A New Method for the Improvement of Data

Abstract

Introduction – The use of the average analytical signal for the construction of curves by the least squares method (LSM) over the standard addition method (SAM) is widespread. It would be advantageous, however, to fi nd a way to avoid intermediary averages, which are known to be the cause of signifi cant increases in standard deviations (SD). Objective – To develop a protocol that uses all gathered data to create curves by LSM over SAM. To use Excel® for the estimation of y = mx + b and R2 rather than using LSM equations for the SD of m, x and b. Methodology – The level of lead (II) in the bark (cork) of Quercus suber Linnaeus was determined using diff erential pulse anodic stripping voltammetry (DPASV). Three current samples were taken for each of the four standard additions. These signals were combined for adjustment by LSM. The results were compared with those obtained after averaging the current for each addition, and the expression of uncertainty in the measurements determined. Results – The new method shows an expanded uncertainty of ± 0.3321 μg/g (nearly 1.42%). The diff erence between the results obtained by the new and the old method is 0.01 μg/g (23.41 and 23.40 μg/g). The limit of detection changed approximately from 4.8 to 4 μg/g and the relative SD approximately from 9 to 6%. Conclusion – The absence of intermediary averages in curves improved the determination of lead (II) in cork by DPASV. Estimation of SD only with LSM equations produced results that were signifi cantly worse. The changes are large enough to transform an apparently internally non-validated procedure (repeatability for precision) into an internally validated procedure.