High-Efficiency and High-Breakdown Methods for Interlaboratory Comparisons

• Activity Number: 694
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Hung-Kung Liu*
• Companies: NIST
• Keywords: Interlaboratory studies ; robust methods ; efficiency ; breakdown point

Abstract

Interlaboratory comparisons (ILC) determine and monitor the performance of participants in specific tests or measurements. In analyzing ILCs, it is desirable to estimate the center of the distribution of results from competent participants without identifying and excluding outliers. However, the performance of classical robust estimators used to achieve this purpose is not well characterized. In general, the performance of any estimator can be judged by its efficiency and breakdown point with a trade off between the two. In this talk, we compare the performance of computationally intensive methods with classical robust methods for ILC analysis in terms of efficiency and breakdown point.