# ANOVA

HideShow resource information
• Created by: jmichel
• Created on: 16-05-16 11:06
When to use ANOVA?
3 or more DVs or more DVs
1 of 26
why not using multiple t-test?
because of the familywise error
2 of 26
F ratio = ?
MSbetween / MSwithin
3 of 26
Between groups
deviation of groups means from the grand mean
4 of 26
within groups
deviation of individual scores from their respective group mean
5 of 26
total variance
deviation if individual scores from the grand mean
6 of 26
What to report with ANOVA
Sum of squares, degrees of freedom, mean squares, F-ratio, p-value, effect size
7 of 26
Repeated-measures ANOVA
equivalent of the one-way ANOVA, but for related, not independent groups
8 of 26
two-way anova
two or more IVs
9 of 26
What is the additional assumption for repeated measure ANOVA?
Sphericity (only if 3 or more conditions)
10 of 26
How do we test for sphericity
Maulchy's test (if < 0.05 assumption violated)
11 of 26
Where do we look when sphericity is violated?
Greenhouse-Geisser
12 of 26
When to look as post hoc tests?
When comparisons are sig.
13 of 26
Two-way ANOVA
compares several means for 2 independent variables
14 of 26
What are the effects you get of two-way ANOVA?
main effect of each variable; interaction between the two
15 of 26
interaction effect
combines effect of two or more predictor variables
16 of 26
main effect
the unique effect of a predictor variable
17 of 26
What to report with 2x2 ANOVA
all the main effects and the interaction
18 of 26
What to check in terms of normality and outliers
skewness, Kurtosis, Histograms, Boxplots, z-values
19 of 26
multivariate normality
normal distribution to multiple variables
20 of 26
multvariate outliers
combination of unusual scores on at least two variables
21 of 26
homogenste of variance
variance of one variable is stable at all levels of another variable
22 of 26
small effect size
.01
23 of 26
medium effect size
.06
24 of 26
large effect size
.14
25 of 26
Assumptions of all ANOVAS
normality, no outliers, multivariate normality. no multivariate outliers, homogeneity of variance, sphericity
26 of 26

## Other cards in this set

### Card 2

#### Front

why not using multiple t-test?

#### Back

because of the familywise error

F ratio = ?

Between groups

within groups