Unformatted text preview: stence of such correlation
may reduce the power of MANOVA (7,19) in detecting differences among independent variables. Remedial measures include
several options such as using repeated measures analysis and principal component analysis. In the repeated measures MANOVA
(19), a set of new variables (D1, D2, and D3) are derived from the
mean differences between successive DS values as follows: D1 =
DS2 – DS1, D2 = DS3 – DS2, and D3 = DS4 – DS3. These new
variables are then used as dependent variables.
Principal component analysis enables the identification of linear
combinations of dependent variables with maximum variance,
which are uncorrelated to each other. These components are then
used as dependent variables in lieu of the correlated original
variables. A drawback of using principal component as a remedial
procedure may be the difficulties in interpreting the results.
Two other matters related to the use of MANOVA include the
handling of a significant effect and the choice of a test statistic...
View Full Document
- Summer '11
- The American, Multivariate statistics, MANOVA