Week 9 - Mediation/moderation

Question 1

What are the 3 uses of HMR?

Answer 1

Normal - control variables at step 1, interesting ones at step 2
MMR - centred IVs in step 1, interaction term in step 2
Mediation -

Question 2

• Explain the difference between mediation and moderation analyses with respect to:
o The research question that is addressed
 (x2)

Answer 2

Moderation is WHETHER different levels of a second variable changes the relationship of the focal to the DV/criterion
Mediation is HOW (the ‘because’…) - the effect goes through it

Question 3

• Explain the difference between mediation and moderation analyses with respect to:
o The relationship between the main predictor and the second predictor (i.e., the mediator or moderator)
(x4)

Answer 3

Moderation has the t-bar
Mediation, the triangle
In moderation, X and Z may or may not correlate
In mediation, they must

Question 4

• Explain what type of research question can be addressed in mediation analysis 


Answer 4

Does one variable explain how another affects the DV?

Question 5

• List the four conceptual steps in mediation analysis 


Answer 5

1. IV should predict mediator
   
   2. IV should predict DV in Block 1
   3. Mediator should predict DV in Block 2 (i.e., when IV = controlled)
   4. Coefficient for IV should decrease to ns (full mediation) or in size but still sig (partial mediation)

Question 6

• List the three steps in testing for mediation, and for each step explain: 

o What analysis you would use
And why you’d report
(x3, x3, x3)

Answer 6

Show path A - IV and mediator related:

• SMR of mediator on IV
• Significant R2 and coefficient (b or beta)

Show path B - IV/DV:

• HMR step 1, predict DV from IV
• Significant R2 and coefficient

Show path C - mediator/DV, controlling for IV:

• Predict DV from IV + mediator
• R2ch, coefficients for IV and mediator

Question 7

• Explain the difference between partial mediation and full mediation. (x1, x1)

Answer 7

In partial, IV retains some independent relationship with the DV
In full, the mediator is essential to link IV/DV together

Question 8

How would you know if your results provide evidence for partial or full mediation? (x2)

Answer 8

If IV/DV path no longer significant at step 2, full mediation
If still significant, but reduced, partial

Question 9

Define ‘direct and indirect relationships’ (x1, x1)

Answer 9

Direct - one predictor is associated uniformly with the criterion
Indirect - there’s a causal chain, a second predictor that provides the pathway to the DV

Question 10

In HMR, what is the effect of giving shared variance to the previous predictor? (x1)

Answer 10

It increases the beta for that predictor (compared to adding all in SMR)

Question 11

Do you always need an IV/DV relationship in mediation? (x2)

Answer 11

Some say no - can just go IV - mediator - DV

But controversial

Question 12

What is the Sobel test? (x4)

Answer 12

After mediation analysis developed, Realised small changes in significance in IV/DV relationship may not be reliable (that dictates full/partial mediation)
So sobel is final test of significance of indirect effect
*Direct comparison of IV before and after controlling mediator

Question 13

With Sobel test going out of fashion, what is common method of testing mediation (that requis no other analyses)(x1)

Answer 13

Bootstrapping