In child welfare knowledge production, administrative data hold a privileged status as the digital residue of state-client interaction, and are used in decision-making, service monitoring and evaluation, and research. Administrative data’s contemporary ascendancy is traced back to the 1990s due to improvements in computational infrastructures, data-sharing capabilities, and policy emphasis. Scenes of Measurement troubles this historicization, arguing administrative data pre-date both computation and formalized Federal child welfare policy, being the raw material of archival annual organizational reports. To this end, Scenes of Measurement conducts an interdisciplinary, close reading of administrative data asking, “what does data do?” towards understanding this process of representation. Guided by work in Black Studies and Media Studies, I perform a discourse analysis of the 1877 - 1923 Annual Reports from the Pennsylvania Society to Protect Children from Cruelty (SPCC) towards (re)defining and (re)contextualizing administrative data. Organizing my findings through Sylvia Wynter’s theory of autopoetics, I demonstrate how SPCC Reformers utilized their annual report as an interface, projecting and structuring representations of themselves, and children and families through Judeo-Christian, and Liberal-Humanist Bourgeois logics which manifested in data presented to would-be donors. Through this recursive structure, the racializing functions of administrative data emerged not as contradiction, but as a condition of utility. I argue that the administrative datum is an assemblage, or singular composite, produced by an entanglement of cosmogony, measurement subject and object, cultural context, and intended media and output, rather than being an objective signifier of people, places, and events. Scenes of Measurement offers historical and theoretical contributions to social welfare and critical data studies through its definition of administrative data, epistemological contributions through its theories and analysis of administrative data, and methodological contributions through its interdisciplinary approach to discourse analysis.