Beck and Sznaider call on ‘methodological cosmopolitanism’ to transcend methodological nationalism and account for an increasingly cosmopolitanized reality. We take up their challenge by drawing on our experiences of conducting a collaborative ethnography of methodological changes in the production of population statistics within and between European national and international statistical institutes. Drawing on debates in science and technology studies, we depart from some conceptual presuppositions of methodological cosmopolitanism to define a ‘transversal method’. Referring to this method as performative and ontopolitical, we reflect on how it requires collaboration and, in our ethnography, gave rise to three practical challenges – (1) going beyond the individual project; (2) using each other's field notes; (3) and working against the national order of things. To meet these challenges, we reflect on how this method required us to practise three modes of care – thinking with others, tinkering with field notes, and dissenting within.