That is a matter that seems to hardly ever be brought up. Yet it would seem clearly important.
I actually do not know much about it. You are right that two different ligands may be bound and that this can result in a different effect than the same ligand being bound, but precisely how this relates to the actions of the selective estrogen receptor modulators, or the actions of say estriol, I don't know. I expect it it is a relevant additional mechanism to their partial estrogen antagonism. Namely, by changing the effects of estradiol binding to the other part, relative to what would have happened had the dimer been homogenous for estradiol.
The only paper I know on it is http://jme.endocrinology-journals.org/cgi/content/full/38/5/569
Hopefully it will be helpful.