The development of the National Fire Incident Reporting System in the late 1970s made detailed, representative national fire statistics possible for the first time. However, calculation rules used to produce these statistics have varied among users. The authors present a detailed consensus procedure for such calculations and the supporting rationale.