Variance Decomposition Methods (1-day)

Instructor

Dr. Wayne A. Taylor

Course Description

Provides a basic introduction to the variance decomposition approach to identifying the root cause of variation.  This approach splits the variation up into components and finds the relative effect of each component on the total variance.  For example, 80% of the variation is from lane differences, 5% between days and 15% within days and lanes.  This allows an investigation to focus on the leading source of variation.  This approach is described in the book Optimization & Variation Reduction in Quality in Part III: Searching for Differences.   The course covers a variety of methods including various charts, variance components analysis and component swapping studies, along with where they fit into the Six Sigma Design and Six Sigma Improvement processes.  An application of a variance components analysis is given in Appendix C of STAT-03, Statistical Techniques for Process Validation, of the book Statistical Procedures for the Medical Device Industry.

Primary Audience

This course is intended for scientist, engineers and other technical staff.

Course Outline

A.M.

  • Searching For Differences
  • Multi-Vari and Control Charts
  • Analysis of Means (ANOM)
  • Analysis of Variance (ANOVA)

P.M.

  • Variance Components Analysis
  • Concentration Diagrams
  • Component Swapping

Prerequisites

Course Materials

Participants receive a comprehensive course manual.

Software

Minitab