How to calculate the average number of employees for the OSHA 300A (and total hours worked)

Every OSHA Form 300A has a small box headed "Employment information" with two blanks in it: "Annual average number of employees" and "Total hours worked by all employees last year." They look like the easy part of the form. They are also the two figures most often filled in from memory, and neither is free-form. OSHA prints a method for each, and the hours figure becomes the denominator of every injury rate anyone calculates from your records.

The form points you to that method on the box itself: "If you don't have these figures, see the Worksheet on the next page to estimate." This guide works through that worksheet line by line. It includes a fill-in version you can copy, a worked example for a seasonal business, and the situations its single page doesn't spell out. For the rest of the form (the case totals, certification and posting), see our complete guide to the OSHA 300A.

What the 300A actually asks for

The requirement is in 29 CFR 1904.32(b)(2)(ii). When you complete the annual summary, you must "Enter the calendar year covered, the company's name, establishment name, establishment address, annual average number of employees covered by the OSHA 300 Log, and the total hours worked by all employees covered by the OSHA 300 Log."

Two phrases in that sentence do most of the work:

• "Annual average." Not your headcount on December 31, not your busiest week, and not the number of people who passed through payroll during the year. It is an average across the year, computed the way the worksheet below describes.
• "Covered by the OSHA 300 Log." Both numbers describe the same group: the people whose injuries go on that establishment's log. Who that is comes from 29 CFR 1904.31, and that one definition settles most of the awkward cases later in this guide. (If you're new to the log itself, start with how to fill out the OSHA 300 Log.)

The hours figure matters well beyond the form. OSHA's own page on calculating incidence rates names "The number of hours all employees actually worked during the year" as the denominator, and sends you to the same place to find it: "Refer to OSHA Form 300A and optional worksheet to calculate this number." Every TRIR and DART rate built from your records is divided by this number, which is why our TRIR and DART guide spends so long on it.

The annual average number of employees, step by step

The page in OSHA's recordkeeping forms package is titled "Worksheet to Help You Fill Out the Summary," and it is marked optional. It opens with a shortcut, then gives the formula.

The shortcut for a steady headcount

"If you pay about the same number of employees every pay period throughout the year (e.g., about 100), then you can use that number as your annual average employment." If your headcount moves around ("e.g., your business is seasonal or your establishment grew or shrunk during the year"), the worksheet says to use the formula instead.

OSHA's four steps

1. Add up the employees paid in each pay period. The worksheet: "Add up and then enter the number of employees your establishment paid IN EACH PAY PERIOD during the year. Be sure to include all employees: full-time, part-time, temporary, seasonal, salaried, and hourly."
2. Count the pay periods. "Count and then enter the number of pay periods your establishment had during the year. Be sure to include any pay periods when you had no employees. For example, enter 26 if you have biweekly pay periods or 52 if you have weekly pay periods."
3. Divide the step 1 total by the step 2 count.
4. Round up. "Round the answer to the next highest whole number," then write it in the blank on the 300A marked Annual average number of employees.

OSHA's own illustration is a construction company, Acme, whose biweekly headcount ran from zero in one pay period to 40 in another. Its 26 pay periods add up to 830 employees paid, and 830 ÷ 26 = 31.92, which rounds to 32.

The one method OSHA rules out by name

The worksheet carries a boxed note: "You CANNOT divide the total number of W2s by the number of pay periods to calculate average employment. You must add up the number of employees paid IN EACH PAY PERIOD and then divide by the number of pay periods." A W-2 count includes everyone who was paid at any point in the year. For any business with turnover, it measures traffic through the payroll, not how many people were working at once.

The worksheet also gives three questions to test the answer: "Is it about the same as the number of employees working at your establishment on any given day? Is it bigger than your smallest number of employees in a pay period? Is it smaller than your biggest number of employees in a pay period?" If any answer is no, "the calculation may be incorrect." The first question is a loose fit for a seasonal business, which has no typical day, but the second and third still catch most mistakes.

