ABSTRACT
In most experimental studies, several different assessments and measurements may be made on each sample studied. These measurements may be of independent characteristics of the samples, but more likely they are different ways of measuring underlying characteristics of the material under investigation and as such are likely to be correlated. For example, the height and girth of trees are both measures of size and are likely to be positively correlated. Multivariate analysis is a term used to cover a range of statistical techniques for describing, simplifying, and analyzing data sets where many different variables are measured on a set of samples or objects. Most multivariate techniques are descriptive and graphical; few multivariate methods have been developed for fitting models or making precise inferences.
