Appropriate Critical Values When Testing for a Single Multivariate Outlier by Using the Mahalanobis Distance

Abstract

SUMMARY The Mahalanobis distance is a well‐known criterion which may be used for detecting outliers in multivariate data. However, there are some discrepancies about which critical values are suitable for this purpose. Following a comparison with Wilks's method, this paper shows that the previously recommended (p(n – 1)/(n – p)}Fp,n–p are unsuitable, and p(n – 1)2 Fp,n–p–t /n(n – p – 1 + pFp,n‐p–1) are the correct critical values when searching for a single outlier. The importance of which critical values should be used is illustrated when searching for a single outlier in a clinical laboratory data set containing 10 patients and five variables. The jackknifed Mahalanobis distance is also discussed and the relevant critical values are given. Finally, upper bounds for the usual Mahalanobis distance and the jackknifed version are discussed.

A socio-interpersonal approach to adjustment disorder: the example of involuntary job loss

“…We performed multivariate outlier analysis using the Mahalanobis distance (Penny, 1996). Five cases were excluded from the analysis because they were multiple outliers on the scales of interest.…”

On the meaning and use of kurtosis.

“…In addition to the univariate statistics, the macro gives for multivariate data (a) Mardia's (1970) multivariate kurtosis; (b) Srivistava's (1984) and Small's (1980) measures and tests of multivariate kurtosis and skew, both of which are discussed by Looney (1995); (c) an omnibus test of multivariate normality based on Small's statistics (see Looney, 1995); (d) a list of the five cases with the largest squared Mahalanobis distances; (e) a plot of the squared Mahalanobis distances, which is useful for checking multivariate normality and for detecting multivariate outliers; and (f) Bonferroni adjusted critical values for testing for a single multivariate outlier by using the Mahalanobis distance, as discussed by Penny (1996), who also noted that the test gives results equivalent to those obtained by using jackknifed Mahalanobis distances.…”

The effect of methods of eliminating spikes in the time series of freight flows on their statistical characteristics

“…Spikes in stationary time series are found like in a usual snap sampling by means of various criteria: Irving, Romanovsky, range of deviation, Dixon, Smirnov, Chauvene, Tityen-Moor, Rosner (Thompson's rule), et al [8][9][10][11][12].…”