Categorical predictors and moderation (L8)

Question 1

What are dummy variables?

Answer 1

A way of recording categorical variables with 2 + categories into a series of dichotomous variables (eg. 0 or 1).

Question 2

How do you make dummy variables?

Answer 2

1) Count groups you want to recode - 1
2) Create as many new variables as calculated in step 1
3) Choose one group as a baseline (the group all others with be compared to)
4) Assign that baseline to 0
5) 1st dummy variable = 1, all others 0
6) 2nd dummy variable = 1, all others 0
7) Repeat until all variables have been tested.

Question 3

What does the linear model for dummy variables look like?

Answer 3

One line connecting all the means. Intercept = mean of 0 coded group; ‘b’/gradient = difference between means of categories.

Question 4

What is moderation?

Answer 4

The relationship between variables changes as a function of a third variable.

Question 5

What is a moderator variable?

Answer 5

Variable that changes the size and/or direction of the relationship between 2 other variables.

Question 6

What is interaction?

Answer 6

Predictor x moderator

If significant as a predictor you have a significant moderation effect.

Question 7

What are simple slope analysis?

Answer 7

Graph showing the relationship at varying levels of moderator.