Exploring Measurement System Study Sample Size and the Power to Detect Production Process Shifts

• Activity Number: 600
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 AM
• Sponsor: Quality and Productivity Section
• Author(s): Laura Lancaster*+ and Christopher Gotwalt
• Companies: SAS Institute Inc. and SAS Institute
• Keywords: Measurement Systems Analysis ; Process Control ; Sample Size ; Power ; Simulation

Abstract

It is important to be able to detect shifts in a production process. The ability to detect these shifts is affected by the ability to measure product. A good measurement system analysis informs the user how much the measurement system will affect their ability to detect shifts, but how much can we trust measurement systems analysis results for a very small measurement system study? We have developed a way to compute prior confidence intervals on the probability of detecting shifts in the production process given guesses for the measurement system variance components. We will explain how we compute these prior confidence intervals and show results of a simulation experiment that explores how the sample size of the measurement system study affects the power of being able to detect shifts of various sizes.