The purpose of this study was to examine the effects of four teaching methods on several measures of motor skill, self-concept, and social attitudes of 130 fifth-grade children (67 boys, 63 girls), who were randomly chosen from five elementary schools in one area. Teaching methods were systematically applied for 60 teaching days (20 weeks). Measurements were assessed on three occasions (pre-, mid-, and post-experiment). A 4 x 2 x 3 multivariate analysis of covariance showed the 'combined' method most effective in improving the students' motor ability, while 'indirect' and 'game-oriented' methods contributed to the improvement of self-concept and social attitudes, respectively. Since each of the four methods had different effects on various measures made to evaluate whether aims of physical education had been met, the teacher should be aware of many methods and able to implement them in various combinations, depending upon the special features and purposes of a lesson.