Harvard - If I say you're out to lunch could I join you? I read the OP's comments and think it's pretty hard to have much confidence in any conclusion. IMO touching a probe to the ground gives a pretty iffy reading. Not useless, but could be pretty different than the reading from a ground bar a few feet long. If the service ground isn't very close to the measurement location, I could see the possibility of a few volts difference in the ground potential (no expert status claimed here). If there's a ground-neutral reversal somewhere in the residence, loads on that circuit would be pulling the safety ground voltage up. All I'd be willing to say is (1) Not feeling a tingle is a good sign but doesn't prove the chassis is correctly grounded, and (2) Not tripping the GFCI is a good sign but doesn't prove the chassis is correctly grounded. It's pretty easy to determine a bad ground by feeling a tingle, impossible to determine a good ground without some sort of test setup. I suspect most of us assume the ground is good unless there's some indication that it isn't. Using a 3-light tester is a good step but still isn't hard proof of a good ground. The only thing I'd consider hard proof would be to pull a few amps through the entire ground circuit with a test setup.
