Year: 2026
Type: article
Abstract: The correlation between density and bending strength of timber is strong across averages for different species and growth areas. This often leads to the mistaken belief that it is also strong for the case of a particular species and growth area; giving the wrong impression of the importance of density when predicting strength. This paper presents an illustrative case study of spruce (mostly Sitka spruce and some Norway spruce) grown in the United Kingdom (with some also from Ireland). The relatively poor correlation between density and bending strength (R2 of approximately 0.18 in this case) is nevertheless shown to be of some limited use for strength grading, even though the calculation of R2 itself is not very informative. As has been known by statisticians for more than a century, the calculated values of Pearson’s correlation coefficient (R) from all but very large studies are very susceptible to random chance. This makes it very difficult to detect any genuine differences in the magnitude of the density-bending strength correlation between timber resources, although it does appear likely that the correlation is slightly different to that reported for European-grown Norway spruce in some regions.
Author: Dan Ridley-Ellis
Institution: Edinburgh Napier University
Language: English
Subfield: Building and Construction
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7143741929