On Statistical Testing Methods for Dimension Lumber Property Monitoring

Summary

No reviewed summary yet. Read the abstract below or open the original publication.

Abstract

The American Society of Testing and Materials (ASTM), D1990 details standard practices for establishing and maintaining design values for dimension lumber. Included in this standard are statistical methods for detecting changes in dimension lumber properties over time. A characteristic (or design) value, also known as the allowable property, is published for each lumber property, and this value must be updated whenever changes in the resource warrant. Currently ASTM D1990 calls for the use of the Wilcoxon Rank Sum test (also known as the Mann-Whitney test) to detect changes in these properties. In essence, this test is a two-sample test designed to detect changes in the underlying populations from whence the samples were drawn. In this work, FPL researchers recommend, with justification, that this practice be revised to the use of one-sample tests that focus on detecting disparities between the current resource properties and the corresponding currently accepted design values. We detail the tests that are recommended for properties whose design values are (based on) either the mean or some quantile. We further examine the impacts, in terms of statistical errors, of these alternative tests in comparison to the current paradigm through simulations using a collection of distributions modeled from actual within-grade softwood lumber modulus of elasticity and modulus of rupture data.

Abstract as supplied by the source. Full text may be restricted by the publisher.