Concerned individuals have been trying to determine the safety of their food since ancient times. In ancient times, people themselves were the ultimate test of food safety, but as human evolution progressed, other techniques such as sensory perception and experimentation on animals were used. Today sophisticated analytical techniques and models are available to measure and predict food safety. These sophisticated techniques and models are dependent not only on the quality of samples that are collected and analyzed but also on how inferences are made from the analytical results to the food being sampled. Unfortunately, the Theory of Sampling and the role of inference have not been fully integrated into prediction of food safety.The basis for many "modern" food sampling protocols was developed prior to the development of the Theory of Sampling. Many of these sampling protocols were based on concepts of acceptance sampling procedures and associated inference. The Theory of Sampling enables the representative sampling of bulk materials and eliminates the reliance of acceptance sampling as the only method for the characterization of food and utilizes a different type of inference than for acceptance sampling. This contribution addresses the differences between inference for acceptance sampling and inference for the sampling of bulk materials and the implications of these differences for food safety.
Publication date: 9 June 2015