What is it about?
Authors have attached great kudos to estimation without bias. We show that this is not conducive to efficient estimation, which we regard as the main goal. Some efficient estimators of variance are derived. They are no more complex than some established (unbiased) estimators.
Featured Image
Why is it important?
Our approach dismisses the general idea that estimation should foremost be (approximately) unbiased, and then have small sampling variance. Bias and sampling variance should be dealt with simultaneously, by minimising the mean squared error, making no compromises on the way.
Perspectives
Mean squared error should be considered without any compromise when estimation any quantity of interest.
Dr Nicholas T Longford
SNTL
Read the Original
This page is a summary of: On the inefficiency of the restricted maximum likelihood, Statistica Neerlandica, January 2015, Wiley,
DOI: 10.1111/stan.12055.
You can read the full text:
Contributors
The following have contributed to this page
