Sampling is more than shoveling material into a bucket. It is even more than using adequate mass, increments, and tools. Sampling is a systematic process that incorporates everything from development of objectives through final decision-making. Many sampling protocols currently in use focus only on the physical sample collection and ignore the preceding steps in the sampling process. The ignored steps include development of the critical decision objectives, integration of sufficient quality control, inferences from test portions to lots, and final decision making, statistical or otherwise. Without this supporting framework, it is impossible to ascertain the validity of the sampling protocols when needs or objectives change. Often, the same sampling protocol is implemented year after year without any consideration to its appropriateness.Proper Sample Quality Criteria (or Data Quality Objectives) are determined from the objectives of the project and must be an integral part of any sampling campaign. The major components of the Sample Quality Criteria are: 1) Question, 2) Decision Unit, and 3) Confidence. The Decision Unit is the specific material to which an inference from the analytical result is made and ultimately to which a decision is made. If the Decision Unit is not precisely determined and integrated into the development of the sampling protocol, the resulting decisions will be incorrect or, at a minimum, will not be cost effective. This contribution addresses development and integration of the Decision Unit into the sampling protocol framework.
Publication date: 9 June 2015