Hey fellows! Is there any standard or protocol in gathering the measurements to be used in GR&R Study regardless of the number of appraisers, parts, etc.? I mean, what are the required steps or behavior of the person responsible for the analysis? Is it really ideal that the appraisers should know nothing or be not aware that there is actually "a study" going on?

Well Henry, I'm not an expert on this subject but I know certain things are needed.

Repeatability studies are series of measurements with as little variation as possible in the course of taking measurements. That generally means same person taking the readings, same setup, same time frame per data set; the person needs to be suitably skilled so that variation would be due to the instrument and not the handling or measuring technique. All of this taking place usually means the person knows an experiment is being done.

Reproducibility studies are designed to introduce enough controlled variability to show the instrument's ability to give the same results if the readings are done at different times of day, by different people, tool warmed up versus not, etc. In this case a given data set (10 or more readings as a rule) is taken by the same person and in the same time frame. Because of this it also seems likely that the person would be aware that an experiment is taking place.

If your question is whether or not measurements can be transferred from records taken during actual inspections (based on convenience), I would say no because the experiment is supposed to be controlled.

Any other takers on this? Surely someone can do a better job of explaining this than I can. :o

Thanks Miss Jennifer. I appreciate your response.

Jennifer is correct. It will not hurt for the operator to know that they are part of a study. The most important thing is to randomly present the parts to them for measure in such a manner that they do not know or remember the original measurements obtained. Therefore, there should be a person coordinating the study that records the data for the operator, so they do not observe their (or others) prior results. To obtain an accurate result, the operators should be instructed to perform the measurements as they would normally. If they take extraordinary care during the study and revert to normal once it is complete, you will end up fooling yourself about the gage.

The next most important thing is to properly select the parts themselves. If the study's intent is to determine suitability for process control, the variation of the samples MUST reflect the actual process variation.