Multiple Criterion Measures

Univariate ANOVA
Repeated Measures ANOVA
MANOVA

N subjects
Economy of subjects (N = # of observations)
s in smallest cell > # of DVs

At least 2 subjects per cell
Imprecision is not with N but with non-correlation of DVs
Sensitive to outliers for small N

Type I error rate can be very large
Type I error rate > alpha but not too bad
Type I error rate of alpha

Easiest: independence, normality, HOV
Hardest: same 3 + sphericity (within subjects)
and compound symmetry (between subjects)
In Between: same 3 + multicollinearity of DVs

Least powerful
Most powerful
More power than univariate ANOVAs

Gives interaction terms
Gives interaction terms
Does not give interaction terms

Tests all DVs separately
Sums DVs together; useless if they're not additive
Takes multiple measures as a coherent package

Back to Statistics course

Lorraine Sherry
http://www.cudenver.edu/~lsherry/multiple_criteria.html
Updated April 1, 1997

Univariate ANOVA	Repeated Measures ANOVA	MANOVA
N subjects	Economy of subjects (N = # of observations)	s in smallest cell > # of DVs
At least 2 subjects per cell	Imprecision is not with N but with non-correlation of DVs	Sensitive to outliers for small N
Type I error rate can be very large	Type I error rate > alpha but not too bad	Type I error rate of alpha
Easiest: independence, normality, HOV	Hardest: same 3 + sphericity (within subjects) and compound symmetry (between subjects)	In Between: same 3 + multicollinearity of DVs
Least powerful	Most powerful	More power than univariate ANOVAs
Gives interaction terms	Gives interaction terms	Does not give interaction terms
Tests all DVs separately	Sums DVs together; useless if they're not additive	Takes multiple measures as a coherent package