Part 39 min

3.3 Hours worked: the denominator nobody checks

Everybody argues about the numerator. Nobody checks the denominator. Yet the hours worked figure is the input most often estimated, most often inherited from last year, and most often quietly wrong — and because it sits under every rate you publish, an error there moves all four at once, in the same direction, invisibly.

1 Hours actually worked

The 300A asks for the total hours worked by all employees covered by the log. Worked is the operative word, and it draws a hard line through payroll.

What belongs in the denominator
Include	Exclude
Straight time actually worked	Vacation, even when paid
Overtime, at actual hours — not at time-and-a-half value	Sick leave, even when paid
Hours of salaried staff covered by the log	Holidays, even when paid
Part-time, temporary and seasonal employees	Any other paid non-work time
Temporary help service workers under your day-to-day supervision, if their cases go on your log	Hours of people whose cases go on somebody else's log

The overtime point catches more people than the leave point. Overtime goes in as hours, not as pay units. A hundred hours of overtime paid at time and a half is a hundred hours in the denominator, not a hundred and fifty.

2 The consistency rule that matters more than the rule

Whichever population your cases come from, the hours have to come from the same population. Count a temporary worker's injury on your log and you must count that worker's hours in your denominator. Leave the case off and leave the hours off too. Doing one without the other is the fastest way to a rate that is out by a third and defended by nobody.

On the job: write the rule down once, in a paragraph, and file it with the 300A worksheet: which populations are in, where their hours come from, and who produces the number. It takes ten minutes and it settles the same argument every February for a decade.

3 When you do not have actual hours

Salaried staff and small businesses without a time-tracking system are the common cases. The OSHA forms package includes a worksheet for exactly this: a method for figuring the annual average number of employees and the total hours worked for the establishment. Use it, keep the completed sheet with the summary, and write down the basis.

A defensible estimate for a full-time salaried employee starts at 2,000 hours a year — 40 hours a week for 50 weeks — and is then adjusted for what you know: part-year hires, unpaid leave, a four-day week, regular overtime. What makes it defensible is not the number. It is that the method is written down and applied the same way next year.

1
Pull actual hours from payroll or the job cost system for everyone who is tracked.
2
For everyone who is not, apply a written estimating rule and show the arithmetic.
3
Add the two, and record the split — actual hours and estimated hours — on the worksheet.
4
File the worksheet with the certified 300A and use the same method next year.

4 Establishment or company: pick one and say so

1904.30(a) keeps the log per establishment, so the 300A hours are establishment hours and the rate they produce is an establishment rate. A company-wide rate is a legitimate management figure — it is just a different one, produced by adding every establishment's cases and every establishment's hours.

The problem is never the calculation. It is quoting a company-wide rate into a form that asks for an establishment rate, or comparing a company-wide rate against a BLS benchmark that is measured per establishment. Both happen routinely in contractor prequalification, and both make the comparison meaningless in whichever direction happens to flatter you.

Common mistake: a contractor with twelve sites reporting a single company TRIR against an establishment-level industry average, and being told it is doing well. It might be. The comparison does not show it.

5 Contractor hours on a jobsite

On a multi-employer site, each employer records its own people's injuries and counts its own people's hours. That is the default and it is the right one, because each employer keeps its own log.

General contractors frequently ask for something different in their own reporting: total site hours, all trades combined, against total site recordables. That is a site metric, not an OSHA rate, and it is perfectly reasonable — as long as everyone knows which is being discussed. The sentence that prevents the confusion is short: “This is a site figure covering all employers and all trades; our own establishment rate is X.”

6 An audit you can run in an afternoon

Once a year, before you certify the summary, run four checks on the denominator. None of them takes long and together they catch nearly every error that matters.

Reconcile to payroll. Total hours should be explainable from gross payroll hours minus paid leave. If the gap is more than a couple of percent, find out why.
Check the overtime treatment. Pull one heavy overtime week and confirm the hours went in at actual, not at premium value.
Check the population. List who is in the denominator and compare it with whose cases went on the log. They must match.
Compare with last year per head. Divide hours by the annual average number of employees. A number that jumps from 1,900 to 2,400 without a change in working pattern is a methodology change nobody documented.

Key takeaways

Hours actually worked: overtime in at actual hours, paid leave out entirely
The population of hours must match the population of cases — both or neither
Where hours are not tracked, estimate with a written method and file the worksheet
An establishment rate and a company rate are different figures; name which you are quoting
On a multi-employer site, a total-site figure is not an OSHA rate — say so when you use one
Reconcile hours to payroll once a year, before the summary is certified

Free educational content — not OSHA-authorized training, no certificate or card issued. Follow your employer's program and the standards cited.