I would assume for our interests in a K, we would want the test for twist to be 90 Deg from what your rendering shows? And we don't want the arm to add its stiffness to the K?
Correct.
As with my experience in applying loads, the part should be tested as installed. In my opinion. As the frame would tie the two mounts together while installed in the vehicle, it should be tested as it was intended to be mounted during use. As long as the test was consistent between the mounts, the % difference would still be accounted for.
As an example, a common motorcycle chassis uses the engine as a structural member. Testing the stiffness of the chassis without the engine would not be representative to it's "as-intended" condition. For this instance, the frame members are going to be there, no matter what K you use.
