A key skill for root cause analysis and verification action is to develop documented evidence that a change in design, materials or process steps has not had an impact on process or product output. For example, a process owner might want to compare the average strength of an adhesive joint before and after an unexpected incoming material viscosity change. The usual statistical tool for comparing averages is the 2-sample t-test, but the 2-sample t-test can only provide evidence of a difference between the means. In order to establish evidence of equivalence of the means, it’s critical to use the 2- sample equivalence test.
In this session, participants will learn the distinctions between the 2-sample t-test and the 2-sample equivalence test, learn how to plan and execute an equivalence study in Minitab ™ software and see real-world case studies where the 2-sample equivalence test was applied. We’ll cover recommendations for the typically most challenging step of determining, from subject-matter expertise, an appropriate range of equivalence. Case studies will include applications to design verification testing and process change assessment.
Karen Hulting Karen is a Distinguished Statistician at Medtronic. Over 25 years at Medtronic, she has applied a broad range of statistical methods in manufacturing, R&D and quality. Her current role is in Corporate Operational Excellence. Karen holds Masters and PhD degrees in Statistics from Iowa State University and an undergraduate degree in Mathematics from St. Olaf College. She is an adjunct instructor in the St. Cloud State University Medical Technology Quality graduate program in Plymouth.