Synthetic Census Data Producing Realities

What is the boundary between imputation and the production of census data? How does this ambiguity influence the way censuses are conducted and how societies perceive their capabilities and limitations? This paper examines how traditional imputation practices are evolving in statistical offices across the global North and South, shifting towards methods that incorporate the production of synthetic data. The lack of explicit norms and organizational controls leads census offices to establish data production parameters based on subjective and often opaque criteria. By analyzing the cases of the United States Census Bureau and the National Institute of Statistics of Ecuador, we illustrate how synthetic data production became central to the census results in the 2020 census round. We then explore the implications of this transition at both sociological and public policy levels.