A branch of statistics that deals with more than two variables at once. This is important, for example, when comparing series of skull measurements, the results of chemical analysis for a number of elements, or a large number of characteristics recorded about a group of artefacts. In multivariate statistics, the skulls, objects, or whatever, are distributed in a hypothetical space, called hyperspace, or Euclidean space, which has a number of dimensions equivalent to the number of variables being studied. Some multivariate techniques (e.g. cluster analysis and discriminant analysis) analyse the distribution of the items under study within the hyperspace, reporting their results as a table or plot. Other techniques (e.g. principal multidimensional scaling) mathematically reduce the number of dimensions of the space. Typically, a multidimensional distribution may be reduced to two or three dimensions, after which it may be plotted or analysed by conventional statistics.