A.6.2 Bulk Material Sampling

A sampling methodology is considered unbiased/correct if all elements/items present in the SU have exactly the same probability/chance of being selected in the sample. In practice, bulk materials (e.g., soils, sediments, ores, mining waste, liquids) are heterogeneous in nature as they are made of particles/molecules of various density, moisture content, shapes, and sizes with many nondistinct particles present at each location of the DU. Bulk material present in a DU cannot be viewed as a set of distinct objects (e.g., people in a meeting room). It is not possible to select items (molecules, particulates) one by one from a batch (DU) of bulk material in an unbiased manner using simple random sampling. This method introduces GSE (and therefore FE) in the collected bulk material sample. Instead, a group of particles making an increment of specified/practical mass is collected using an appropriate sampling tool (Gerlach and Nocerino 2003). Sampling bias or nonrepresentativeness is introduced due to the distribution (e.g., segregation or grouping) of material particulates contained in the DU. Elder, Thompson, and Myers (1980) showed that even when sampling from bulk materials (consisting of many coexisting and nondistinct particles), the use of simple random sampling yields unbiased estimates of the mean with minimum variance (Smith 2006). The simulation experiments reported in this appendix support Elder, Thompson, and Myers’ findings.