Bias Equivalency in the Laboratory
In many testing laboratories, it can be necessary to qualify a new test method, instrument or lab itself with respect to one or more performance characteristics. One such key characteristic is bias, or mutual agreement of average test results run on the same material. A proposed ASTM standard will aid in the determination of bias equivalency.
WK37659, Guide for Testing for Bias Equivalency in the Laboratory, is being developed by Subcommittee E11.20 on Test Method Evaluation and Quality Control, part of ASTM International Committee E11 on Quality and Statistics.
According to Thomas Murphy, independent statistical consultant, T.D. Murphy Statistical Consulting LLC, and E11.20 chairman, anyone concerned with the development or application of a test method would be a potential user of WK37659.
"This proposed standard would apply to most, if not all ASTM committees having test methods," Murphy says. "Guidance will be given on the proper design of equivalence studies, including proper numbers of tests required, assistance in deciding what constitutes sufficient agreement, and selection of the proper statistical test for evaluating the data. Methodologies will be given for determining equivalence at a single level of a property or over a range of property levels."
Murphy says that the task group developing WK37659 is seeking participants who have knowledge of and experience with bias equivalency and who are interested in taking an active role in the writing and final editing of the proposed standard. In addition, the task group is interested in obtaining data sets from previously conducted equivalence testing since actual examples will be included to illustrate the methods covered in WK37659.
CONTACT Technical Information: Thomas D. Murphy, T.D. Murphy Statistical Consulting LLC • Morristown, N.J. • Phone: 973-292-0637 | ASTM Staff: Katerina Koperna • Phone: 610-832-9728 | Upcoming Meeting: Oct. 17-19, ASTM International Headquarters, West Conshohocken, Pa.