Week 7 Flashcards
What happens when we have a third IV?
The same principles for analysing a two factor ANOVA apply. We just have more possibilities for interactions.
A 3 way interaction, 3 two way interactions and 3 one way interactions
What is the error term for a 3 way interaction in a between subjects design?
abc(n-1)
Where a, b and c are the dof for each of those IV and n is the number of participants in each group.
If it is a between subjects design, we need __ column for each variable and one for the _____
one
dependant variable
What does APA specify in terms of standard errors?
You have to be using standards deviation or confidence intervals. You can’t just use standard error.
How can we get standard deviations?
IN DESCRIPTIVES
If a main effect isn’t involved in anymore complex interactions, what do we do?G
Give a more detailed analysis, can do tests of post hoc.
What is a significant interaction actually telling us?
That the overall effect for IVA isn’t consistent (the same) across all levels of IVB.
A two factors interaction in the absence of a three factor interaction is assessed through an analysis of:
simple main effects
A significant three factors interaction between IVs A, B & C can be thought of as asking “is the interaction between IV A and IV B the same at all levels of IV C? How do we test this?
Examine the simple interaction effect of IV A X IV B separately for each of the levels of IV C
What are two things we think about to go about how we conduct analysis of a significant three factor interaction (i.e., do we test IVA and IV B against IVC, or IVC and IVA against IVB)?
- theoretical decisions as to which are the most relevant (plots may help to make this decision)
- if one factor only has two levels, this can help to simplify
If we have a significant 3 way interaction, and we have split this up to identify a two way interaction on the basis of theory, what is the next step?
To perform follow up tests to clarify this interaction. (simple effects, and do Bonferroni).
If BOTH of the interactions you have choosen to split it by have significance, you have to do simple main effects on both! If there are 6 tests, you have to divide the alpha by 6 for this family of comparisons.
What should go in the simple effects box in Jamovi?
The thing you are testing against. The thing you have theoretically justified as possibly producing the interaction.
One way to help understand a 3 way interaction is to look at what?
a plot of the AXB interaction at all levels of C.
The more follow up tests of simple effects you do, the more chance you have of making a:
type 1 error
