Research on group creativity has focused on the ability of group members to improve their divergent thinking by interacting with others who have different perspectives. I argue that this research has ignored the issue of how groups converge around highly creative ideas. I suggest that the processes that facilitate divergent thinking can also reduce the ability of the group to build on and integrate information in creative ways. Moreover, differing perspectives disrupt the cognitive, affective, and social processes necessary for the group to select creative ideas. Thus, in organizational groups charged with developing creative outputs, diversity of perspective can actually reduce the group's overall level of creativity. In contrast, when group members share a common framework for thinking about the task (i.e., have a shared representation), they can generate creative ideas by building on and integrating the ideas of multiple members and recognize and select the most creative of these ideas. I develop a theory of group creativity based on the challenges for groups in making intuitive, creative judgments. I argue that shared representations of the group task - rather than diversity of perspectives - are the basis for group creativity, because they help groups to converge around highly creative ideas. I test these ideas with a multi-method approach. The first exploratory study in the asset management industry reveals the importance of shared representations to group creativity. The second field study tests the role of shared representations in individual cognitive, affective, and social processes. The third study is an experiment that demonstrates these affects and their impact on creativity at the group level. The final study presents quantitative evidence of these effects in the asset management industry