Two high school general mathematics classrooms were differentially taught a unit on percents, one with a cooperative and the other with an individualistic goal structure. A pre-post-test design with a three-way ANOVA analysis of treatment by time within subjects was used. Neither group was found to be significantly different from the other on the pre-test. Although both groups obtained significant gains on their post-test scores as contrasted with their pre-test scores, the cooperatively goal structured classroom demonstrated significantly higher achievement post-test scores than the individualistic group. The data strongly support theories concerning the effectiveness and motivation associated with inter-group competition of small cooperating groups. (Author)