ABSTRACT
Summary values of multivariate data are often referred to as profiles and contain averages, variation measures, correlations and other summaries. The core of modern data science deals with a mathematical modeling of multivariate patterns in the data. Very often in medical context, it is necessary to compare mean profiles across several groups of cases to rule out chance occurrence of findings.
This chapter explains methods of comparing the mean values of a vector (panel) of variables among several groups instead of comparing each variable separately. The need and importance of comparing multivariate means using the Hoteling’s T2-test discussed as a starting point. Illustrative notes are provided for two important situations viz., one sample comparison (with a hypothetical vector) and the two-sample test between independent groups. The use of Real-statistics add-in of Excel is demonstrated with a step wise tutorial to perform Hoteling’s test using live data. The chapter also unfolds the importance of confidence interval to verify the truth of a hypothesis apart from the p-value of a test. (169)
