The design for the comparative evaluation of the efficacy and toxicity of phenobarbital, phenytoin, primidone, and carbamazepine is outlined. A double-blind prospective study of a sufficient number of patients can determine the optimum drug to use initially for partial and generalized tonic-clonic seizures in adults. The rationale for methods defines the major parameters that should be addressed in order to determine optimum drug for longterm seizure therapy. Major problems in the function of such a project include aspects of sample size attainment, screening/recruitment, non-drug-related losses, and adjustments to the ongoing protocol. The design, with modifications, can be used to study other antiepileptic drugs and other types of seizures. © 1983 American Academy of Neurology.