An elementary model for statistical lot inspection and its application to sampling by variables

Recently some critical papers have initiated a discussion pros and cons of standard sampling by variables (sampling by measurement). This discussion demonstrates the need for clarification by a sound stochastic model of statistical lot inspection. On the basis of such a model, variables sampling for lot proportion nonconforming and attributes sampling can be compared in terms of the theory of statistical hypothesis testing. This comparison refutes the well-known dogma of ‘same protection at lower sample size with variables plans’. Sampling by measurement for lot means, however, emerges as a perfectly valid method of statistical lot inspection.