The classical experiment where the concentrations of two analytes are determined in a mixture by measuring its absorbance at two different wavelengths is reviewed with specific emphasis on prediction uncertainty. It is shown that the standard prediction error on each concentration can be estimated by an equation that bears a strong resemblance to the well-known expression for the univariate case. Aqueous mixtures of potassium permanganate and potassium dichromate at an appropriately regulated pH are studied, and the concentrations are calculated along with the proper uncertainty estimates. Clues are given as to how the different experimental error sources propagate to the reported concentration values.
Cabezón et al. (Tue,) studied this question.