["Journal of Agrarian Change, EarlyView. ", "\nABSTRACT\nWhy have Indonesia's decennial agricultural censuses defined their key unit of analysis, the ‘agricultural household’, differently? How do changes to that category shape knowledge production among those who rely on census data? This paper compares definitions of the agricultural household across seven censuses with international recommendations to show Indonesia's state causes the shifts in meaning endogenously. Undertaking a critical assessment of the classifications and results reveals both the persistence in the scale of Indonesia's small‐scale farming and the state's construction of an entrepreneurial and male agrarian nation. Doing so is useful for situating studies of local agrarian change and as a basis for comparing how other states see their countrysides.\n"]