This paper deals with the application of optimal design theory in a ty
pical experiment in economics. It describes general principles for con
ducting efficient experiments for model discrimination and illustrates
the benefits of the techniques by an empirical comparison.