Re: Should we check the normality of measurements when the calculation of Gage RR is
Star18,
Sorry that this slipped by for this long without an answer.
Due to the small sample sizes and the calculations involved, normality of the measurements is not a concern. The one area where you want to be cautious of normality is in the selection of the parts used in the study.
The parts selected should represent the same variation as the parent distribution. If that distribution is not normal (it does NOT have to be Normal), it may make it more difficult to select samples that represent that distribution.
If the samples reflect the non normal distribution, you are good to go.