Research groups are a historically contingent expression of the fact that what we simply call “research" is an involved effort that takes effect through multiple forms of collaboration (Katz & Martin, 1997; Maien- schein, 1993). While social epistemologists and philosophers of science have only just begun to show interest in the structures of collaboration other than specialist communities, social scientists have long studied research groups as a diverse, multifaceted empirical phenomenon. For this reason, this section draws on the work of social scientists to characterize research groups as a distinct, but utterly varied, empirical phenomenon of scientific collaboration.
The significance of scientific collaboration for contemporary science plays out in numbers. As collaborative authorship is arguably a robust measure for close research collaboration (Beaver & Rosen, 1978; Clarke, 1964; Meadows & O’Connor, 1971), bibliometric studies suggest that close collaboration has become the preferred way of conducting research in many fields. The last century witnessed a continuous increase in coauthored papers in almost every scientific discipline, and especially in experimental natural science multi-authored papers dominate by far. Scholars have also found that research collaboration is co-related with individual productivity and visibility (Beaver & Rosen, 1979b; Lee & Bozeman, 2005; Zuckerman, 1967).
Research groups are a crucial element of scientific collaboration, as Edward Hackett observes: “Research groups are an elemental form of scientific collaboration and knowledge production" (Hackett, 2005, p. 788). Scientific collaboration can take many forms but “small batch production" in research groups has become typical for the way scientific knowledge is created at Western universities at least since the mid-twentieth century (Hagstrom, 1974a, p. 757). In a similar vein, Henry Etzkowitz calls small and medium-sized research groups of “moderate scale" the distinctive feature of science at American universities as it has emerged in the last century (Etzkowitz, 1992, p. 28).
Usually, a university-based research group has between three and thirty members. In terms of their academic personnel, most research groups would consist of at least one senior scientist, commonly in a tenured position, who assumes group leader responsibilities, a number of students and post-doctoral fellows as well as further senior scientists. Additionally, technicians and administrative staff may support the research group. The size of research groups can change quickly when funding runs out or new research opportunities open up (Etzkowitz, 1992, p. 36). Yet, for a group to function, it should be small enough so that its members can develop ongoing face-to-face relationships. A group is too small when it cannot ensure enough varied interaction among group members. A group is too large when its members’ activities cannot be “kept on track," when their activities cannot be sufficiently coordinated, especially when junior scientists cannot be sufficiently supervised and trained and when they may produce significant results that senior group members miss out on(Etzkowitz, 1992, p. 38). The less routine the work, the smaller a group has to be to ensure an accurate exchange of information among its various members (Hagstrom, 1974a, p. 756).
Research groups can be coordinated in different ways. A great variety of factors shape research group leadership, such as the degree of interdisciplinarity involved, work ethos and employment culture, but also group leaders’ professional experience, their academic biography and personality. While some groups centralize leadership, other groups will distribute its responsibilities among its members (Cohen, Kruse, & Anbar, 1982). A crucial responsibility of group leadership is to manage the scientific focus of group members’ research activities. Here, group leaders focus on an ambivalent choice, as Hackett observes: “tighter focus may intensify competition, while looser focus may dissipate energy and relinquish fruitful synergies" (Hackett, 2005, p. 818). While some competition among group members may increase their initiative, too much competition constrains collaboration among group members and may prove detrimental to the group’s research performance. Furthermore, group leadership needs to balance continuity in research focus against “the continual, incremental freshening of a group’s membership and capabilities that expands the boundaries of its sphere of inquiry" (Hackett, 2005, p. 794), continuously adapting to the ways in which scientific fields evolve.
Research group members, as Hackett observes, typically “work face-to- face, sharing work space, materials, technologies, objectives, hypotheses and, to a significant degree, a professional reputation and fate" (Hackett, 2005, p. 788). Shared resource access, as the historian Jane Maienschein notes, is a major motivation for an individual scientist to enter a research collaboration in the first place, as they are “eligible for resources that individual researchers could not obtain" (Maienschein, 1993, p. 167). The sharing of resources helps increase both the efficiency of scientific labor and the credibility of research outcome. Increasing efficiency may “simply be a matter of needing more hands doing the same kind of work, or it may involve bringing together specialists who provide different types of expertise" (Maienschein, 1993, p. 167). To bundle specialist expertise through collaboration can have a positive impact on the credibility of a scientific knowledge claim: “collaborations among different individuals may produce greater credibility because each brings to the project his or her own credentials and acceptability in a different research community" (Maienschein, 1993, p. 167). For research groups, it can be a challenge to keep these cognitive, technical, social and financial resources “in stock" despite the constant fluctuation of group members.
Beyond resource sharing, research groups also facilitate hands-on learning. Face-to-face collaboration helps group members to acquire the tacit skills that are an important element of scientific expertise, helping junior scientists mature professionally—a responsibility that particularly research groups in academic contexts attend to. In fact, for university research groups, the interaction between senior and junior scientists will often be one of the collaborative relations that determine the way in which a group divides labor among its members.
The collaborative relations within a research group can be multifarious, and a number of typologies have been suggested to describe them. Drawing on scientometric studies, Subramanyam (1983), for example, lists teacher-pupil collaborations, collaborations among colleagues, supervisor-assistant collaborations and researcher—consultant collaborations. Toomela (2007) distinguishes “dialogical collaboration" and “unidirectional collaboration." While the first designates a truly collective research effort, the latter is ultimately an individual enterprise where only one person determines the direction ofresearch. This person hires in other “helping hands" and “basically absorbs knowledge provided by others" (Toomela, 2007, p. 202). For philosophy and epistemology of science,
Thagard (1997) suggests a typology that consists of “employer/employee,” “teacher/apprentice,” “peer-similar” and “peer-different.” A monodisciplinary collaboration takes place among peers with a similar, largely overlapping expertise, while an interdisciplinary collaboration involves scientific peers with utterly different fields of expertise (see also Sect. 2.3). Such typologies, however, say little about the collaborative practice by which individual efforts are entwined. In fact, the task of later chapters (such as, e.g., Chaps. 6, 7, and 8) in this book will be to investigate how different types of relation frame dependence and facilitate the trust that it takes for group members to rely upon one another.
-  Scientometric studies have produced a large body of data testifying to the increase in multi-authored publications (Balog, 1979/1980; Beaver & Rosen, 1978,1979a,b; Braun, Gomez, Mendez,& Schubert, 1992; Cronin, 2005; Cronin, Shaw, & Barre, 2004; Meadows, 1974; Meadows &O’Connor, 1971; Moody, 2004; Wagner & Leydesdorff, 2005). For an overview of the literaturesee Subramanyam (1983) and Katz and Martin (1997). Note that the number of colleagues thankedin acknowledgments has also increased (Cronin, 2005; Cronin et al., 2004).
-  For competition among members of a research group see Edge and Mulkay (1976); Hagstrom(1974b), and for the complicated mediation between collaboration and competition in researchgroups see Traweek (1988, p. 88); Poulsen (2001).