Marlot Kuiper

275 Conclusion that treats routines as collective entities relies on analogies between routines as organisational level constructs, and individual skills, but provides very little analysis of their links (Felin & Foss, 2009; Salvato & Rerup, 2011). Concepts from the Sociology of Professions literature like socialization and hierarchy proved helpful in understanding collective processes, but future research could focus more on how individual ideas and skills play out at the collective level. Lack of longitudinal data Thirdly, routines are about stability and change. I conducted episodic observations at different points in time, which allowed me to see change. For instance, when I returned at Plainsboro after a while, the whiteboards representing the checklist had been introduced and I could observe how this new artefact found its way into practice. However, gradual change was difficult to see. A longitudinal study might be worthwhile to closely see how routines evolve over time. Besides, as routines are argued to be “stable for now” (Feldman et al., 2016; Pentland & Feldman, 2008) we should be aware that the patterns observed in this study, might have evolved into different patterns today. Multiple research sites Fourthly, for this study I conducted fieldwork at two research sites. I would like to underline that every case is unique, and that the aim of this study was not strict comparison. Rather, I used data from two research sites to generate a more comprehensive understanding of the research phenomenon. Data collection was not equally divided among the research sites. I started off with observations at Plainsboro, and later on decided to move to St. Sebastian’s to look for overarching storylines, and thought-provoking differences. This dissertation shows that many of the mechanisms (re)creating routines hold for both research sites. There clearly are bigger storylines, for example when it comes to how professionals deal with conflicting routine demands. As said, differences were often more clearly visible at the team level than at the organisational level, for example when it comes to entrusting team members. The most evident difference between the research sites, was the way in which artefacts were used to model the routine. Chapter 7 therefore most clearly differentiates between the research sites. Despite artefacts’ different affordances, similar patterns of action could emerge, for example ticking off boxes in the software system our drawing a continuous line across the items on paper. In making sense of similarities and differences that occurred, it proved helpful that some of the research participants ‘appeared’ at both research sides. For educational purposes, residents for example worked in St. Sebastian’s after they had worked in Plainsboro, so we could discuss what was going on at both 8