8 - Multiple Regression and Interactions

Question 1

How do you express interaction in multiple regression?

Answer 1

By multiplying variables (*) instead of adding them.

Question 2

What does it mean if 2 variables interact?

Answer 2

It means that one influences the effect of the other. 2 variables interact when one influences the effect of the other.

Question 3

True or False: Interaction is a form of moderation effect. Note: moderation effect = when a third variable changes the nature of the relationship between a predictor and the outcome.

Answer 3

True.

Question 4

Sometimes, multicollinearity issues arise when working with interaction terms. What can be done to avoid said issues?

Answer 4

Centering.

Question 5

What 3 strategies can be used to further understand interactions?

Answer 5

1. Decompose interaction into lower-order components
2. Probe (use hypothesis testing to evaluate significance of interactions)
3. Plot/ visualize the interaction as simple slopes

Question 6

FIGURE OUT SIMPLE SLOPES VS SIMPLE INTERACTION TERMS

Answer 6

FIGURE OUT PLOTTING SIMPLE SLOPES

Question 7

What R functions can be used to plot simple slopes?

Answer 7

• plotSlopes() from rockchalk package
• interact_plot() from interactions package