The dose-response models for full agonists and for a particular type of partial agonist can be described by sigmoidal curves and bell-shaped curves, respectively. The methods currently used to evaluate the interaction of a full agonist and a partial agonist require a large number of experimental units and base their analysis on nonlinear regression analysis, which may not be statistically appropriate. We propose an appropriate design and a global nonlinear model to evaluate such interactions. The new model allows us to estimate the interaction parameters and the parameters that characterize the individual partial agonist curve and the full agonist curve.