This paper shows that the single-cage model is not the right choice to represent squirrel-cage induction motors in dynamic studies. Therefore, the aggregate model of a group of motors must be based on their double-cage models. Four aggregation methods that use the double-cage model have been studied. Various tests have been performed on several groups with a different number of motors. The accuracy of the aggregation methods has been calculated at different points of the torque-speed and currentspeed curves. The highest error is always produced near the maximum torque point.Index Terms-Aggregate motor, double-cage model, single-cage model, squirrel-cage induction motor.