Book review: Robust Estimation and Testing by R G Staudte and S J Sheather

Robust procedures are most informative when their results are compared to those from classical procedures. In other words, statistical scientists should always think about doing at least two analyses of their data, and comparing the results. The book under review makes the important subject of robustness more accessible to statisticians; for those who know little but would like to learn more, it is the first place that I would recommend they look