Answers are grounded in this paper and cite their sources.
Suppose that a sample of size n from a continuous and symmetric population with an unknown parameter is given. We consider a fictitious random subsample of size k drawn from the original sample and construct the best linear estimator based on the subsample. Applying the Rao-Blackwell type argument, we get an estimator which uses the information contained in the whole sample and is supposed to be uniformly efficient for a wide class of distributions. Monte Carlo experiments established that this estimator is highly efficient for small samples of size 10 to 20.
No takes yet. Share an insight, caveat, or question.
Kei Takeuchi (1971) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: