Mediation Analysis

Question 1

What types of effects are seen in mediation?

Answer 1

Direct effect and indirect effect

Question 2

What variables are there is mediation?

Answer 2

Predictor, mediator and outcome.

Question 3

What are the different variables called in PROCESS?

Answer 3

Y = outcome 
X = predictor 
M = mediator

Question 4

What is percentile bootstrap confidence interval levels?

Answer 4

It’s the upper limit confidence interval (ULCI) and lower limit confidence interval (LLCI) which states that there is 95% confidence that unstandardized coefficient for the indirect effect of the predictor on the outcome will have a value included in this confidence interval.

Question 5

Why is bootstrap confidence interval different for every time you run the test?

Answer 5

Because it’s based on taking random samples from the data set.

Question 6

What is the common number of bootstrap samples?

Answer 6

1000, 2000 or 5000.

Question 7

How many outputs does PROCESS give for mediation analysis?

Answer 7

4 outputs.

Question 8

What is output 1 in mediation?

Answer 8

It looks at the effects of the predictor on the mediator. Simple regression

Question 9

What is output 2 in mediation analysis?

Answer 9

Multiple regression looking at the effects of both the predictor and the mediator on the outcome variable

Question 10

What is output 3 in mediation analysis?

Answer 10

It’s the total effect model which shows simple regression of the effects of the predictor on the outcome (without the mediator)

Question 11

What is the 4th output in mediation analysis?

Answer 11

Total, direct and indirect effects.

Question 12

What is BCa?

Answer 12

It stands for bias corrected accelerated and it refers to the bootstrap CI (confidence interval) method used

Question 13

What does some researchers believe about mediation?

Answer 13

That it should only be used on cross sectional data because mediation implies causal process unfolding over time