Talk:ANOVA gauge R&R/Archives/2011

Design of R&R Study
I HAVE A QUESTION: CAN ANYONE TELL ME WHAT ARE GENERALLY ACCEPTED MINIMUM GUIDELINES FOR AN EFFECTIVE GR&R STUDY WITH RESPECT TO THE SAMPLING MATRIX? THIS WOULD BE IN A FACTORY PRODUCTION SETTING MEASURING MACHINED METAL PARTS WITH CALIPER, MICROMETER, ETC. MY EXPERIENCE HAS BEEN WITH 3 X 3 X 5 MATRICES (3 OPERATORS, 3 REPETITIONS OF MEASUREMENT PER OPERATOR ON 5 SAMPLES), AS WELL AS 3 X 3 X 10. WOULD 4 OPERATORS X 3 REPETITIONS X 3 SAMPLES BE STATISTICALLY INAPPROPRIATE?

THANK YOU,

--Jackdaiuto 20:12, 11 April 2007 (UTC)

I need to know what kind of study will help me to evaluate an attribute gage. I know I can't use gage R&R but I still need to know how reliable my gage is. Thanks for your help. L.R. 02:43, 20 July 2007 (UTC)

Gage R&R does not restrict number of operators. For variable GR&R, the minimum requiement is 2 operators, 10 parts & 2 repeats. You can use as many operators along with as many parts you think would be necessary to capture the total part variation. --136.1.1.101 23:36, 5 August 2007 (UTC)


 * see section added in article Mlonguin (talk) 17:59, 8 January 2008 (UTC)

Hey M, I've got references that say the minimum requirements need to be operators x trials x parts >= 60. Where do the 2, 10, 2 come from? Thanks! Homerspy (talk) 02:31, 23 August 2008 (UTC)

With only two operators, the Degrees of freedom (statistics) is only one: not enough to estimate the variation between operators. Three does not help much. If you really want to get an estimate of the variation component for operators, use MORE than ten. Rlsheehan (talk) 13:29, 23 August 2008 (UTC)

Hello I'm thai person I'm very confuse about calculation GRR my question if I have 10 sample I will use real value which I will already measure by vernier such as measure by size by vernier sample 1 (100.00) sample2 (100.50) sample3(100.30) ,.... sample 10(100.20) size tolurance (+,-0.5) if so finally I will calculate I should fill in what value in calculation It should be 100.00 mm.(becuase spec customer 100.00+,-0.5) or real value of each sample +,- 0.5 —Preceding unsigned comment added by 116.68.150.68 (talk) 10:01, 12 July 2009 (UTC)

Misconceptions
The misconception section is filled with unsupported opinions. Either this needs a major rewrite or it needs to be removed. Rlsheehan (talk) 01:17, 9 August 2010 (UTC)
 * Removed Rlsheehan (talk) 19:08, 16 August 2010 (UTC)