Lecture 13: Factorial ANOVA

Question 1

What is an independent factorial ANOVA

Answer 1

Two or more independent variable with two or more categories, one dependent variable

Question 2

What are the assumptions for a factorial ANOVA

Answer 2

1. Continuous variable
2. Random sample
3. Normally distributed
4. Equal variance within groups

Question 3

T/F: It becomes easier to asses the individual effect of one of the predictor variables if there is an interaction effect

Answer 3

False! It becomes harder

Question 4

T/F: an ANOVA looks at the variance in the dependent variable and tries to explain it by adding more predictor variables

Answer 4

True

Question 5

When looking at the interaction between multiple IV’s, the model error (= error sum of squares) … (decreases/increases), and that leads to the … (decrease/increase) of the model accuracy

Answer 5

Decreases, increase

Question 6

What is the term for when the model error and model accuracy are added together

Answer 6

Total sum of squares

Question 7

What are post-hoc comparisons and how do they differ from contrasts

Answer 7

Unplanned comparisons that explore all possible differences and adjust the T-value for inflated type 1 error (correct for the p-value which controls for the type 1 error)

Question 8

Does it make sense to include a variable with two levels in your post hoc analysis and why/ why not

Answer 8

It does not, because you already know that the difference is by definition between those two levels. When you have 3 or more levels, you can’t be sure which levels exactly differ from each other, and to find this out you can look at the post hoc tests