Estimates of Precision in a Standard Additions Analysis

Using the method of standard additions, two different methods of estimating the standard deviation and associated confidence interval for the concentration of an analyte are quoted in commonly used textbooks. It is shown, numerically and algebraically, that a method using the law of propagation of errors invariably gives lower values for the standard deviation than a second technique, which uses extrapolation. On examining the statistical reasoning used in the law of propagation of errors method, it is found that certain covariance terms are assumed to be negligible. Only when combining variables that are truly independent is this assumption valid, which is not the case in standard additions analysis. Consequently the equation used in the extrapolation method is recommended in estimating the standard deviation and confidence interval.