3. Two-way ANOVA + contrasts Flashcards
What is the formula for total number of groups in two-way ANOVA?
J x K
What is a factorial design?
A design of two or more IVs, where every level of each IV is combined with every level of every other IV.
How many cells are there in a 3x5 design?
15
If there are n subjects per group, how many subjects are there in total (N)?
N = n x J x K
How do you calculate SS for a main effect in two-way ANOVA?
For a main effect, instead of multiplying sum of squared deviation scores by n, you multiply by n AND the number of groups for the OTHER IV. So multiply by nJ or nK.
How do you calculate SS for an interaction in two-way ANOVA?
Just multiply by n as usual.
What are degrees of freedom for an interaction in two-way ANOVA?
(J-1)(K-1) or dfA x dfB
What are the degrees of freedom error for a two-way ANOVA?
N–JK
What should contrast coefficients always add to?
The same integer for positive and negative coefficients. I.e. 2 and -2, 1 and -1.
What is the null hypothesis for contrasts?
That the weighted means are all equal to each other.
How is type I error rate controlled in two-way ANOVA?
By families. We control the family-wise error rate.
How many families are there in two-way ANOVA?
Three.
IVA
IVB
IVA x IVB
What are simple main effects?
The effect of one IV on the DV at one level of the other IV.
What control of EER should be used for simple main effects?
Scheffé - as simple main effect contrasts are not planned (if they were, why use a factorial design).
Are there families of simple main effects?
No. So just use EER control – i.e. Scheffe –for them.
