Impact of Sampling Concepts on the Effectiveness of Microbiological Methodologies

Abstract

As discussed in Chap. 6, there are a number of underlying assumptions that are typically made when developing microbiological sampling plans. For example, it is generally recognized that a basic statistical assumption used for lot evaluations is that the samples are taken randomly to minimize sampling biases. However, there are number of other methodological assumptions that are often made for which there is less understanding in relation to their impact on the effectiveness of microbiological detection. For example, it is commonly assumed in selecting sampling plans that the microbiological methods used are fully effective at recovering the target organism if it is present in a food sample. However, in reality one often has to deal with type I errors (true negatives testing positive) and type II errors (true positives testing negative (AOAC 2006). While great efforts are made to develop and employ methods that meet these methodological assumptions, there are a number of conditions where deviation from these assumptions could have a significant impact on the ability of sample plans to achieve the desired level of confidence. The current chapter will explore several of the methodological factors that could impact the effectiveness of sampling plans, and approaches for quantitatively estimating the decrease or increase in the level of confidence provided by the sampling plan under those circumstances.