|
Activity Number:
|
98
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 4, 2003 : 9:00 AM to 10:50 AM
|
|
Sponsor:
|
Section on Statistics & the Environment
|
| Abstract - #301877 |
|
Title:
|
Comparing Lognormal Distributions Resulting from Two Respirable Dust Air Sampling Methods
|
|
Author(s):
|
William K. Sieber*+ and Alan Echt
|
|
Companies:
|
National Institute for Occupational Safety and Health and NIOSH
|
|
Address:
|
7524 Kingstonview, Cincinnati, OH, 45255-2495,
|
|
Keywords:
|
lognormal distribution ; direct reading instrument ; estimation ; MVUE ; percentile ; tolerance limit
|
|
Abstract:
|
The availability of instruments providing real-time measurement of exposure levels provides data from which the distribution of employee exposures may be ascertained. Other methods to determine exposure levels involve estimation from a number of samples. To compare results from such procedures with real-time data, respirable dust exposures at a construction site were measured in a series of consecutive runs during which exposures were measured simultaneously both in real time and by collecting personal breathing zone air samples. Parameters estimated for a lognormal distribution were compared using data collected by each method. Of particular interest were estimates of the mean, percentile distributions, and tolerance limits. The Minimum Variance Unbiased Estimator of the arithmetic mean of the lognormal distribution from the sample data was not significantly different from the arithmetic mean of the real-time data. Estimates of the 95th percentile values of the sample data were significantly lower compared to real-time measurements (p< .001). This suggests the estimated distribution may minimize observed variation in exposure levels.
|
