The aim of this study is to find out which one of the following two different techniques is more effective to produce the desirable structural change of the small group of children in which an egocentric and autocratic leader exercises influnece. Procedure : (1) The subjects of this experiment were 36 school children of the fourth-year-class who were divided into 9 groups of 4 each. One of the 4 members of each group was the above-mentioned leader. Of the 9 groups, 4 were the 1st experimental groups, one was the control group and the other 4 were the 2nd experimental groups. (2) Members of each group were asked to rank 5 socially good conducts in their own individual orders of preference. Group rankings were then formulated. A final ranking was then secured from each individual. At the end of the 1st day experiment the experimenter explained to the leaders of the 1st experimental groups the ideal way of the democratic management of a group (the 1st technique). Whereas each leader of the 2nd experimental groups was made to observe members of a democratic group discussing with one another (the 2nd technique). This democratic group was composed of 4 children of the fourth-year-class. (3) Twenty-four hours later members of each group were asked to rank 5 socially wrong conducts in their own individual orders of preference. Group rankings were then formulated. A final ranking was then secured from each individual. (4) A couple of weeks later the same procedure as (3) was repeated. Result : (1) On the 1st day followers of both experimental groups were likely to withstand the impact of group opinion under autocratic leadership, which probably indicates that no strong group opinion was formed under this condition. (2) On the 2nd day the final rankings of the followers were found to agree with group rankings on the whole. The difference between the final rankings of the followers and final rankings of their own on the 2nd day was found to be greater than that of the 1st day, which probably was due to the fact that they had been much more content with and influenced by the group rankings. The important fact is that the degree of the above-stated agreement and the difference was higher in the 1st experimental groups than in the 2nd experimental groups.