From clay to sand, I have yet to meet terrestrial soil that did not respond to a dedicated ground stake and several gallons of water. Testing consisted of 3 repeat direct shorts line to earth through a 30 amp breaker.
I have encountered far too many marginal receptacle "grounds" that "tested" good then went open circuit with more than say a quarter-ampere applied short. I know of no effective way to thwart this - and it is definitely dangerous. I had a line to earth test circuit within Quicksilver. With a complete circuit it held the NC relay open. Loss of ground triggered a Sonalert. Alarms became such a pain down here I switched to my own ground rod.