Staff Injury Rate in a Trampoline Park: 6.4 per 100 Workers
The US Bureau of Labor Statistics puts amusement and theme parks at 6.4 recordable injuries per 100 full-time workers. Amusement arcades sit at 1.3. Fitness and recreational sports centers sit at 1.8.
A trampoline park is all three of those things under one roof, and the payroll clerk picks one code for the whole building.
That choice is not paperwork. It sets the benchmark your safety record gets compared against, and it sets the class rate your workers compensation premium is built on. The gap between the codes is wider than most operators expect: 6.4 is 4.9 times the arcade figure and 3.6 times the fitness figure.
This article works through four numbers. The published rate. The number of cases that rate implies at a given headcount. What one case costs in wages. And what the same case costs once it reaches an experience modifier.
The short version: the wage cost of a staff injury is small enough to be rounding, and the premium cost is not. Operators budget for the wrong one.
1. Two Codes, One Building
The BLS Survey of Occupational Injuries and Illnesses reports incidence rates per 100 full-time equivalent workers. The 2024 reference year was published on 22 January 2026, and the detailed industry table is public.
Six lines from that table matter to anyone building a park. All figures are cases per 100 full-time workers.
| Industry (NAICS) | Total recordable | DART | Days away from work | Job transfer or restriction |
|---|---|---|---|---|
| Amusement and theme parks (71311) | 6.4 | 4.8 | 1.5 | 3.3 |
| Amusement arcades (71312) | 1.3 | 0.5 | Not published | 0.4 |
| Fitness and recreational sports centers (71394) | 1.8 | 0.9 | 0.5 | 0.4 |
| Child day care services (6244) | 1.9 | 1.2 | 0.8 | 0.4 |
| Amusement, gambling and recreation (713, all sub-industries) | 3.3 | – | – | – |
| All private industry | 2.3 | – | – | – |
DART means cases involving days away from work, job transfer, or job restriction. It is the number underwriters read first, because it removes the cases that only needed a bandage.
Amusement and theme parks carry a rate 2.8 times the private-industry average. That is not a surprise once you look at what the code contains: ride operators, maintenance crews, food stands, and crowds standing for hours on wet or uneven floors.
The arcade figure is the interesting one. At 1.3, a building full of cabinets and prize counters looks like an office park on paper. The physical risk has been engineered out of the equipment and into the floor staff’s back and shoulders.
A trampoline park with 12 courts and a 4,000 sq ft (372 m²) arcade sits in both worlds. Its injury profile looks like 71311. Its payroll department may file it as 71312 or as 71394, because the classification follows the principal activity described on the policy, not the riskiest thing in the room.
2. What the Split Between Days Away and Restricted Duty Says
Total recordables hide a second decision. Of the 4.8 DART cases per 100 workers in amusement and theme parks, only 1.5 involved days away from work. The other 3.3 were job transfer or restriction cases, meaning the person came back on light duty.
That is 69 percent of DART cases held inside the building rather than sent home. Fitness centers run at 44 percent and child day care at 33 percent.
| Industry | DART cases | Restricted-duty cases | Share held on light duty | Share with days away |
|---|---|---|---|---|
| Amusement and theme parks | 4.8 | 3.3 | 69% | 31% |
| Amusement arcades | 0.5 | 0.4 | 80% | Not published |
| Fitness and recreational sports centers | 0.9 | 0.4 | 44% | 56% |
| Child day care services | 1.2 | 0.4 | 33% | 67% |
The calculation is the restricted-duty rate divided by the DART rate. For amusement parks, 3.3 divided by 4.8 equals 0.688. For fitness centers, 0.4 divided by 0.9 equals 0.444.
Two readings are possible and the data cannot separate them. The first is that parks are good at keeping injured staff on the floor: greeters, ticket scanning, party hosting, and stock work absorb a sprained ankle that would end a shift at a gym. The second is that a young seasonal workforce reports strains that a career fitness instructor works through without filing anything.
Both readings point at the same operating requirement. A park needs a written list of light-duty tasks before it needs another safety poster, because 69 percent of its DART cases will need somewhere to go. Without that list, a 3.3 restricted-duty rate turns into a days-away rate, and days-away cases are what move a modifier.
One caution on the arcade line: 0.5 DART cases per 100 workers is a small-count estimate, and BLS suppresses the days-away figure because the sample cannot support it. Treat the 80 percent share as directional, not as a target.
