week 11: mediation

Question 1

What is mediation?

Answer 1

refers to a situation when the relationship between a predictor variable and an outcome variable can be explained by their relationship to a third variable (the mediator)

Question 2

Explain the interaction between variables in mediation

Answer 2

3 way interaction (think of a venn diagram). **the relationship between the mediator, the predictor and the outcome explains some of the relationship between the predictor and the outcome

Question 3

what letter symbol represents the total effect (predictor, outcome, and mediator)?

Answer 3

c

• total effect: simple regression Not controlling for mediator

Question 4

what letter symbol represents the 1st part of the indirect effect (predictor -> mediator)?

Answer 4

a

• simple regression

Question 5

What letter symbol representsthe 2nd part of the indirect effect (mediator -> outcome)?

Answer 5

b

• multiple regression controlling for predictor

Question 6

What letter symbol represents the direct effect (predictor -> outcome)?

Answer 6

c’

• multiple regression controlling for mediator

Question 7

Baron & Kenny (1986) suggested that mediation is tested through what 3 linear models?

Answer 7

1. a linear model predicting the outcome from the predictor (total: c)
2. a linear model predicting the mediator from the predictor (1st indirect: a)
3. a linear model predicting the outcome from both the predictor (direct: c’) and the mediator (2nd indirect: b)

Question 8

What are the 4 conditions of mediation?

Answer 8

1. the predictor must significantly predict the outcome (model 1: is c significant?)
2. the predictor must significantly predict the mediator (model 2: is a significant?)
3. the mediator must significantly predict the outcome, controlling for the predictor (model 3: is b significant?)
4. the predictor must predict the outcome less strongly in model 3 than in model 1

Question 9

what is the equation for the total effect?

Answer 9

total = direct + indirect

c = c’ + ab

Question 10

What is the equation for the direct effect?

Answer 10

c’ = c - ab

Question 11

which is larger, the total or direct effect?

Answer 11

total is larger

Question 12

what is full mediation?

Answer 12

direct effect of the predictor on the outcome becomes nonsignificant when the mediator is entered into the analysis

Question 13

What is partial mediation?

Answer 13

Direct effect of the predictor on the outcome is reduced but still significant when the mediator is entered into the analysis

Question 14

What is bootstrapping?

Answer 14

• a statistical procedure that resamples a single data set to create many simulated samples
• useful when the data is highly skewed with low N
• the higher your N, the more similar it is to normal stats (remember central limit theorem)

Question 15

How does bootstrap resampling work?

Answer 15

resample from your sample with replacement thousands of times (default = 5000 random resamples)
- order the estimates and work out the limits within which 95% of them fall
- with replacement, put each score back before a new one is drawn from the sample

Question 16

Why use bootstrapping?

Answer 16

• it can give you more reliable estimates of a standard error (more reliable confidence intervals)
• bootstrapping is often used for the indirect path (ab) in mediation

Question 17

How do you interpret confidence intervals?

Answer 17

95% of the time a replication of the study should give you a result within the confidence interval (CI)
- Lower limit (LLCI) and Upper limit (ULCI) have equal distances from the ab coefficient
- if both LL and UL are positive or both negative, the ab path is significant (different from zero)
- if one is + and one is -, then ab path is not significant

Question 18

how do you report confidence levels and significance?

Answer 18

Ba*b = 0.005, 95% CI [0.002,0.08]
(EXAMPLE)