DISCLAIMER: I am NOT an Industrial Hygienist, so seek help elsewhere if you find this post intriguing.
How many of you use the 95th Confidence Interval in your “Data Sampling”?
Whether it’s personal sampling, area sampling, or any sampling looking for safety data, do you attempt to achieve the 95th confidence interval?
A 95% confidence interval is a range of values, calculated from sample data, that is likely to contain an unknown population parameter. It means if you repeat an experiment many times, 95% of the calculated intervals will hold the true value.
For example, there is a department within the business that you suspect of being a High-Noise area. The safety pro obtains a calibrated sound level meter and does a baseline noise survey of the area. It is established that some locations exceed the 85-decibel “action level”. This is NOT a TWA, but a point-in-time measurement, indicating that it may be possible for a worker to exceed the 85-decibel TWA. So we set up some personal monitoring and have a single worker wear a Noise Dosimeter for a single shift. The data shows that the TWA is below the Action Level.
Are we satisfied that there are no harmful noise exposures in that area with the Noise Survey and the single noise dosimetry data? How “confident” are we that these two (2) pieces of data are telling us the entire noise hazard story of that area?
I see this quite often in our audits: take a single sample and claim no hazard is present. We typically find the sample was even taken on a “light production” shift, swaying the data even worse.
How many samples across all shifts do you feel are enough, such that you are confident of what the data is telling you? How formal are you in documenting your data so when that STS Hearing Loss Case pops up in your annual audiograms, you can confidently challenge it?
Lastly, if your business is working 12-hr shifts, EVERY SINGLE PEL(noise, chemical, etc.) has to be recalculated from the PELs that are based on an 8-hr TWA. We see this fundamental error in a lot of audits, and once we recalculate the data, the facility finds it has some exposure issues.