Manova and Mancova PDF

Preview

Summary

Notes about MANOVA and MANCOVA and how to conduct the analysis using SPSS...

Description

MANOVA This is used when the study measures 2 things that are theoretically similar. (2 DVs) Large samples, small effect. MANOVA can be used in those circumstances Small effect does not result in significant results, but when you add all small effects across the study, the result end up being statistically significant. This is why MANOVA can be used. For individual analysis a different test would be used. MANOVA can be complicated, mixed designs etc.

T-test is equivalent to Hotelling’s T if you have multiple DVs

Assumptions (refer to BB PowerPoint) Check for similar group sizes. Variance is a calculation of deviation from the mean.

(pp slide – 2. Homogeneity of covariance-covariance assumption) Variance-covariance***

Use Wilks Lambda and Hotelling’s T.

Box-m value needs to be reported in the assignment. In the MANOVA table (multivariate tests) ignore the intercept, look at group.

Use Shapiro –wilk if the sample size is small. One-way anova. Only look at the group sig. table.

Repeated measures MANOVA: In a way this is easier because you only have one group with multiple DVs. Should still check for normality for all the variables. Can look at the effect of treatment – before and after.

MANCOVA: At least 2 dvs. Approx. normal distribution. Reasonable correlation between the dvs. Reasonable correlation between the covariate and the dvs. (Check PowerPoint) 1.000 sig = not significant....