Inferences about Robust Measures of Location

Abstract

For most purposes, simply estimating a measure of location is not sufficient. There is the issue of assessing the precision of an estimate, and of course there is the related problem of testing hypotheses. How do we test hypotheses or compute confidence intervals with a trimmed mean or an M-estimator of location? To what extent do such methods address the problems with Student’s T listed in Chapter 5? These issues are discussed in this chapter.