It is dicult to detect outliers in p-variate data when p > 2 because one can no longer
rely on visual inspection. While it is still quite easy to detect a single outlier by means
of the Mahalanobis distances, this approach no longer suces for multiple outliers because
of the masking eect, by which multiple outliers do not necessarily have large Mahalanobis
distances. It is better to use distances based on robust estimators of multivariate location and
scatter (Rousseeuw and Leroy 1987, pages 265{269). In regression analysis, robust distances
computed from the explanatory variables allow us to detect leverage points. Moreover, robust
estimation of multivariate location and scatter is the key tool to robustify other multivariate
techniques such as principal component analysis and discriminant analysis.