Factorial ANOVA

A

ANOVA

B

factorial ANOVA

C

MLR with dummies

D

MANOVA

E

ANCOVA

F

RM ANOVA

DV and IV measurement level in ANOVA

DV interval/ratio IV nominal

3X2

A 3lvl, B 2lvl

Does watching the pictures have the same effect on men as it has on women (condition, gender) explain effects

1. main effect condition: are self ratings the same in all three conditions?

2. main effect gender: are the self ratings the same for men and woman?

3. interaction effect: are the differences between men and women the same in all conditions?

Factorial ANOVA steps

1. assumptions

2. check main and interaction effects (F-test)

3. if significant make an interaction plot, if not rerun without interaction

4. effect size of significant effects

5. post-hoc comparison for factors with more than 2 lvls

ANOVA assumptions

1. normality of DV: Q-Q

2. homogeneity of variances: Levene a=0.01, NS=H

How to test which groups are significantly different?

1. post hoc

2. simple main effects

ANOVA simple main effects

testing for differences between groups on one factor

F test

1. F=explained variance/unexplained variance = MSm/MSr

2. F=between group variance/within group variance = MSm/MSr

ANOVA measure of effect size

1. eta squared = SSm/SSm+SSr

2. .01/.09/.25

Factorial ANOVA measure of effect size

partial eta squared = SSeffect/SSeffect+SSresidual

Adjusted vs Unadjusted means

1. unadjusted ANOVA mean is larger

2. adjusted ANCOVA mean is smaller

Bayes factor extreme evidence (BF10)

1. H0: BF<0.01

2. H1: BF>100

Bayes factor very strong evidence (BF10)

1. H0: 0.01-0.03

2. H1: 30-100

Bayes factor strong evidence (BF10)

1. H0: 0.03-0.1

2. H1: 10-30

Bayes factor moderate evidence (BF10)

1. H0: 0.1-0.3

2. H1: 3-10

Bayes factor anecdotal/weak evidence (BF10)

1. H0: 0.3-1

2. H1: 1-3