The invention of victimization surveys is often presented as a synthesis of the two theoretical attitudes that, supposedly, dominated the 1960s debate over official crime statistics: realism and social constructionism. This paper turns this genesis story on its head. Using original archives, I argue that victimization surveys responded to organizational opportunities in the field of applied research. It was only after the fact that two of their architects seized the debate on crime measurement to broadcast their invention. In so doing they strategically recast the terms of this debate into a binary division between two antithetical social ontologies. This case is used to discuss how social scientists come to reinterpret and misunderstand their history.