Overall rail incidents reached 10,578 in 2023 — higher than in any of the immediately preceding years. Railroad deaths totaled 970, a 7% increase from 2022 and significantly above the levels recorded a decade ago.
And then there's the number that should reshape every safety conversation at your railroad: human factors account for nearly two-thirds of all rail accidents and are the leading cause of derailments. In rail yards specifically, over half of all incidents are attributed to human error.
These are not equipment failures. These are not acts of weather or infrastructure decay. These are preventable events — rooted in how people communicate, how they make decisions under pressure, and what the operating environment signals about whether it's safe to slow down and speak up.
What "Human Error" Actually Means in Rail Operations
The phrase "human error" appears with regularity in FRA accident reports and post-incident reviews. It is also, almost universally, incomplete as an explanation.
When an incident is labeled human error, what the label is actually pointing to is the last link in a longer chain:
- A roadway worker who was fatigued after back-to-back maintenance windows with inadequate crew rotation
- A conductor undertrained on a specific piece of track or equipment configuration they encountered for the first time under time pressure
- A car inspector under schedule pressure to turn a consist quickly who documented a concern but didn't escalate it because the response to previous concerns had been "we'll handle it"
- A signal maintainer who knew a communication protocol wasn't being followed correctly but didn't push back on a supervisor with twenty more years of seniority
- A crew member who knew the right procedure, was not confused about the rules, but operated in an environment where slowing down to follow them exactly felt riskier to their standing than the procedure shortcut itself
"Miscommunication. Fatigue. Failure to follow procedures. Shortcuts taken under time pressure. These aren't character flaws in your workforce. They are symptoms of a culture that hasn't been designed to catch them before they become incidents."
The distinction matters enormously for what you do next. If you read "human error" as a statement about individual failure, the response is more training, more enforcement, more accountability. If you read it as a statement about the operating environment, the response is a fundamentally different set of questions about what your culture signals to people when safety and speed are in tension.
The Data Visibility Problem
There is an additional complication embedded in the FRA data that most railroad leaders don't encounter unless they go looking for it.
As the DOT OIG report notes, the FRA's employee injury reporting form doesn't currently define people as on-duty roadway workers — making it difficult to filter safety data by the specific circumstances of each injury, including whether RWP procedures were in effect at the time. The broad trends are visible in the aggregate numbers. The granular picture — the actual story of what was happening in the moments before each incident — largely remains invisible to regulators.
That story is visible only to the people who were there. Which is precisely why those people need to feel safe enough to tell the organization what actually happened.
An operating environment where workers underreport near-misses, where incident investigations focus primarily on accountability rather than learning, and where people self-censor about what they observed doesn't just create a culture problem. It creates a data problem. The numbers the FRA sees are a floor, not a ceiling. The actual frequency of close calls and hazard conditions on your property is almost certainly higher than what gets documented.
The Inspection Gap: 32% Fewer Eyes on the Ground
In that same five-year period, total incidents and fatalities trended upward. The FRA has acknowledged the staffing constraints that drive the inspection gap — inspectors cover vast territories, and as the OIG report notes, roadway workers are inherently difficult to locate in the field during regular safety reviews. These are real operational constraints, not explanations that diminish the significance of the trend.
A railroad that waits for FRA inspection cycles to surface safety conditions is operating on a lag that the data shows is getting wider, not narrower. The organizations that will lead on safety in the coming decade are the ones building internal capacity to catch what external oversight increasingly cannot.
What the Data Demands: Three Unavoidable Conclusions
Human factors are the dominant safety risk, and they cannot be addressed through compliance alone. You cannot regulate your way to zero incidents when the root cause is how your crews communicate, how they make decisions under time pressure, and whether they feel safe raising concerns. These are not rulebook problems. They are culture problems.
The most preventable incidents are the most common ones. Slips, trips, and falls. Miscommunications during switching operations. Procedure shortcuts under schedule pressure. These are not dramatic, unpredictable failures. They are the daily accumulation of small risks in environments where speaking up feels harder — or more professionally risky — than staying quiet. The path to incident reduction runs directly through the culture that makes those small risks visible before they compound.
External oversight is decreasing as incidents increase. This is not an argument for ignoring regulatory requirements — it's a statement about where the real leverage is. The railroads that will differentiate on safety over the next decade are the ones that build internal self-correction capacity. Cultures where frontline crews feel empowered and expected to surface problems before they escalate — not because someone is watching, but because that's how this railroad works.
The Compliance vs. Culture Framework
| Safety Approach | Compliance-Driven Model | Culture-Driven Model |
|---|---|---|
| Primary Safety Tool | FRA inspections and regulatory requirements | Internal monitoring and peer accountability |
| Incident Response | "Who violated the rule?" | "What in our system allowed this?" |
| Worker Motivation | Fear of citation or discipline | Commitment to each other's safety |
| Data Quality | Underreporting due to fear of blame | High reporting rates because people believe it matters |
| Near-Miss Visibility | Low — conditions remain invisible until incidents | High — problems surface early, before they escalate |
| Adaptability | Reacts to regulatory requirements | Proactively identifies and addresses risks |
| Outcome | Meets minimum standards; incidents follow patterns | Prevents incidents before they reach the data |
Self-Assessment: How Does Your Railroad Compare to the Data?
Data-Driven Culture Assessment
Benchmark your safety culture against these indicators. Check all that apply:
- You track near-miss reporting rates as a leading indicator alongside incident rates — and the ratio is monitored by leadership, not just the safety department
- Your incident data shows year-over-year improvement that you can attribute to specific cultural or operational changes, not just statistical variation
- You have a reliable way of knowing whether your roadway workers, dispatchers, and yard crews feel comfortable raising safety concerns — through anonymous surveys, regular frontline conversations, or both
- Your post-incident reviews produce operational or cultural changes that frontline crews can see and point to
- You have processes to capture near-misses and "what almost went wrong" events before they appear in FRA incident data
- Your frontline crews — track supervisors, yardmasters, dispatchers — can articulate your safety priorities in their own words, not just cite the rulebook
- You track both lagging indicators (incidents, injuries, FRA violations) and leading indicators (near-miss reports, safety concerns raised, crew-initiated hazard identification)
Your Score: /7
Discuss Your Results With Our Team
Share your contact information and we'll reach out to walk through what your score means for your railroad — and where the highest-leverage opportunities are.
From Data to Action
If you're a COO, VP of Safety, or operations leader at a Class II or III railroad, these numbers are about your property. Your crews. The people who show up at 4 a.m. to inspect the consist, walk the track segment, or manage the dispatch board through a busy freight window.
The question the data raises isn't whether the industry has a safety problem — that's settled. The question is a more specific one: do your people feel safe enough to help you find the risks that exist right now on your railroad, before those risks find them?
If the honest answer is that you're not sure — that's the data point that matters most. And it's one you can actually do something about.
Your track crews are waiting to tell you what they see. Are you ready to listen?
Sources & References
- U.S. DOT Office of Inspector General Report ST2025029, "Federal Railroad Administration's Oversight of Roadway Worker Protection," May 14, 2025
- ASCE 2025 Infrastructure Report Card
- FRA Safety Data (publicly available datasets)
- Boston Consulting Group, research on psychological safety in the workplace