Abstract
Two main approaches to estimating parameters of dominant height curves (derived by algebraic difference approach (ADA) and generalized ADA (GADA)), the dummy variable method and the difference approach (the error-in-variable (EIV) method along with ordinary least-squares (OLS) fits)are compared using second-rotation loblolly pine (Pinus taeda L.) data from a designed experiment. It was found that the EIV method consistently performed better than the dummy variable method, and the two unbiased estimation methods were not necessarily better than the biased OLS differencemethods. In general, the performance of these two approaches was similar in height prediction given site index, but in combination with site index prediction given heights, the difference methods generally performed better. The nonoverlapping forward difference data structure (III), is recommendedto fit difference models.