Run a test on natural ground considered to be OK for a reasonable bearing value. My experience shows it around 85% of standard Proctor generally.
Hey, it isn't a simple test procedure with "no sweat" problems. Take an extreme case at the St. Lawrence Seaway job some 70 years ago. The earth there was a very dense glacial til. Perhaps there was a break-down of coarser particles in the lab Proctor tests, maybe a bad stone correction number. I can't recall all the details, but my boss (a former Corps of Enineer)then was hired by the earthwork contractor to help him with meeting the Corps of Engineer's spec. Apparently it was not possible to meet the spec. with all sorts of effort. It got so heated a problem for the contractor that he committed suicide, even with the best specialist soil engineer helping him. Now that fussyness is really not needed on any job.