This paper demonstrates how analytical plurality can aid the study and understanding of emotions. Specifically, it describes an ethnography to explore emotions in palliative care and how they shape what clinicians do, why they do it, and how they do it. Furthermore, it illustrates how multiple analytical approaches were coalesced to examine and critique the role of individual and collective emotions in this context. This was aided by using critical incidents to catalyse inquiry into the emotional aftereffects that were experienced, individually and/or collectively. Qualitative data about these effects were collected via: interviews with clinicians, particularly at the time of a critical incident, to respectfully probe their thoughts, emotions, and behaviours; fieldwork notes when shadowing clinicians, as a form of naturalistic data; as well as a journal to capture researcher experiences, particularly the emotional effects associated with fieldwork. The datasets were analysed jointly to craft narratives. These narratives helped to map the ways in which critical incidents give rise to emotions that shape practices, within and beyond a health service, as well as the experience of palliative care for clinicians, patients, and carers. The narrative analysis developed a concrete, coherent view of emotions in this context, however it was also important to look beyond this neat frame and ‘undo’ the tidy narratives and attend to what might be shaping the constitutive practices. As such, each dataset was analysed discretely, which involved constructing themes with a focus on the discursive and performative patterns of emotion and the ways they create communities and legitimise political decisions through the work of affective economies. As this paper demonstrates, analytical plurality helped to recognise emotions as culturally-relevant, public performances that can be used to negotiate power relations and social practices. The value of this paper is twofold. First, it clarifies how ethnography can be respectfully appropriated in emotionally-charged contexts, like palliative care. Second, it demonstrates how analytical plurality can be used to blend different datasets on individual and social emotions and promote scholarship on emotions.