The present study conducts an experimental investigation to compare the efficiency of the cooperative learning method with that of the traditional learning method. A total of 42 engineering students are randomly assigned to the two learning conditions and are formed into mixed-ability groups comprising three team members. In addition to the regular daytime classes, homework sessions are arranged such that the out-of-hours learning method and learning time can be effectively controlled. The students' academic achievement is evaluated by means of unit tests in the daytime classes and homework tests in the out-of-hours sessions. As an alternative method for resolving the multiple grouping criteria problem, the analysis of covariance method is used to compensate for the initial difference in the prior knowledge of the students in the two learning conditions regarding the course contents. The results show that given an equivalent learning time, the students in the cooperative learning condition outperform those who study alone in both the unit tests and the homework tests. Therefore, it is concluded that cooperative learning has a higher efficiency than the individualistic learning method. (Contains 5 tables.)