Re: 3 people 10 parts 12 positions on each part
That can be an issue, especially with smaller number of samples. Usually people don't have to memorize data points, because they are presented with a form and write in the data themselves - so they can see the other data. If someone else is writing down the data, it might be very hard to remember 12 points. But, since they are measuring 12 independent points on 10 independent parts, they are really doing 120 independent data points. It would have been better to pick fewer parts and do the 12 points (if the geometry is so compellingly different - and it may be if it is an awkward shape) and repeat the measurement on those parts. Blending in the gage variation with the part to part variation really is over-complicating a simple process. I don't see it as a "shortcut" of any kind, myself.
Bottom line - an ndc of 1 (or 1.3 in miner's data) shows no matter how fine of a comb you use, this gaging isn't working.
Miner is correct. if you don't have each operator repeat each measurement (randomly without knowing what value they got the first time they measured that part and location) we cannot determine measurement error.
Bottom line - an ndc of 1 (or 1.3 in miner's data) shows no matter how fine of a comb you use, this gaging isn't working.