In addition to conventional sample size tables, few formulae were developed according to risk-based approaches and used for calculating the size of food commodity samples for inspection purposes. The current paper hypothesized the dependency of the sample size on both the risk level of the commodity or establishment and the confidence level of sampling. Accordingly, a sampling formula was developed using the commonly used 95% confidence level as fixed attribute. Application of the developed formula on populations and lots selected from three different sources, sample size tables, official authorities’ information, and calculated number of units for a fixed lot of hay (2,400 tonnes), revealed no significant difference between the sample sizes at the two selected risk probabilities 0.99 and 0.75. The findings of the current paper strongly support the use of lot units as basis for calculating the sample size rather than the lot weight, further the use of individual lot sizes to calculate sample sizes is more realistic than the common use of groups of lot and populations. The developed formula could confidently be used for calculating sample sizes for commodities of known risk probabilities.
Keywords: Risk Analysis, confidence level, inspection, risk probability, risk-based sampling (RBS), sample size