3. From a Rate to Cases at a 40-FTE Park
Rates are hard to act on. Case counts are not. The conversion is straightforward, and it is where most operators discover that their headcount is not their FTE count.
Start with FTE, not bodies. A park with 80 part-time staff at 20 hours a week carries 40 FTE. A park with 45 full-time staff plus 60 weekend staff on 8-hour shifts carries close to 57.
The formula is FTE multiplied by the rate, divided by 100.
Expected recordable cases per year = FTE x rate / 100 40 FTE x 6.4 / 100 = 2.56 recordable cases 40 FTE x 1.8 / 100 = 0.72 recordable cases
The difference between those two lines is 1.84 cases a year, produced entirely by which code the payroll sits under.
| Headcount | Recordable cases | DART cases | Days-away cases | Restricted-duty cases |
|---|---|---|---|---|
| 25 FTE | 1.60 | 1.20 | 0.38 | 0.83 |
| 40 FTE | 2.56 | 1.92 | 0.60 | 1.32 |
| 60 FTE | 3.84 | 2.88 | 0.90 | 1.98 |
| 100 FTE | 6.40 | 4.80 | 1.50 | 3.30 |
Two things follow from the table. The first is that a 40-FTE park is expected to log between two and three recordables a year. If it logs five, the safety program is not the only explanation: the mix of tasks, the age of the staff, and the reporting culture all feed the number.
The second is that below 25 FTE the rate stops being a management tool. One case at 16 FTE equals 6.25 per 100, which is the industry average, and one case at 8 FTE equals 12.5. At that size you count cases, you do not compute rates.
The reciprocal is worth remembering: 100 divided by 6.4 equals 15.6. One recordable per 15.6 full-time workers is what the industry average looks like in plain language.
4. What One Recordable Costs in Wages
Here is the part that gets overbudgeted. Take a days-away case. The BLS case-characteristics data for 2023-2024 puts the median at 8 days away from work across private industry. The median for job transfer or restriction cases is 15 days.
Those hours are paid at the median wage for amusement and recreation attendants, which O*NET reports at $15.46 an hour from 2025 BLS wage data. Then there is a coverage premium, because the shift still has to be covered.
| Cost line | Calculation | Amount |
|---|---|---|
| Wages for days away | 8 days x 8 hours x $15.46 | $989 |
| Coverage at 1.5x | 64 hours x $15.46 x 0.5 | $495 |
| Total per days-away case | Direct wage plus coverage premium | $1,484 |
| Total per restricted-duty case | 15 days x 8 hours x $15.46, no replacement | $1,855 |
| Annual exposure at 40 FTE | 0.60 x $1,484 plus 1.32 x $1,855 | $3,339 |
| As a share of payroll | $3,339 against 40 x 2,080 hours x $15.46 | 0.26% |
Payroll for 40 FTE at the median wage is $1,286,272 a year. The wage cost of the injuries that rate predicts is $3,339, or about 0.26 percent of payroll.
That number is real and it is also irrelevant to the decision. Three reasons. First, the median wage is not your wage; a park paying $19 an hour sees 23 percent more. Second, the 8-day median is a private-industry figure, not an amusement-park figure, and BLS does not publish a park-specific median. Third, and this is the one that matters, the wage line is not where an injury is expensive.
Where this model came from: the rate is from the BLS 2024 incidence table, the day counts are the 2023-2024 private-industry medians, and the wage is the O*NET median for amusement and recreation attendants. The coverage multiplier of 1.5 is an assumption you should replace with your own overtime rate, and the payroll total is arithmetic on the same wage.
5. Where the Money Actually Goes
Workers compensation premium is not priced off your injury rate directly. It is priced off manual premium, adjusted by an experience modifier that compares your losses with the losses expected for a business of your size and class.
The modifier is retrospective in a way that catches new operators. Claims count for three policy years. A single bad year follows you into two renewals after the year itself, which means the cheapest month to fix a task is the month before the claim, not the month after.
The arithmetic is simple enough to run on a quote. Every 0.10 of modifier applied to $60,000 of manual premium is $6,000 a year, and across the three-year look-back it is $18,000. Set that against the $3,339 of direct wage cost from the table above and the ratio is about five to one.
