Studies examining students often marginalized in the sciences are adopting sociocultural theories to explain learning as a social process. Some sociocultural theories contend that learning occurs as individuals participate in communities of practice (COPs) (Wenger, 1998). Instead of looking at cognitive or conceptual changes, these theories examine learning as a process of identity formation and transformation through participation and membership in COPs. Wenger (1998) defines three dimensions (joint enterprise, mutual engagement, shared repertoire) to help describe and identify COPs, as well as to illustrate what a sociocultural theory might look like in practice. Few empirical studies, however, analyze the appropriateness of this theoretical framework. Rather, researchers generally assume the framework is appropriate and adopt it to explain particular aspects of learning. Additionally, environmental education research rarely employs existing learning theories to explain learning. Thus, this study critically examines the applicability and usefulness of the COP framework in explaining learning of students in science classrooms and in an after-school environmental education program, the Environmental Club (EC), by asking the following questions: 1) How do COP dimensions (joint enterprise, mutual engagement, shared repertoire) manifest themselves in science classrooms and ECs? 2) How does learning as participation, membership and identity formation manifest itself in science classrooms and ECs? I examined two science classrooms and three ECs in south Texas, using a case study methodology. The data collected included classroom and EC observations; student focus group, individual student, teacher, and principal interviews; student journal entries; student drawings; and secondary data on schools. I manually coded interview transcripts and other data using the COP dimensions and student perspectives on community. I then determined where findings from multiple methods converged in each case and employed analytic induction to compare findings across cases. Findings illustrated Wenger's dimensions manifested themselves in four of the five contexts. However, differences in contexts contributed to the way in which dimensions evolved and developed and consequently to how learning was manifested as a social process. The framework was applicable and useful for understanding learning as a social process when Wenger's dimensions evolved within the context, were agreed upon by participants and the researcher, and when the majority of participants felt they belonged to a COP. Findings imply that the applicability of the framework should not be assumed for all learning contexts but that it is useful for providing a sociocultural view of learning in environmental programs.