Page 3 - Moderation and Mediation Theory and Concepts

Question 1

What are conditional effects in moderation

Answer 1

The relationship between a predictor and the outcome variable is dependent on the other predictor (moderating predictor (b3XMi) having a specific value (0)

Question 2

When should we use centering

Answer 2

When a predictor value of 0 is not possible (normally on b2 [moderator]

Question 3

What is grand mean centring

Answer 3

For each variable, subtracting the mean from each score

Question 4

Does centering assist the interpretation of ‘lower order’ (conditional) effects?

Answer 4

Yes, it alters the b value.

Particulary when there is no moderation effect

Question 5

Does centering assist the interpretation of ‘higher order’ (interaction) effects?

Answer 5

Yes, it does not alter the b value. Particulary when there is a moderation effect

Question 6

How are centring variables and multicollinearity related?

Answer 6

Reduces the strength of the correlation between the predictor(s) and interaction term

Question 7

How do you perform centring in SPSS

Answer 7

PROCCESS function

*Identify the grand mean from descriptives

Question 8

The combined effect of two or more predictor variables on an outcome is know as…

Answer 8

Moderation (conceptually)
Interaction effection (Statistical term

Question 9

What does the Johnson-Neyman Technique do?

Answer 9

Same analysis as simple slopes, but with more detail rather than just three levels of the moderator (1SD, 0, -1SD)
*Produces zones of significance instead

Question 10

What is the Zone of significance

Answer 10

Region in which values of the moderator results in a significant relationship between the predicitor and the outcome

Question 11

What is the Zone of significance

Answer 11

Region in which values of the moderator results in a significant relationship between the predicitor and the outcome

Question 12

A mediator is a variable that helps explain…

Answer 12

WHY or HOW a predictor variable relaties to an outcome variable

Question 13

What are the key objectives od mediation

Answer 13

Quantify what proportion of the relationship between the predictor and outcome is;
A. Direct from the predictor
B. What proportion flows indirectly to the outcome via the mediator

Question 14

What are the three effect pathways of mediation?

Answer 14

Total effect (c) 
Indirect effect (a.b)
Direct effect (c' or c prime) =

Question 15

What is the Total effect (c)

Answer 15

The effect of the predictor on the outcome variable

Question 16

What is the Indirect effect (ab)

Answer 16

The effect of the predictor opn the outcome variable, through the mediator

Question 17

Direct effect (c’ or c prime)

Answer 17

The effect of the predictor on the outcome variable, controlling for the mediator

Question 18

What does path ‘a’ represent

Answer 18

The effect of the predictor on the mediator

Question 19

What does path ‘b’ represent

Answer 19

The effect of the mediator on the outcome, controlling for the predictor

Question 20

Baron and Kenny (1986) steps to mediation

Answer 20

1. Determine predictor significantly predicts the outcome
2. Determine that the predictor significantly predicts the mediator
3. Determin the presence of the mediator in a regression model with thepredictor and outcome reduces the predictive strength of the predictor in the model

Question 21

What is the Sobel test

Answer 21

*Added to Baron and Kenny (1986) to test the significance of the indirect effect not provided view Baron and Kenny

Question 22

When can we state that the mediator has a significant effect of the relationship between the IV and DV in SPSS?

Answer 22

When the Confidence interval for the indirect effect (A x B correlations) does not cross 0
Examples
Significant = [0.09, 0.47]
Non=significant = [-.07, .047]

Question 23

When can we state that the mediator has a significant effect of the relationship between the IV and DV in theory

Answer 23

If the proportion of variance attributable to the indirect effect (A.B) via the proposed mediator is significant we say we have statistical support for the mediator

Question 24

Full mediation

Answer 24

If the proportion of variance attributable to the direct effect of the predictor is not significant we say we have support for full mediation

Question 25

Partial mediation

Answer 25

If the proportion of variance attributable to the direct effect of the predictor is significant we say we have support for partial mediation

Question 26

Is the Baron and Kenny (1986) method flawed?

Answer 26

Yes, Sobel test of indirect effect is more appropriate

Is causal, when significant mediators do not ensure causation - need solid theoretical base, RCT, longitudinal

Question 27

What does the Sobel test inappropriately assume?

Answer 27

Assumes normal distribtution for the indirect effect

*unstable results for small samples

Question 28

What did Preacher and Hayes develop?

Answer 28

Boostrapped test for the indirect effect which does not assume normality, thus avoid Sobel test issue
*Used through PROCESS macros in SPSS

Question 29

A moderator is a variable that helps explain/answer…

Answer 29

for WHOM and WHEN a relationship occurs between the predictor and the outcome variable

Question 30

What is moderation?

Answer 30

A technique used to explore whether the relationship between variable changes as a function of a third variable (moderator) [interaction effect

Question 31

What does centring involve?

Answer 31

The transformation of scores around a designated point that makes interpretation meaningful