The purpose of this study was to determine if group cohesiveness has a positive effect on learning. The specific hypothesis for experimentation was that the performance of elementary school students in mathematics would be higher for those in cohesive than in noncohesive groups. This hypothesis was based on the assumption that the students were task-oriented and motivated to obtain the group goals. It was intended that the research findings be a contribution to the fields of education and sociology. The subjects for this study were 156 second, third, fourth, fifth and sixth grade pupils from an elementary school within the East County of San Diego, California. At the beginning of the six-week period of experimentation, thirty-two groups of four or five members were formed on the basis of the sociometric data obtained from the children. Fifteen groups were defined as cohesive and seventeen as noncohesive. In addition, every effort was made to match the two types of groups according to I.Q., mathematical aptitude and present knowledge of the material which was to be covered during the experimental period. During the time of investigation the teachers adhered to the following schedule. First, the teacher presented the material for that day's lesson. Following the instruction, the students worked in their groups on assignments related to the teacher's presentation. Weekly mathematical tests, based on the material which had been covered, were administered to the individual students. The scores from these tests served as the measurement of the dependent variable of learning. Since cohesiveness constituted the main independent variable, it was important to establish its existence throughout the experiment. Therefore, five sociometric tests were administered during the investigation to measure the degree of group cohesiveness from week to week. At the conclusion of the experiment, a significant difference was found between the test scores of students in the cohesive and noncohesive groups. Thus the research hypothesis was supported that the performance of elementary school students in mathematics is higher for those in cohesive than noncohesive groups. From the research findings, it appears that to form groups on the basis of student selection is an effective way of grouping children and an aid to student achievement. One can also conclude from this study that there is a relationship between the need-satisfaction a child experiences in a group and his subsequent learning. The results of this investigation suggest that teachers should consider utilizing the property of cohesiveness when deciding upon a grouping system for their classroom, because the findings provide evidence that group cohesiveness does have a positive effect on learning. Within the field of sociology, the results of this research support findings in previous small group experimentation. Within an educational environment, it has been demonstrated that cohesiveness has a positive effect on productivity, which has been defined in this thesis in terms of learning.