Ladies and Gents, I've been having a few problems at work regarding the Var GRR as performed in Minitab v14.

The problem relates to the way in which the ANOVA table calculated. In Minitab the F statistic is calculated based on the variance of the repeat measurement, not as in the general heirarchical (nested) or crossed ANOVA.

Can anyone proved a process or statistical reason for this difference. I need an argument to convince a Chief Engineer and Statisical professional of the methods validity.

Mike Phillips

:frust:

Welcome to the Cove, Mike:bigwave:
I fear you have posted your inquiry at the beginning of the weekend doldrums here in the Cove. Many of our experts access from work and you may not get the answer you need until next week. Be patient!

Mike -

Can you maybe give an example?

Of course the F test is formed by the repeat measurement or "test to test" or "pure error". Why is your "Chief Engineer and Statisical professional" questioning this?

What do they think it should be?

Dave