Laserfiche WebLink
Reproducibility data are presented in Table 8. Of the 30 ambient air <br />duplicate pairs, the reproducibility differed by more than 20 percent <br />in only 2 duplicate pairs, or 7 percent. Both the large differences <br />resulted because the compounds were detected at very low concentra- <br />tions. At or near the detection limit, very small fluctuations in <br />quantity result in large percentage differences. Concentrations near <br />the detection limit are also more vulnerable to fluctuations because <br />the limits of linearity are approached, and instrument noise and <br />background interference become quality factors. Concentrations near <br />the detection limit would thus be expected to be outside the 0 to <br />20 percent acceptable range of reproducibility. Of the 10 well <br />duplicate pairs, the reproducibility ranged from 0 to 79 percent. <br />Mean reproducibility varied from 0 to 35 percent, with standard <br />deviation from 0 to 40 percent. <br />EFITWE�t <br />For this investigation, data precision is a measure of the percent <br />difference between the ambient air concentrations of a specific <br />constituent from a primary sampler and the concentration from a co - <br />located, duplicate sampler. Since the approved ASWAT proposal did not <br />include co -located samplers, no precision data were determined. <br />I. iTU -�- <br />According to the criteria specified in Guidelines, the data set was <br />100 percent complete. <br />PJ9 9390405.00D 5-7 Rev. 0 09/07/88 <br />EMCon Associates <br />