State of Noise at Work Index · v0.1 - What 8 real noise assessments actually found
A first look at what noise at work actually looks like on the ground, built entirely from 8 real, anonymised assessment reports already public on our Resources page — not a national survey, and not dressed up as one.
- sites measured at or above the upper action value
- 8 of 8
- sites exceeded the exposure limit value itself
- 7 of 8
- the single worst reading recorded (aluminium machining)
- 98.8 dB(A)
Site by site
Sourced from claude/Case-Studies-and-Report-Access.md — the same 8 case studies already public on the Resources page, each from a real, anonymised Impact Acoustics assessment report.
| Sector (anonymised) | Region | Worst documented reading | vs. 85 dB(A) UAV | vs. 87 dB(A) ELV |
|---|---|---|---|---|
| Production line (robotic welding) | South West | 94.0 dB(A) | Exceeds | Exceeds |
| Aluminium machining | Gloucestershire | 98.8 dB(A) (5 machines ≥85) | Exceeds | Exceeds |
| Chilled-food factory | Cornwall | 93.9 dB(A) (plant room) | Exceeds | Exceeds |
| Joinery workshop | West Midlands | ≥85 dB(A) on 10 of 13 tools No single worst-case figure published for this site — carried through as-is rather than inventing one. | Exceeds | Not stated |
| Marine engineering | South Coast | 97.1 dB(A) (hammering) | Exceeds | Exceeds |
| Brewery (keg washing) | West Midlands | ~96 dB(A) group exposure; 134.3 dB(C) peak The ~96 dB(A) figure rests on a deliberately conservative task-time basis pending confirmation. The 134.3 dB(C) peak is 0.7 dB below the 135 dB(C) peak action value, so it is not itself shown as a peak-metric breach — only the separate exposure metric breaches here. | Exceeds | Exceeds |
| Metal fabrication | Dorset | 90.0 dB(A) (shaking-out) | Exceeds | Exceeds |
| Wood processing | West Midlands | 95.4 dB(A) (hammer mill) | Exceeds | Exceeds |
UAV = upper exposure action value, ELV = exposure limit value, both under Regulation 4 of the Control of Noise at Work Regulations 2005 (SI 2005/1643): 80 dB(A)/135 dB(C) lower action value, 85 dB(A)/137 dB(C) upper action value, 87 dB(A)/140 dB(C) exposure limit value.
What this is, and what it deliberately isn’t
Eight assessments across eight sectors is a real but small, non-random sample of our own client base. It says something true and citable about those eight sites — it can’t support a claim like “UK manufacturing averages X dB(A).” A genuine nationally-representative index would need a much larger, defined sample and a published methodology; this is a first, honest step toward that, not a finished one.
“In a review of 8 real UK workplace noise assessments across 8 different sectors, every single one had at least one point exceeding the legal action value that triggers mandatory hearing protection — and 7 of 8 breached the exposure limit value itself.”
A defensible headline for external use — true, sourced, and striking without claiming national representativeness.
What it would take to make this a real, citable annual index
- A signed-off anonymisation standard. Aggregating the full assessment tracker needs the same sector + region + dB-banding approach already used for these 8 published case studies, applied consistently and signed off before any wider figures go out.
- A real query across the full client base. Not just these 8 — the complete assessment history, properly anonymised, to move from a proof-of-concept to genuine statistical weight.
- A defined annual cut-off and sample-size floor. So the index reads as a genuine year-over-year series, not a one-off snapshot dressed up as a trend.
- A published methodology note. Sample size, what “assessment” means, and that this is one assessor’s own client base — the same honesty this first draft is built on, just formalised alongside the numbers.