ABSTRACT
There are ways of adjusting signicance levels to allow for many tests being carried out at the same time, but it may be preferable to conduct a single test that uses the information from all variables together. For example, it might be desirable to test the hypothesis that the means of all variables are the same for two multivariate populations, with a signicant result being taken as evidence that the means differ for at least one variable. These types of overall test are considered in this chapter for the comparison of means and the comparison of variation for two or more samples.
