I agree, that works good against steel, but if you use them against anything with "give" you may need to revert to feel.
The one thing people often mess up - especially in the shop if they are in a hurry - is impact. The part will measure much smaller if you run the gage down fast against the part. The smaller ratchet know can really get it going fast, too. Usually do not pick that up in GR&R studies, because they are being watched, they are relaxed and slower than on the line.
I saw a TV movie a long time ago where they were training a dog to be a bird dog. To train the dog not to clamp down on the bird, they would play catch with a ball made of tape and nails. The dog learned to bring it back with a more gingerly grip in the jaws. Training people to understand the issue of feel, I use rubber and plastic blocks. It gives them an idea just how much the torque of a screw can affect their measurement.
To tell you the truth, I have never done this experiment, but I would love to see it done: take 5 plastic cubes. perform a good, blind GR&R on those parts. Then, write on each one a "supposed" correct number and do another GR&R. Do that with calipers and micrometers. I bet the results would be telling.