Ask workers “what could go wrong?” and they spot roughly half the hazards in their work area (Albert et al., 2017). The bottleneck is the question, not the person.
Energy-source methods find over 30% more hazards by swapping that open question for eight targeted prompts (Albert, Hallowell and Kleiner, 2014). Pressure, chemical, and radiation hazards show this gap most clearly. Recognition rates fall below 10% when no structured method is used (Perlman, Sacks and Barak, 2014).
Key findings
- Energy-source methods find over 30% more hazards than generic prompts in controlled field tests.
- Workers recognize roughly 47% of gravity, electrical, motion, and temperature hazards but under 10% of pressure, chemical, and radiation hazards.
- Length of experience does not predict hazard-finding ability; the method matters more than the person.
- Groups with diverse roles find a wider range of hazards than groups of experienced specialists.
- Learning from Normal Work’s Hazard Hunt+ produces a 50% increase in found hazards and a clear gain in description quality (Psychology Applied, 2022-2026).
Here is what the research shows and how to apply the energy-source method to a task your team has already assessed.
Why does the question matter more than the person?
The research points away from the common view that workers miss hazards because they lack focus.
A 2013 study of underground mining found that length of experience did not predict the ability to spot hazards (Bahn, 2013). Some of the longest-serving supervisors found few hazards. Groups with diverse roles found the most.
Across 3,000 hours of field study, workers could not spot over 55% of hazards in their work area (Albert et al., 2017). When the method changed from a generic question to an energy-source scan, finding rates improved by roughly 30%.
Learning from Normal Work builds on this through the Hazard Hunt+ tool. It replaces experience-based hazard spotting with a structured, repeatable method.
Structured prompts find more hazards regardless of experience. The method is the bottleneck, not the person.
The eight energy sources give that structure where experience alone falls short.
What are the eight energy sources?
The method gives each hazard type a name and a prompt. Hazard Hunt+ checks each energy source against each step of the task.
| Energy source | What to look for | Field example |
|---|---|---|
| Gravity | Falls, falling objects, collapse | Unsecured materials on a scaffold above the work area |
| Electrical | Contact with live circuits, arc flash | Worn cable insulation near a metal handrail |
| Chemical | Exposure to harmful substances | Welding fumes in an enclosed space without extraction |
| Mechanical | Pinch points, rotating parts, struck-by | Unguarded belt drive on a conveyor |
| Pressure | Hydraulic lines, pneumatic tools | Degraded hose on a pneumatic wrench |
| Thermal | Burns, heat stress, cold exposure | Uninsulated steam pipe at shoulder height |
| Radiation | UV, ionising, non-ionising, laser | Welding arc exposure to workers in adjacent bays |
| Biological | Bacteria, viruses, insects, mold | Legionella risk in stagnant cooling-water systems |
Workers can learn this method in a single session (Albert, Hallowell and Kleiner, 2014) because the energy types do the mental work that experience alone cannot. A structured prompt replaces the blank question with a guided scan.
Running this method step by step is straightforward.
How do you run it step by step?
Pick a task your team assessed recently. Hazard Hunt+ guides the scan with prompts for each energy source.
The process follows five steps:
- Walk through the task with the person who does the work. Divide it into five to eight steps.
- At each step, run through the eight energy sources. Ask how that energy could cause harm.
- Record the specific injury that could result, not the energy type label.
- Note what conditions make the harm more likely.
- Compare the results against the first assessment to see what was missed.
The key discipline is recording the specific harm pathway, not the type label.
“Pressure” on a risk assessment form tells the reader nothing they can act on. “Lung damage from breathing in welding fumes when the extraction unit sits behind the welder” tells the reader who is exposed and where the control has failed.
On one program, a team completed a routine task scan in 30 to 45 minutes (Psychology Applied, 2022-2026). Across clients, the data shows a 50% increase in found hazards and a shift from generic labels to detailed descriptions.
The gap between the two approaches shows up in what teams write down.
What does better hazard finding produce?
The quality of what teams record reveals the gap between generic prompts and the energy-source method.
