A job hazard analysis, step by step, with a worked example

The four steps

1. Pick the job. Start with tasks that have a history of injuries, new tasks, and tasks where one mistake is severe.
2. Break it into steps. Write the steps in the order the crew does them. Keep each one short, a verb and an object.
3. Find the hazard in each step. Ask what can hurt someone, how, and who is exposed.
4. Set the controls. Pick the strongest practical control for each hazard, then name who owns it.

Keep the steps and the hazards in separate columns. Crews skim a table far faster than a paragraph.

Worked example: laying pipe in a trench

All numbers here are invented for practice. The job: a crew lays a pipe section in a trench about 6 feet deep.

Step	Hazard	Likelihood	Severity	Control
Mark and call before digging	Striking a buried line	Medium	High	Locate and mark utilities first, hand dig near marks
Excavate	Equipment tipping or striking people	Low	High	Keep workers out of the swing and travel zone, use a spotter
Enter the trench	Cave-in	Medium	High	Protective system in place, competent person inspects first
Lower pipe	Pinch or crush from the load	Medium	High	Rated rigging, no one under the load, tag line
Backfill	Heavy equipment near the edge	Low	High	Keep spoil back from the edge, stop-and-check at the lip

For the cave-in line, the rule that governs is 1926.652, which requires a protective system for trenches at the depth in this example unless the walls are in stable rock, and the daily inspection duty sits in 1926.651(k). Read both in your own copy. The excavations guide explains the subpart, and the trench sloping calculator turns soil type and depth into a slope.

Rank the risk

A simple grid keeps you honest. Score likelihood and severity from 1 to 3 and multiply. Anything at 6 or above needs a control before work starts. Anything at 9 needs a stronger control than a warning sign. In the table above, the cave-in step scores highest, so it gets the strongest control, a protective system, and the inspection comes first.

Pick controls from the top of the hierarchy

For each hazard, work down the list: eliminate, substitute, engineer, administer, then protective equipment. In the example, you cannot eliminate the need to enter the trench, but you can engineer a protective system and use a trench box so workers are shielded. Hard hats and reflective vests are the last layer, not the plan.

Common mistakes

• Writing hazards as generic words. "Falling" says nothing. "Worker falls from the edge while reaching for a pipe" tells the crew what to watch for.
• Skipping the steps that feel routine. The routine steps are where habits form.
• Writing it once and filing it. A JHA is for the crew on that day, in those conditions. Review it when the job, weather, equipment or crew changes.
• Not telling the crew. A JHA no one has heard is paperwork. Walk through it in the morning talk.

How the CHST tests it

Expect scenario questions that ask for the first step, the best control or what the analysis missed. The pattern is a situation with a hazard and four plausible actions. The strongest answer usually works higher on the hierarchy and acts before exposure. Practice by writing your own JHA for a task you know, then for one you have never done. The unfamiliar one is the better exam preparation. Pair it with the study guide week plan.

Try it on the exam style

Take the free 10-question check and watch for the questions that ask what to do first or what control is strongest. Every answer is explained and it takes about ten minutes.

FAQ

Is a JHA required by OSHA?

OSHA publishes guidance on job hazard analysis, but the construction standards do not set one universal JHA form. Many employers and owners require them by contract or policy.

What is the difference between a JHA and a JSA?

Most employers use the two terms for the same thing: job hazard analysis or job safety analysis. Check what your employer's program calls it.

How many steps should a JHA have?

As many as the job has. If you have more than about ten, split the job into two analyses.