TABLE OF CONTENTS The multivariate statistical inferences discussed in Chap. 3 can only conduct comparative study for up to two multivariate populations. In Example 3.7, we compared the impact-absorbing qualities for two sides of the bicycle protection helmet. If we are given two similar welding technician training programs (not one) as we discussed in Example 3.1, and both programs offer the same three courses, we can certainly use the multivariate statistical inferences techniques that are discussed in Chap. 3 to compare the performances of these two programs. However, what if we are asked to compare more than two, say, five similar welding technician training programs in a state?
This situation is called multiple comparisons, that is, comparison of three or more similar items. When each item has only one characteristic to compare, this multiple comparisons problem is a univariate problem, and it can be dealt with by analysis of variance (ANOVA). When each item has more than one characteristic for example, in the welding technician training program problem, there are three courses, the grade of each course is a characteristic, and so we have three characteristics for each item this kind of multiple characteristics and multiple comparisons problem will have to be dealt with by multivariate analysis of variance (MANOVA).
As seen in Chap. 3, there are usually similarities between each multivariate statistical method and its univariate counterpart, and comparing the similarities and differences between the univariate method and the corresponding multivariate method can greatly enhance our understanding of the...
