Reduction of the relative error in the standard additions method

Abstract

Reduction of the relative error in the standard additions method is suggested by presenting the equations and graphs that enable the coefficient of variation of the estimate to be computed easily. The graphs show the most efficient way of computing data and for the design of measurement for a given set of method, apparatus and effort, which correspond to three cases of the variance character: the variance of the signal is either constant, proportional to the signal or proportional to the square of the signal. Adoption of the best procedure described will reduce the coefficient of variation of the estimate by a factor of two over the common method, with the same expenditure of effort.