Workshop on Variance Estimation in Stereology

22-26 November, 2004, Aarhus University, Denmark

Building 252, Room 204

Partially funded by MaPhySto.


The precision of stereological estimators is a question of great practical importance and has been investigated actively by stereologists, statisticians, geologists, materials scientists, cartographers, computer scientists and others. Variance is a complex problem for several reasons. Firstly, the basic theory of stereology, like survey sampling theory, only guarantees that parameter estimates are unbiased, and has little to say about variances in general. Secondly, there are no simple, exact formulae for the variance of a stereological estimator. The variance depends on covariance terms that depends on the spatial organization of the material.

Recent research in design-based variance estimation in stereology has focused on developing procedures that are generally applicable. There is a need for further investigation of these variance estimation procedures and comparison with what can be obtained by probabilistic modelling or statistical modelling and data analysis. Covariance modelling for spatial stochastic processes is an important issue that may imply new variance estimation procedures in stereology.

This workshop brings together scientists from the model-based and design-based approaches. A number of scientists applying variance estimation, especially in the biosciences, are also invited.


Additional information can be obtained from:

Oddbjørg Wethelund
Department of Mathematical Sciences
University of Aarhus
Ny Munkegade 118, Bldg. 521
DK - 8000 Aarhus C.