A worked example: a seasonal landscaping company

Pine Hollow Landscaping is a fictional company: one location, a biweekly payroll, owned by a sole proprietor. From January into early March, and again in December, the owner runs the office alone. Crews come on in spring, peak in midsummer and are let go by Thanksgiving. A salaried crew manager is on the payroll from pay period 5 through pay period 24, and during the summer peak the company brings in agency temps whom its crew leads supervise day to day. The next section explains why the temps are counted and the owner isn't.

Here is the year, pay period by pay period:

• Step 1: the employees paid across all pay periods add up to 447.
• Step 2: there are 26 pay periods, six of them with nobody on the payroll.
• Step 3: 447 ÷ 26 = 17.19.
• Step 4: round up to 18. That's the number for the 300A.

OSHA's checks pass. 18 is bigger than the smallest pay period (0) and smaller than the biggest (48).

The same year produces several wrong answers just as easily:

The cases the worksheet doesn't spell out

Temps you supervise day to day

The worksheet's first step counts "employees your establishment paid," and an agency temp is paid by the agency. The regulation's wording settles it, though. Both 300A figures cover employees "covered by the OSHA 300 Log," and 1904.31(b)(2) puts supervised temps on your log. If you obtain workers from a temporary help service, employee leasing service or personnel supply service, "You must record these injuries and illnesses if you supervise these employees on a day-to-day basis." The worksheet's hours section points the same way, telling you to include hours "worked by other workers subject to day-to-day supervision by your establishment (e.g., temporary help service workers)."

Read together, temps you supervise day to day belong in both numbers. Temps the agency supervises belong in neither, because their injuries go on the agency's log. Pine Hollow counts its summer temps in the pay periods they worked, using the agency's invoices or timesheets for the headcount and the hours. Our guide to temp and staffing workers under 1904.31 covers who counts as supervising whom.

Owners, partners and contractors

Some people who work at your site are not employees for recordkeeping purposes, so they stay out of both numbers. Under 1904.31(a), "If your business is organized as a sole proprietorship or partnership, the owner or partners are not considered employees for recordkeeping purposes." That is why Pine Hollow's winter pay periods count as zero even though the owner was working. Self-employed individuals are out too ("self-employed individuals are not covered by the OSH Act or this regulation," 1904.31(b)(1)). So is a contractor's employee who works under the contractor's day-to-day supervision, because the contractor records that person's injuries (1904.31(b)(3)).

Two payroll schedules

The worksheet assumes one run of pay periods. Plenty of small employers run two, often weekly for hourly staff and semimonthly for salaried staff. OSHA's worksheet doesn't address that. One approach that keeps its logic, which is our suggestion rather than OSHA text, is to average each group over its own pay periods: the weekly group's total over 52, the semimonthly group's total over 24. Then add the two averages and round up once, at the end. Whatever you do, write down how you did it.

More than one location

The 300A is a per-establishment form. Under 1904.30(a), "You must keep a separate OSHA 300 Log for each establishment that is expected to be in operation for one year or longer." The summary carries that establishment's name and address, so each one needs its own average and its own hours. A company-wide payroll report has to be split by location before you start. The details are in our guide to multi-establishment recordkeeping.

Total hours worked: hours worked, not hours paid

What goes in and what stays out

The worksheet: "Include hours worked by salaried, hourly, part-time, and seasonal workers, as well as hours worked by other workers subject to day-to-day supervision by your establishment (e.g., temporary help service workers). Do not include vacation, sick leave, holidays, or any other non-work time, even if employees were paid for it."

That last clause is where most payroll exports go wrong. Payroll systems report hours paid, and hours paid include paid time off. Take the figure from your time records, or start from hours paid and subtract the vacation, sick-leave and holiday hours.

