The successive remeasurement of trees and plots is common practice in forestry research. But the usual regression techniques are not generally suitable for deriving prediction equations, variances, and tests of hypotheses from systematic remeasurement data. Assuming a model of the form: u = a + Bx, an analytical method is proposed based upon the fact that each experimental unit (plot) provides estimates of a and B. Thus, the variances of these estimates may be calculated directly; and the variance of estimated u may be considered as a linear combination of variances and covariances. The method is easily extended to the case of two or more independent variables.
Document Type: Journal Article
Silviculturist, Northeastern Forest Expt. Sta., Forest Service, U. S. Dept. Agric, Upper Darby, Pa.