This preview shows page 1. Sign up to view the full content.
Unformatted text preview: . In
the presence of a significant effect, two of the possible approaches include (i) dependent variable contribution and (ii)
multivariate contrasts. In the dependent variable contribution approach, dependent variables are removed one at a time and
MANOVA is performed with the remaining dependent variables
to examine how much change is induced in the significance of an
effect by the removal of a dependent variable (18). Besides dependent variable contribution, multivariate contrasts may be constructed to assess differences among significant independent
Several criteria used for gauging multivariate differences among
independent variables are Roy’s gcr, Wilks’ lambda (or U statistic), Hotelling’s trace, and Pillai’s criteria. Although Wilks’
lambda is most often used and reported in research articles, the
choice of any criterion is guided by the degree at which the
assumptions underlying MANOVA are satisfactorily met (7,18).
Correspondence analysis. Correspondence analysis describes
relationships among two or more cross-tabulated categorical variables (con...
View Full Document
This document was uploaded on 11/20/2013.
- Summer '11
- The American