A retrospective analysis of the effects of electroconvulsive therapy (ECT) was performed for two groups of 11 patients matched according to age (mean age, 52 years), sex, and diagnosis. Group 1 received ECT according to the age–dose protocol; group 2 was treated according to the titration method. A higher dose relative to the seizure threshold appeared to shorten the seizure duration. At the first treatment, the correlation between stimulus intensity and seizure duration was negative. In the titration group, the initial mean charge of 91 mC resulted in a seizure duration of 51 s, whereas in the age–dose group the seizure duration of 31 s was significantly shorter despite a higher mean charge of 312 mC. Seizure duration decreased during the ECT course in the group treated first at low dose (titrated) and then at 2.5 times the initial threshold. High stimulus intensity represented adequate treatment, although it produced short seizures. Thus, seizure duration proved to be an unreliable guideline for effective treatment. Furthermore, focus on seizure duration led to frequent high-dose restimulation in the elderly. The titration method obviates inadequate or excessive charges because the seizure threshold must first be determined.