A Critical Examination of the Statistical Techniques Available for Estimating Exclusion Limits from In-Grade Tests

摘要

The strength of wood products, exhibits variability that is in general the sum of both random and nonrandom components. The latter can be associated with such factors as variations in the wood supply, technological innovations or other changes in processing and hence related to specific combinations of location, time and manufacturer. In this paper the relevant statistical theory is developed, and it is shown that, within this framework, and apart from exceptional circumstances, there does not seem to be a satisfactory solution to the problem. In other words, under what could be called the conventional approach, statistical theory in itself does not have the power to provide definitive answers.