The $60,000 is a placeholder. Ask your broker to substitute your manual premium and your expected loss rate, then rerun it. The point that survives the substitution is the shape: premium moves in thousands, wages move in hundreds.
Two details make the modifier move further than operators expect.
Classification follows the policy. If the building is rated as a fitness center at 1.8, a year of claims that would be unremarkable at 6.4 becomes conspicuous against the expected losses for 1.8. The lower rate is not a discount. It is a higher bar.
Restricted duty counts. A transfer or restriction case with no days away still enters the loss run. It costs less per case than a days-away claim, but it appears in the count, and the count is what a frequency-driven modifier reacts to. That is why the 69 percent light-duty share in Section 2 is not automatically good news.
6. What the Data Does and Does Not Show
Four limits belong in front of any decision made from these rates.
Limit 1: there is no trampoline park code NAICS has no line for trampoline parks. A park files as 71311 amusement and theme parks, 71312 arcades, or 71394 fitness and recreational sports centers depending on how its principal activity is described. The published rate for 71311 also contains theme parks, water parks, and ride operators, so 6.4 is an average over a broader population than your building.
Limit 2: rates are per worker, not per guest Nothing here converts to guest injuries. The incidence table counts employees, and a park with 1,200 visitors on a Saturday contributes one exposure unit per full-time employee, not per visitor. Guest injury data comes from a different source entirely, and mixing the two produces nonsense.
Limit 3: the survey undercounts BLS has funded research showing the Survey of Occupational Injuries and Illnesses misses cases inside its own scope. The size of the miss depends on the method used. Recordable counts are a floor, not a census, and small establishments report with more noise than large ones.
Limit 4: the median days are national The 8-day and 15-day medians cover all private industry in 2023-2024, not amusement parks. A park-specific median is not published. If your staff do heavier maintenance work than a typical park, the days will run longer than 8.
One thing the data does establish: NAICS 7131 does not appear in Appendix A to Subpart B of 29 CFR 1904, the list of partially exempt industries. Neither does 71394. So once an establishment passes the 10-employee threshold, the OSHA 300 log applies. That matters twice over: the log is the document an inspector reads, and it is the document your modifier is built from.
The honest conclusion from the published numbers is a comparison, not a forecast. A park rated as amusement and theme parks should expect more than twice the recordables of an identical park rated as a fitness center, for reasons that have nothing to do with how well either one is run.
7. If the Code Is Wrong, the Benchmark Is Wrong
The most common classification problem in this sector is not fraud. It is drift. A park opens with courts and a small arcade, gets rated as an amusement facility, then adds a ninja course, a climbing wall, a party floor, and a kitchen. Two years later the policy still describes the building that opened.
The reverse happens too. A soft play center with two small trampolines is described as an arcade because the original broker wrote it that way, and the owner then wonders why the loss run looks bad against a 1.3 benchmark.
Both directions have a cost, and they are not symmetrical. Filing too low raises your expected-loss ratio and pushes the modifier up faster for the same claims. Filing too high means paying manual premium on a class that may overstate your exposure, which is money out of the door every month regardless of claims.
There is no published rule that says which code a trampoline park must use. The classification is a negotiation between your broker, the carrier, and the state rating bureau, and it is documented in the policy declarations rather than in the statute.
Ask for the classification in writing at binding, with the class code and the description, and check it again at each renewal. It is one page, and it decides which of the industry lines in Section 1 you are measured against.
8. Decision Rules
The point of the tables is to give you a rule you can apply without a consultant. Five of them follow.
| If this is your situation | Read it as | Do this |
|---|---|---|
| Under 16 FTE | One case equals the industry average | Track case counts and tasks, ignore the rate |
| 40 FTE and no written light-duty list | 3.3 restricted-duty cases per 100 will have nowhere to go | Write the list before the next hire |
| Policy rated 71394 but the floor is courts | You are measured against 1.8, not 6.4 | Get the class code and description in writing |
| More than one days-away case a year at 40 FTE | Above the 1.5 per 100 pattern | Read the 300 log by task, not by person |
| Two claims from the same task in one year | Frequency problem, not a training problem | Change the task or the equipment |
The fourth rule is the one operators resist. Reading the log by person produces blame; reading it by task produces a fix. A park that logs three strains in six months across three different attendants has one problem, and it is usually a mat edge, a trampoline bed that has lost tension, or a lifting point with no trolley.
