Robust Confidence Intervals for a Location Parameter: The Configural Approach

From MaRDI portal





This article discusses interval estimators for a location parameter in small to moderate samples. The approach uses the technique of conditioning on ancillary statistics, thus reducing the complexity of the problem. Two new types of confidence interval estimators using two- dimensional numerical integrations within location-and-scale- configurations emerge; firstly, strong confidence intervals which keep the conditional coverage probabilities above the nominal confidence coefficient for a finite number of distinct models (conditional version of robust validity) and secondly, bioptimal confidence intervals, which are robust in terms of efficiency. Robust intervals compromising the Gaussian and the heavy-tailed slash distributions serve as examples. A comparison with a variety of other confidence intervals shows the nonrobustness of rank-based interval estimators. For samples of size 5 Student's t-interval is nearly the best one can do, and for samples of sizes 10 and 20 biweights with tuning constants suitably chosen seem preferable among the interval estimators based on M-estimators.




Cited in
(26)








This page was built for publication: Robust Confidence Intervals for a Location Parameter: The Configural Approach

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3747502)