A note on the use of the non-central t-distribution in setting numerical microbiological specifications for foods

An exact procedure is described and used to determine k values and probabilities of rejection used in the application of variables sampling plans to foods. Comparison of quantities so obtained to those calculated by Kilsby et al. (1979), who used an approximation procedure, reveals that for small samples the approximation can lead to probabilities of rejection which are larger than those obtained using the exact procedure. It is recommended that the exact procedure be used in the application of these plans.