Volume 16, Issue 1 (January 1988)
Sampling Variance: Bias and Reproducibility
Product quality necessitates the collection of a representative sample for analysis. The representativeness of a sample is measured by its accuracy or variance. Sampling variance is discussed in two parts, bias and reproducibility. Bias is an error that consistently results in either high or low analysis. Test methods for bias are described in detail. Reproducibility is a measure of expected fluctuation around the true result. Reproducibility can be reduced by additional increments, analysis, or both.