Page 212 - Six Sigma Demystified

192 Six SigMa DemystifieD

Interpretation

The end result of the affinity process is a collection of problems grouped by
main ideas. These main ideas, represented by the header cards, may provide key
drivers that need to be addressed for achievement of the goal.
It’s important to remember that the affinity diagram uses only subjective
opinions of issues. As with any of these types of tools, we can use the tool to
gain focus, but we then must substantiate these ideas with objective evidence
using properly analyzed data.
These issues may be further developed in other tools, such as a prioritization
matrix.

ANOVA

ANOVA, the acronym for the analysis of variance, is a tabular presentation of
the sum-of-squares (SS) variance attributed to a source, the sum of squares at-
tributed to error, and the total sum of squares from the data. F statistics on the
significance of the source relative to the error are included.

When to Use

Measure Stage
• To isolate sources of measurement error, particularly when Repeatability
and Reproducibility (R&R) studies cannot be done (such as in destructive
testing).

Analyze Stage
• To look for differences between subsets of data as a source of variation in
a process
• To investigate statistical significance of a regression model to uncover
potential process drivers

Methodology

ANOVA provides a means of comparing the variation within each subset of
data to the variation between the different subsets of data. The between-subset
variation is a reflection of the possible differences between the subset averages.
The within-subset variation, for each subset, is a reflection of the inherent
variation observed when sampling from the subset repeatedly.