Suppose that the quality of products of a manufacturing process is measured on a continuous scale and that the frequency distribution of quality so measured can be appropriately described by a Gaussian distribution with unknown mean and standard deviation. Suppose quality is considered unsatisfactory when the quality measure is greater than some specified value or less than some other specified value. An approximate procedure, at any specified level of confidence is presented for estimating an upper limit to the proportion of unsatisfactory items that the process will produce. The procedure is approximate in the sense that the level of confidence is greater than specified. This work supplements the results available in statistical literature which give exact procedures when quality is considered unsatisfactory only if the quality measure is too large or only if the measure is too small.
展开▼