AIAG's MSA Reference Manual, page 95, "Gage R&R Confidence Intervals"

[Ben]

Reference AIAG's MSA Reference Manual, page 95, "Confidence Intervals"

Does anyone know of a supplemental text I can go to in order to figure out how to determine confidence intervals for a gage R&R study?

I apologize, but I cannot interpret the phrase "where MS denotes percentage point for a mean square distribution (i.e., a Chi-square deviate with the same degree of freedom divided by degrees of freedom, v)"

Thank you.

 

[Marc]

Any statistical gurus out there with a comment or 2 on this?

 

[max]

The formulas might be in some SPC texts. Here is a journal reference:

"Gauge Capability Analysis and Designed Experiments. Part II: Experimental Design Models and Variance Component Estimation" by Montgomery and Runger, Quality Engineering, Vol. 6, No. 2, OCTOBER 1993, pp. 289-305.

(part I: Quality Engineering, Vol. 6, No. 1, JULY 1993, pp. 115-135)

As for "where MS denotes percentage point for a mean square distribution (i.e., a Chi-square deviate with the same degree of freedom divided by degrees of freedom, v)": if you use ANOVA to get your results from the study, there will be mean squares for part, operator and so on. Each mean square has a specific degree of freedom associated with it. These numbers (the df) are needed to compute the confidence interval.

For example, we performed an R&R study with 16 parts. There are 16-1=15 degrees of freedom for the part source of variation in the study. If I wanted to get the 5th percentage point for a Chi-square distribution with 15 df, Excel would give it to me with CHIINV(0.05,15), which is 24.99.

Consult the references above. They are written for non-statisticians.