The gold standard for cohesive soils is very high quality undisturbed samples with CU, UU, CD triaxial and odeometer tests. Often we don't get high quality undisturbed samples, and budgets may not allow for triaxial tests. So you are stuck using a few UUs, maybe UCS and correlations between SPT, Atterberg limits.

For granular material, SPTs, PSD, and shearbox tests are used, you wont get any advanced triaxial tests as its nearly impossible to get economical undisturbed samples of SAND.

For the most part this level of assessment is conservative enough to not be too cost prohibitive.

Re disturbance, I think it should generally leave you with more conservative parameters, lower shear strength, higher compressibility

Thanks EireChch

So for cohesive soil CLAYs, SILTs, lab tests are more accurate and where I practice, lab tests seem to be quite cheap.

for granular soil SAND, In-situ tests or shearbox are suitable as it is hard to get a good undisturbed sample.

More or less, but I would still be cross checking lab results with correlations between other tests such as SPT, CPT, atterbergs etc. There shouldn't be a big big difference between correlations and lab results.

The CPT is really your best tool for trying to compare insitu test to lab results. There are many webinars from Robertson and Gregg drilling which you can check out.

Quote (EireChch)

nearly impossible to get economical undisturbed samples of SAND

This is true for most situations but can be achieved for a manageable cost in fine grained sands to silt material, common in tailings using a specialized piston sampler to get undisturbed tubes, the name of which escapes me currently.

GEG - I have heard that before, and my company even use similar sampling methods, however I think they are just less disturbed than other sampling methods. Not suitable for doing a triaxial test on, IMO.

The name is a GUS sampler.