Equipment condition belongs in that conversation. A bed that has gone soft changes the landing and the effort required to get back up, and the maintenance schedule that keeps tension in spec is cheaper than the claim it prevents. The equipment list behind a 22,000 sq ft park layout sets how many beds one attendant inspects per shift, which is a staffing number before it is an engineering one.
Supervision ratios belong here too, for a different reason. A park that staffs to a court ratio puts more people on the floor at peak, and more people on the floor means more staff exposure hours in the busiest part of the building. The ratio that answers a legal question can raise an insurance one. The supervision ratio a court named is a benchmark, not a floor plan.
9. Three Moves Before Your Next Renewal
1. Get the class code in writing. Ask for the declarations page and find the code, the description, and the manual premium. If the description says arcade and the building has twelve courts, you have found the cheapest correction available this year. Our safety standards page lists the documentation we hand over for a compliance file, which is the same paperwork a carrier asks for.
2. Run the three-year loss run against 6.4. Divide your recordable count by FTE, multiply by 100, and compare. If you are under 40 FTE, compare counts instead. Anything above the published rate for your class needs a task-level explanation before it needs a new incentive scheme.
3. Build the light-duty list. One page, ten tasks, each with a stated restriction. Party hosting, front desk, stock, cleaning, and inspection rounds absorb most upper-body restrictions. Without it, your 3.3 restricted-duty cases become days-away cases, and those are the ones that set next year’s modifier.
If you are at the design stage rather than the renewal stage, the sequence runs the other way: the equipment plan decides the task list, the task list decides the light-duty options, and the options decide what your loss run looks like in year two. The design and quoting process is where that starts, and the trampoline range shows what the inspection load per court actually is.
A park with clean mats, documented tension checks, and a written light-duty list still logs injuries. It logs fewer of them as days away, and that is the line the modifier reads.
FAQ
What is a good OSHA recordable rate for a trampoline park?
Use 6.4 per 100 full-time workers as the reference point, because that is the published rate for amusement and theme parks. Anything below that is better than the industry average, and anything above it needs a task-level explanation. Below 16 FTE the rate is too noisy to compare, so count cases instead.
Does a trampoline park have to keep an OSHA 300 log?
Yes, once the establishment has more than 10 employees. NAICS 71311, 71312, and 71394 are all absent from the partially exempt industry list in Appendix A to Subpart B of 29 CFR 1904, so the recordkeeping duty applies. Fatality and severe injury reporting applies regardless of size.
Which NAICS code should a trampoline park use for workers compensation?
There is no published rule. The code is set by the classification on the policy, agreed between your broker, the carrier, and the state rating bureau. What matters is that the description matches the building, because the code sets the expected-loss benchmark your modifier is measured against.
How many staff injuries should a 40-full-time-equivalent park expect in a year?
At 40 full-time equivalent workers, the published rate implies 2.56 recordable cases, of which 1.92 would involve days away, restriction, or transfer. Area does not enter that calculation. Headcount does, and part-time hours have to be converted to FTE first.
Why does a staff injury cost more in premium than in wages?
Because the wage cost is one case, and the premium cost is three years of modifier. A single days-away case at the median wage costs about $1,484 in wages and coverage. The same case sits in the loss run for three policy years, where 0.10 of modifier on a $60,000 manual premium is $6,000 a year.
Sources
- US Bureau of Labor Statistics, “Employer-Reported Workplace Injuries and Illnesses, 2023-2024”, released 22 January 2026 – bls.gov/news.release/osh.nr0.htm
- US Bureau of Labor Statistics, Table 1, incidence rates of nonfatal occupational injuries and illnesses by detailed industry, 2024 – bls.gov/web/osh/table-1-industry-rates-national.htm
- US Bureau of Labor Statistics, Survey of Occupational Injuries and Illnesses supplemental tables index – bls.gov/web/osh.supp.toc.htm
- O*NET OnLine, summary report for 39-3091.00 Amusement and Recreation Attendants, 2025 BLS wage data – onetonline.org/link/summary/39-3091.00
- Occupational Safety and Health Administration, recordkeeping requirements at 29 CFR Part 1904 – osha.gov/recordkeeping
- Occupational Safety and Health Administration, Appendix A to Subpart B of Part 1904, partially exempt industries – osha.gov/laws-regs/regulations/standardnumber/1904/1904SubpartBAppA
