The Behaviour of Estimators of the Parameters of Various Yield-Density Relationships

Abstract

The yield-density equations of Bleasdale and Nelder (1960) and Holliday (1960), are compared in a series of simulation studies. The results show that the maximum likelihood estimators (MLE's) of two of the parameters of the Bleasdale-Nelder equation can produce markedly biased estimates with positively skewed dlstributions. These deficiencies are only partially overcome when a reparameterized Bleasdale-Nelder equation is used. The estimators of the parameters of the Holliday equation are effiectively unbiased and normally distributed and this equation appears to be generally suited for yield-density studies. The method of Box (1971) is recommended as a useful predictor of bias in least squares estimators of parameters of non-linear regression models.