(Statist) An error range r around a value such that a measurement of the value is equally likely to be within or outside the range of ±r