The Precision Associated with the Sampling Frequency of Log-Normally Distributed Air Pollutant Measurements

The Statistical Form for the Ambient Participate Standard Annual Arithmetic Mean vs Annual Geometric Mean

“…A different program must be used for analyzing the TSP data and in computing the confidence intervals about the geometric mean of the data set. 18 " 20 In addition, the confidence intervals about the arithmetic mean are symmetrical, 21 but the confidence intervals about the geometric mean are asymmetric. 18 The author has seen several reports and papers where a geometric mean of TSP data and an arithmetic mean of SO2 data are computed and both are referred to in the text as annual means.…”

“…18 " 20 In addition, the confidence intervals about the arithmetic mean are symmetrical, 21 but the confidence intervals about the geometric mean are asymmetric. 18 The author has seen several reports and papers where a geometric mean of TSP data and an arithmetic mean of SO2 data are computed and both are referred to in the text as annual means. As an example of the confusion, Holland, et al, 2 cite the Winkelstein TSP data as "the two-year average isopleths of total suspended particulate" and later as "two-year total suspended particulate levels" but without mention that they are two-year geometric means.…”

Occupational Inhalation Health Risk Assessment of TCE Exposure in the Korean Manufacturing Industry

“…The industry classified the measured values using the Korean Standard Industrial Classification (KSIC 10th) [39]. Only industries with more than 30 measured values were chosen as research subjects to assume a log-normal distribution [40,41]. Excluding STEL measurements, this study used 6311 measurements in total.…”

Statistical Dependence of Hourly Carbon Monoxide Measurements

“…1 - 3 Several investigators have evaluated the precision of various sampling frequencies for replicating the monthly mean concentrations of total suspended particulates, using confidence intervals estimated from the variance of the data. 4 " 6 In these applications, statistical independence of the data has been implicitly assumed and the usual analyses will lead to biased and unrealistic estimates of the sample variance.…”