The presentation has three parts. First, I describe different forms of cooperation in science. Second, I discuss Thagard's (1997) and Wray's (2002) epistemic standards for assessing scientific practices. Finally, I examine students' and young scientists' epistemic position in the research groups from an educational point of view.Science is a product of individual minds and of complex social organizations (Thagard 1994). At present collaboration plays a significant role in natural and social science. For example, Hardwig (1991) claims that teamwork is an inescapable feature of modern science. Although cooperation has become prevalent in modern science, there has been little discussion on the nature and the significance of collaboration. In this paper I try to examine this phenomenon.The institutional structure of a scientific organization creates the circumstances for research work. As in the information society generally, scientific organizations have moved from a hierarchical form to an equal form organization. However, the social context in which scientific knowledge is produced is an important aspect, when collaborative work is considered. Crane (1972) has argued that beside the formal organizations in science there has been an unstructured, informal communication network. Wray differentiates collaborative research from collective research. Collective research has been a traditional convention in the sciences and it is typical that credit and responsibility rest only with one person. In contrast, however, at present collaborative research is more common than collective research. In collaborative research every member of the research group has an equal epistemic role and credit and responsibility are shared (Wray 2002). In these situations the interaction between researchers is reciprocal and all members have an equal epistemic position. Thus, an equal role between young scientists and senior scientist is also educational goal for research groups. Thagard (1997) emphasizes that not all collaboration is similar in nature. His analysis of the different forms of collaboration is more detailed than Wray's analysis. Whereas Wray argues that equal roles between members of the research group is the basic starting point for collaborative research, Thagard emphasizes that technical assistants and young students in research groups cannot have an equal epistemic role in the research groups. Scientific practice involves many cognitively demanding skills such as statistical analyses and highly constrained publication formats (Klahr 2000). The acquisition of these skills spans a long time and participation in the research group's work is important from an educational point of view. In this respect the nature of interaction between senior scientists and young students is inevitably educational. Thagard (1997) assesses the value of collaboration using Goldman's (1992) five standards for epistemic practice. In addition, Thagard adds a sixth standard, explanatory efficacy of a practice. By contrast, Wray assesses broadly scientific practices based on Kincaid's (1996) functional analyses. Both Thagard and Wray regard scientific journals as main indicators of the quality of scientific practices. However, these kinds of analyses reveal scientific practices from an institutional point of view. More research is needed relating to the quality of interaction between scientists, when they are working in a real social context. Especially, we need more research into how senior scientists can support the student's and young scientist's role as an active learner in research groups.