ESTIMATOR BIAS IN A LAKE PHOSPHORUS MODEL WITH OBSERVATION ERROR

We investigated observation-error-induced parameter bias using least s quares and Stein-corrected least squares estimators in a model for pre dicting lake phosphorus. The Stein-corrected estimator performed bette r than the uncorrected estimator from bias and ''closeness'' perspecti ves, though the corrected estimator was still biased. Examination of t he model structure revealed that parameter bias is strongly related to both the parameter space and sample space. Additionally, the model is robust to parameter bias over a large portion of the sample and param eter space, indicating that this model may be particularly useful for estimation and prediction. Analogous structure in other models could b e an important consideration for model selection.