Correcting for nonrandom errors in a fixed-effect model
- 1 December 1989
- journal article
- research article
- Published by Canadian Science Publishing in Canadian Journal of Forest Research
- Vol. 19 (12) , 1550-1554
- https://doi.org/10.1139/x89-236
Abstract
In this article, a statistical method of estimating the microsite component (a nonrandom error) in a fixed-effect model is presented and discussed. The approach uses generalized least squares to partition the model variance into two components: one part due to microsite and the other one due to random errors. An iterative procedure is then used to solve for the maximum likelihood estimate of the microsite component. Application of the technique to a slash pine (Pinuselliottii Engelm. var. elliottii) spacing trial indicates that 30% of the variability in total tree height at age 6 years was due to microsite, whereas the remaining 70% was due to random errors.Keywords
This publication has 0 references indexed in Scilit: