R Statistics Cookbook

How it works...

There are two main functions here: nonpartest() and ssnonpartest(). The former provides global statistics and p-values as in MANOVA. In our case, we clearly rejected the null hypothesis of equality of means between the different groups. The latter function (ssnonpartest), identifies significant subsets of variables and later pinpoints which is the variable driving the differences between the groups. In our example here, the initial difference that we got using nonpartest() is caused by the fact that the setosa-virginica and setosa-versicolor vectors are different. We then found out that all the variables (Petal.Width, Petal.Length, Sepal.Width, and Sepal.Width) are driving that difference that we got using nonpartest().