Answers are grounded in this paper and cite their sources.
For the problem of efficiency-robust estimation of location, approximate versions are developed of optimally robust Pitman-type estimators. These are shown to have full asymptotic efficiency for a prototype family of distributions, used to define the estimators. The asymptotic efficiency under other distributions of interest is also characterized. For sample sizes n = 20, 30, 40, 50, and 100 the efficiencies were estimated by Monte Carlo methods under the following distributions: normal, logistic, double-exponential, and contaminated normal (one percent, five percent, ten percent). Over all these shapes the efficiencies obtained are approximately 88 percent or more for all n; they rise to approximately 91 percent or more for n = 100. Some theoretical and numerical comparisons with other estimators are given.
No takes yet. Share an insight, caveat, or question.
Birnbaum et al. (1970) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: