Interpretation of agonist affinity estimations: the question of distributed receptor states