Composite estimators for growth derived from repeated plot measurements of positively-asymmetric interval lengths
The statistical properties of candidate methods to adjust for the bias in growth estimates obtained from observations on increasing interval lengths are compared and contrasted against a standard set of estimands. This standard set of estimands is offered here as a solution to a varying set of user expectations that can arise from the jargon surrounding a particular data aggregation procedure developed within the USDA’s Forest Inventory and Analysis Program, specifically the term “average annual” growth. The definition of a standard set of estimands also allows estimators to be defined and the statistical properties of those estimators to be evaluated. The estimators are evaluated in a simulation for their effectiveness in the presence of a simple distribution of positively-asymmetric measurement intervals, such as what might arise subsequent to a reduction in budget being applied to a national forest inventory.