Best Linear Unbiased Estimators for Stereology

Precise criteria have been published recently for obtaining unbiased ratio estimators of structural parameters, defined in an n-dimensional opaque specimen, from observations in lower-dimensional sections. In this paper, the possibility is shown of obtaining linear unbiased estimators of minimum variance whenever the data can be described by a linear regression model through the origin. Estimation from singleand two-stage sampling is discussed; the former is illustrated by an example.