Interactions (Moderations)

Question 1

What is an interaction?

Answer 1

The effect of one variable depends on the level of another.

Question 2

Thus far, we assumed additivity, what does this mean?

Answer 2

Adding each predictor to interpret main effects. But now we add interaction term.

Question 3

What type of vector do we need to include the interaction term?

Answer 3

A cross-product vector.

Question 4

How do we do a cross product of 2 continuous variables?

Answer 4

Multiplying 2 variables together.
X1X2 = Cross product

ex) person 1 scores a 3 in variable 1 and 4 in variable 2 (interaction of continuous variables).

3x4 = 12

Question 5

How many vectors do we need to represent a 2 category variable?

Answer 5

I only need 1 vector to represent a 2 category variable.

Question 6

I have 2 vectors to code for 1 variable, how many categories do I have?

Answer 6

3 categories.

ex) 5 categories - 4 vector codes
2 vector codes - 3 categories

Question 7

• How do you represent interactions (two-way and higher-order) in a regression analysis? What model do you fit?

Answer 7

o Two Continuous variables - To include the interaction term to regression analysis, we need to multiple the 2 continuous variables to make a Cross-product (X1X2).
• So if person scored a 4 on X1 (variable 1) and 3 on X2 (variable 2), then interaction term is 12.
• Note: Continuous is notated as X
o Categorical vs. Continuous –
• Note: Notation of Categorical is Dummy Coded (0 and 1s). Categorical is notated as D