Sampling is not gambling
Analytical results form the basis for decision making in science, technology, industry, society – and they must therefore be relevant, valid and reliable.
But analytical results cannot be detached from the specific conditions under which they originate. Sampling comes to the fore as a critical success factor before analysis, which should only be made on documented representative samples.
Sampling: Counteracting heterogeneity
All lots, materials and process are heterogeneous. Heterogeneity is the arch enemy of all sampling. The pathway from a typical lot in science, technology and industry is a multi-step mass-reduction process with sampling rates of 1 : 106 to 1 : 109 or higher.
But while analytical results are only representative of the miniscule aliquot mass, decisions must relate to the million/billion times larger original lots. A single ‘grab sample’ can never be representative! Being able to conduct representative sampling across six to nine orders of magnitude is the primary objective of the Theory of Sampling (TOS).
Taking into account all factors impacting on sampling accuracy (bias) and precision (uncertainty) requires the relevant competence in all steps “from-lot-to-aliquot”. Sampling must at all stages counteract heterogeneity in a fully documented fashion. TOS provides the only fully developed scientific and practical framework available.