On one program, a team assessed a routine warehouse task using their standard form (Psychology Applied, 2022-2026). The hazard list read “trip hazard” and “manual handling.”
When the same team re-assessed using the energy-source method, one entry changed from “trip hazard” to a far more specific finding.
“When a forklift enters the bay and the container door is open, it can hit the door, which closes with force and injures the worker.”
Another entry changed from “pressure” to “lung damage from breathing in welding fumes.” Details like these lead to better controls. They name the failure pathway rather than a label that fits any task.
A generic label tells the reader nothing they can act on. A specific finding names who is harmed and how.
The hazard types that generic prompts miss most are also the ones most relevant to process industries.
Which hazard types does the generic prompt miss most?
Workers spotted roughly 47% of gravity, electrical, motion, and temperature hazards but under 10% of pressure, chemical, and radiation hazards (Perlman, Sacks and Barak, 2014).
These missed types are the leading hazard sources in refineries and chemical plants. This gap helps explain why formally assessed tasks still produce incidents.
The Kleen Energy power plant explosion in 2010 killed six workers. Roughly 290,000 cubic feet of natural gas was released near active hot work during a routine gas-blow procedure (U.S. Chemical Safety Board, 2010).
The Chemical Safety Board called the practice “inherently dangerous,” yet it sat within an accepted assessment framework.
No single method can prevent every incident. An energy-source scan, though, would have flagged the chemical energy in the gas blow and the thermal energy from welding nearby. Hazard Hunt+ addresses this gap by directing attention to pressure, chemical, and radiation sources at every step.
Generic prompts miss the hazard types most relevant to process industries.
Group mix also shapes what the method finds.
How does group mix affect what you find?
Groups with diverse roles found a wider range of hazards than groups of specialists (Bahn, 2013).
Research on diverse problem-solving groups supports this: they can outperform groups picked purely for high ability (Hong and Page, 2004).
In practice, Learning from Normal Work recommends a specific team mix for the energy-source session.
- The person who does the task daily brings knowledge of how the work really happens.
- A colleague from a different shift or trade adds views the regular team has normalized.
- A supervisor or safety lead runs the structured prompts and keeps the scan on track.
The facilitator runs the energy-source scan while the team brings different views to each prompt. When both elements are in place, hazard finding extends beyond the hierarchy of controls into the conditions that shape how work is done.
Try it this week
Pick a routine task your team assessed in the last month. Re-run it with the energy-source method: break the task into steps and check each energy source at each step. Record the specific harm rather than the label.
Compare the two lists side by side. Look for pressure and chemical sources that the first assessment labeled broadly or missed. This is the core practice behind Learning from Normal Work’s Hazard Hunt+ tool.
Understanding why signed-off risk assessments still miss critical hazards starts with seeing that the question you ask decides what you find.
References
Albert, A., Hallowell, M.R. & Kleiner, B. (2014). Enhancing construction hazard recognition with energy-based cognitive mnemonics. Journal of Construction Engineering and Management, 140(2), 04013042. https://doi.org/10.1061/(ASCE)CO.1943-7862.0000790
Albert, A., Hallowell, M.R., Skaggs, M. & Kleiner, B. (2017). Empirical measurement and improvement of hazard recognition skill. Safety Science, 93, 1-12.
Bahn, S. (2013). Workplace hazard identification and management: the case of an underground mining operation. Safety Science, 57, 129-137. https://doi.org/10.1016/j.ssci.2013.01.010
Hong, L. & Page, S. (2004). Groups of diverse problem solvers can outperform groups of high-ability problem solvers. Proceedings of the National Academy of Sciences, 101(46), 16385-16389.
Perlman, A., Sacks, R. & Barak, R. (2014). Hazard recognition and risk perception in construction. Safety Science, 64, 22-31.
U.S. Chemical Safety Board. (2010). Investigation Report: Kleen Energy Natural Gas Explosion. Report No. 2010-7-I-CT.
Psychology Applied. (2022-2026). Implementation data: Measured outcomes from structured hazard identification across enterprise clients.