When you only have hours-paid records

The worksheet covers this case directly: "If your establishment keeps records of only the hours paid, or if you have employees who are not paid by the hour, please estimate the hours that the employees actually worked. If this number isn't available, you can use this optional worksheet to estimate it." The optional hours worksheet runs:

1. "Find the number of full-time employees in your establishment for the year."
2. "Multiply by the number of work hours for a full-time employee in a year." The worksheet labels the result "the number of full-time hours worked."
3. "Add the number of any overtime hours as well as the hours worked by other employees (part-time, temporary, seasonal)."
4. "Round the answer to the next highest whole number," and write it in the blank marked Total hours worked by all employees last year.

The worksheet doesn't fix the full-time figure in step 2; you use your own schedule. For reference, OSHA's rate formula is built on a conventional full-time year: "The 200,000 figure in the formula represents the number of hours 100 employees working 40 hours per week, 50 weeks per year would work." That works out to 2,000 hours a person.

Pine Hollow's hours

• Hourly crews: 29,540 hours worked, from the time clock. Payroll showed 30,150 hours paid; the 610-hour difference is paid holidays and PTO, which stays out.
• Agency temps the crew leads supervised: 2,860 hours, from the agency's timesheets.
• The salaried crew manager doesn't clock in, so Pine Hollow estimates 40 weeks at 45 hours: 1,800 hours.

Total: 29,540 + 2,860 + 1,800 = 34,200 hours.

A quick check on the two numbers together

Divide the hours by the average: 34,200 ÷ 18 = 1,900 hours per average employee, about a full-time year. A result above 8,760, the number of hours in a 365-day year, means at least one of the two figures is wrong. So does a result far below what your mix of full-time and part-time schedules could produce.

Pine Hollow's log has two recordable cases for the year. On 34,200 hours, its TRIR is 11.70 (2 × 200,000 ÷ 34,200). Suppose it had estimated hours from its busiest pay period instead: 48 people × 2,000 hours = 96,000. The same two cases would then show a TRIR of 4.17. That's well under half the real rate, from the denominator alone.

Three headcounts that are easy to mix up

The 300A average is one of three employee counts in Part 1904, and each answers a different question:

OSHA's own FAQ for the electronic submission says it outright: "you need to determine the establishment's peak employment during the last calendar year." Pine Hollow shows why the difference matters. Its 300A average is 18, under the 20-employee line, but it had 48 employees at its peak. If its industry is on Appendix A to Subpart E, it has to submit its 300A through OSHA's Injury Tracking Application by March 2, even though the number on its form starts with a 1. Our ITA submission guide covers that process, and the free ITA checker asks for the peak for this reason. The size exemption is in our guide to the recordkeeping exemptions, and how all three fit with the rest of Part 1904 is in our overview of OSHA recordkeeping requirements for small employers.

The worksheet, ready to fill in

Copy this into a spreadsheet or print it, one copy per establishment.

Part A: annual average number of employees

Part B: total hours worked

Keep the filled-in worksheet with that year's records, so you can explain how you got the two numbers if anyone asks.

Where the two numbers go

Once the 300A is certified and posted, from February 1 through April 30, the same two figures travel. The hours become the denominator of your TRIR and DART. If your establishment has to submit electronically, both go to OSHA with the rest of the 300A by March 2.

LogStead doesn't read your payroll, so these two numbers still come from a worksheet like the one above. What it does is keep them straight. Each establishment carries its annual average, its peak employment and its total hours as three separate figures. The 300A uses the average and the hours, TRIR and DART are computed on the hours you entered with the formula shown, and the electronic-submission check uses the peak, not the average. The case totals are tallied from the log itself, so the employment figures are the part you supply. To see what your hours do to your rates first, the free rate calculator runs TRIR and DART and compares them with the BLS average for your industry.

Two blanks, one method for each, and a single rule underneath both: count the people whose injuries go on your log, the way OSHA's worksheet counts them.

This post is general compliance information, not legal advice. Verify current regulatory text against eCFR and your state plan's requirements.