The Field Guide › Paper
A methodological companion to Edmondson and McManus's (2007) broader fit framework, addressing a specific problem that framework doesn't fully anticipate: most team research methods assume teams are bounded, stable, and self-aware, meaning they have a fixed, identifiable membership, that membership persists across the period being studied, and that members themselves recognise who is and isn't on the team. Increasingly, real teamwork happens in forms that violate one or more of these assumptions: fluid membership that rotates in and out as a task unfolds, permeable or contested boundaries, ad hoc configurations assembled for a single episode, and cross-functional or distributed arrangements where members may not share a settled, common understanding of who else is 'on the team' at any given moment. Applying standard team-research methods, which typically start by fixing a roster and then measuring properties of that fixed group, to teamwork of this kind risks studying an idealised, stable version of the phenomenon rather than the messier, dynamic reality practitioners actually navigate, and can systematically miss exactly the improvised, boundary-crossing coordination that makes fluid teamwork distinctive in the first place. The paper works through the specific methodological strain points this mismatch creates: defining the unit and level of analysis when the boundary itself is unstable or contested; timing data collection when team composition may have already changed by the time a survey or interview reaches participants; and interpreting team-level constructs like psychological safety when respondents may not agree on who the relevant 'team' even is. In response, it argues for adaptive research strategies that let team boundaries and membership emerge from the data rather than being fixed by the researcher in advance: qualitative and ethnographic methods that follow the actual flow of interaction and coordination as it happens; sampling and timing strategies sensitive to when composition is likely to shift; and a general shift from treating 'the team' as a stable unit to be measured toward treating teaming as an ongoing, adaptive activity to be traced, echoing Edmondson's broader distinction between team as noun and teaming as verb. The underlying argument is that as more real-world work moves toward exactly these fluid, boundary-crossing forms, sticking with methods built for stable, intact teams doesn't just under-measure the phenomenon, it risks quietly redefining the research questions the field asks toward whatever happens to still be measurable with old tools.