Abstract
In this paper we consider two methods for comparing variability of two measurement processes. The comparisons are based on statistics commonly computed in an R&R study. Computer simulation is used to compare methods and to support our recommendations. The basic conclusions are that: i) a simple method proposed by Cochran (1951) works well across a wide variety of conditions typically encountered in R&R studies, and is recommended over a modified large sample method; and ii) in studies where operators are randomly selected, samples of more than three operators are needed to provide useful results.
Additional information
Notes on contributors
Richard K. Burdick
Dr. Burdick is Professor of Statistics in the College of Business. He is a Member of ASQ. His email address is [email protected].
A. Elizabeth Allen
Ms. Allen is a Statistician at IMS Health.
Greg A. Larsen
Mr. Larsen is the site Statistician and a Senior Member of ASQ.