You are here:
ENVIRONMENTAL TECHNOLOGY PROTOCOL VERIFICATION REPORT, GENERAL VENTILATION FILTERS (WITH APPENDICES)
Citation:
Elion, J. M., J. Hanley, D. S. Ensor, AND D. L. Franke. ENVIRONMENTAL TECHNOLOGY PROTOCOL VERIFICATION REPORT, GENERAL VENTILATION FILTERS (WITH APPENDICES). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-00/111.
Description:
The results of the electrostatically charged filters tested by three labs were analyzed following the method described in "Standard Practice for Conducting an Interlaboratory Study to Determine the Precision of a Test Method" (ASTM E691). The statistical analysis examines the consistency of the test results from laboratory to laboratory and the consistency of results within a laboratory from one laboratory to another.
Such a statistical analysis is necessary because not every factor that may influence the outcome of a test can be completely controlled. Some differences may result from unavoidable random errors while others may result from the inherent variability of the test procedure itself.
Different factors that may contribute to variability in the application of a test method include the operator, the equipment used, the calibration of the equipment, and the environment (temperature, humidity, etc.). The reproducibility measure "h" considers these four factors as sources of variability and reflects what precision might be expected when randomly selected samples are sent to random "in-control" laboratories. The repeatability measure "k" reflects what precision might be expected when these four factors are kept or remain nearly constant.
The "h" and "k" values for the three labs were determined for all twelve particle sizing channels. The values for the initial efficiency test results are shown in Tables 3 and 4. The critical values for "h" and "k" were taken from Table 12 of ASTM E691, where "p" is the number of laboratories and "n" is the number of samples. For P=3 and n=3, h(critical)=1.15 and k(critical)=1.67. Values that exceed the critical values are in bold typeface. Values that approach the critical values are underlined. It should be noted that the relative inexperience with the test method is likely to skew the results towards higher values.