The article discusses tests comparing one or two sets of average values, using a method called Hotelling's T2 distribution. This method is more powerful than testing each value separately because it considers correlations between variables. When results differ, the multivariate test should be used. The article provides a table of critical values for Hotelling's T2 distribution and discusses its properties. It also explains how to compute T2 using different statistical methods. The article includes examples with real data and problems for practice.