That wouldn't test for whether we evolved for it. We could, at best, simulate early human societies with various levels of violence and see if one led to large jaws. (Though currently we have insufficient information to model such societies.)

Most people grossly underestimate the forces of genetic drift and personal preference on phenotypical evolution. The null hypothesis for any gender-linked trait having evolved should probably be "it made them more attractive to a fertile population of the opposite sex."

Even if you can show experimentally that lots of face punching will lead to stronger jaws, that doesn't prove that it's what happened historically. We could have evolved the same strong jaw for a completely unrelated reason.