Summary
The essential problem in representative sampling is a comparison of analytical and sampling errors. Application of analysis of variance (ANOVA) to this problem seems less appropriate than regression analysis based on Wilsons formula, because in the first approach the total analytical error is commonly underestimated. Furthermore, regression analysis allows derivation of confidence intervals for representative sample weights. In this context a parametric and a non-parametric procedure are described.
Similar content being viewed by others
References
Aston SR, Oregioni B (1987) At Spectrosc 8:155–158
Doerffel K, Eckschlager K (1981) Optimale Strategien in der Analytik. Verlag für Grundstoffindustrie, Leipzig
Danzer K, Doerffel K, Ehrhardt H, Geissler M, Ehrlich G, Gadow P (1979) Anal Chim Acta 105:1–7
Skogerboe RK (1971) Appl Spectrosc 25:259
Krause B, Metzler B (1988) Angewandte Statistik. Verlag der Wissenschaften, Berlin
Kratochvil B, Goewie CE, Taylor JK (1986) Trends Anal Chem 5:253–257
Wilson AD (1964) Analyst 89:18–30
Ingamells CO, Switzer P (1973) Talanta 20:547–568
Stoeppler M, Kurfürst U, Grobecker KH (1985) Fresenius Z Anal Chem 322:687–691
Dörfel H (1976) Biom Z 18:651–659
Efron B (1979) Ann Statist 9:1218–1228
Robinson J (1987) In: Puri ML, Vilaplana JP, Wertz W (eds) New perspectives in theoretical and applied statistics. Wiley, New York, pp 243–255
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Henrion, R., Henrion, G., Szukalski, K. et al. Parametric and bootstrap estimations of confidence intervals for representative sample weights. Fresenius J Anal Chem 340, 1–5 (1991). https://doi.org/10.1007/BF00324381
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF00324381