Upgrading from:

Statistics`MultiDescriptiveStatistics`

Univariate descriptive statistics have been added to the builtin Mathematica kernel.
Multivariate functionality from this package is included in the newly created Multivariate Statistics Package.
MultivariateMode is replaced by the system function Commonest.
CovarianceMatrix and CorrelationMatrix are replaced by the builtin functions Covariance and Correlation.
RegionProbability is replaced by EllipsoidProbability.
ConvexHullMedian and ConvexHullArea are included in the newly created Computational Geometry Package.

CentralMoment, GeometricMean, HarmonicMean, InterquartileRange, Kurtosis, MeanDeviation, MedianDeviation, QuartileDeviation, Quartiles, QuartileSkewness, RootMeanSquare, Skewness, and TrimmedMean are now part of the built-in Mathematica kernel:

Version 5.2 << Statistics`MultiDescriptiveStatistics`;
GeometricMean[{{10, 50}, {10, 15}, {20, 25}, {10, 15}}]

MultivariateMode is replaced by Commonest:

Version 5.2 << Statistics`MultiDescriptiveStatistics`;
Mode[{{10, 50}, {10, 15}, {20, 25}, {10, 15}}]

SampleRange can be computed using Min and Max:

Version 5.2 << Statistics`MultiDescriptiveStatistics`;
SampleRange[{{10, 50}, {10, 15}, {20, 25}, {10, 15}}]

ZeroMean can be computed by subtracting the mean from the data:

Version 5.2 << Statistics`MultiDescriptiveStatistics`;
ZeroMean[{{10, 50}, {10, 15}, {20, 25}, {10, 15}}]

Other univariate measures can be computed as shown in the Statistics`DescriptiveStatistics` compatibility pages:

Version 5.2 << Statistics`MultiDescriptiveStatistics`;
KurtosisExcess[{{10, 50}, {10, 15}, {20, 25}, {10, 15}}]

Covariance matrix for data:

Version 5.2 << Statistics`MultiDescriptiveStatistics`;
CovarianceMatrix[{{10, 50}, {10, 15}, {20, 25}, {10, 15}}]

MultivariateKurtosisExcess can be computed from MultivariateKurtosis:

Version 5.2 << Statistics`MultiDescriptiveStatistics`;
MultivariateKurtosisExcess[{{10, 50}, {10, 15}, {20, 25}, {10, 15}}]

Alternate covariances can be computed from Covariance:

Version 5.2 << Statistics`MultiDescriptiveStatistics`;
CovarianceMLE[{10, 50, 10, 15, 20}, {20, 35, 5, 25, 40}]
Version 5.2 << Statistics`MultiDescriptiveStatistics`;
CovarianceMatrixMLE[{{10, 50}, {10, 15}, {20, 25}, {10, 15}}]

AssociationMatrix, CovarianceMatrixOfSampleMean, DispersionMatrix, MultivariatePearsonSkewness1, and MultivariatePearsonSkewness2 were available in an add-on package in previous versions of Mathematica and are now available on the web at library.wolfram.com/infocenter/MathSource/6831.