Published online by Cambridge University Press: 06 March 2019
Sampling is a necessary part of the chemical analysis of particulate matter where the objective is to characterize bulk properties since it is usually undesirable or impossible to test an entire lot. The sample must be a miniature replica of the bulk material at least in respect to those features being tested. In other words, the sample must be representative to permit extrapolation from the sample to the bulk.
There is error associated with sampling particulate material, but this sampling error is only one component of the total error associated with an analytical result. Sample preparation and determination of the property being tested are two other major sources of error, These three errors combine as the squares of their standard deviations, i.e., their variances, to produce the total analytical error. Very often, sampling error is the largest of the three and, therefore, contributes a disproportionately large share to the total error.