NIYA, it would help if you better described how you modeled this. As johnbridge asks, did you use springs based on subgrade modulus of soil? What supports/restraints did you use "without soil"? Obviously soil has some flexibility which can change load distribution and displacements.
In addition, if you're analyzing a raft foundation, it may be necessary to model soil springs as nonlinear compression-only in order to properly account for load distribution.