Scribd
Upload
49 views32 pages

PSY40630 Advanced Topics

This document provides an overview and summary of topics related to mediation and moderation analysis, as presented in a session on quantitative data analysis taught by Dr. Laura Taylor. Key points include: - Mediation involves how an independent variable (X) affects a dependent variable (Y) through a mediator, while moderation looks at when or under what conditions X affects Y. - Analytical approaches for testing mediation and moderation hypotheses include regression analysis in SPSS and the PROCESS macro add-on. - Advanced topics discussed include multiple mediators, moderated mediation, and mediated moderation. Diagrams are provided to illustrate concepts like parallel mediation, chain mediation, and moderated

Uploaded by

Marge Sierra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
49 views32 pages

PSY40630 Advanced Topics

This document provides an overview and summary of topics related to mediation and moderation analysis, as presented in a session on quantitative data analysis taught by Dr. Laura Taylor. Key points include: - Mediation involves how an independent variable (X) affects a dependent variable (Y) through a mediator, while moderation looks at when or under what conditions X affects Y. - Analytical approaches for testing mediation and moderation hypotheses include regression analysis in SPSS and the PROCESS macro add-on. - Advanced topics discussed include multiple mediators, moderated mediation, and mediated moderation. Diagrams are provided to illustrate concepts like parallel mediation, chain mediation, and moderated

Uploaded by

Marge Sierra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 32

@lauraktaylorPHD

#PSY40630
#PSY40400

Mediation & Moderation


Advanced Topics

PSY40630: Quantitative Data Analysis

Dr Laura Taylor

Session 3
Overview
1. Understanding the differences between mediation and
moderation

2. Analytical approaches to testing mediation and moderation


hypotheses
– Based on regression analysis in SPSS
– PROCESS add-on

3. Advanced topics
– Multiple mediators
– Moderated mediation
– Mediated moderation
Mediation vs. Moderation
Mediation Moderation
• How does X affect Y? • When does X affect Y?
• What is the psychological • Does the relationship between X
mechanism through which X and Y change at different levels of
affects Y? M?

M
M

X Y X Y

• Significant interaction term (X*M)


• Significant indirect effect (a*b) • Decomposition of the effect of X
• Small direct effect (c’ < c) on Y at different levels of M (e.g.,
simple slope analysis)
ADVANCED TOPICS
Parallel Mediation
• Multiple mediators
– Often the case that there would be multiple
mediators of the relationship between X and Y
– Using PROCESS add-on in SPSS, we can then
determine the size and significance of each indirect
effect
Security

Life
Income satisfaction

Freedom
Chain Mediation
• Chain mediation:
– allows us to model a more complex mediation
(causal) model with intervening variables
– Indirect effect of income on life satisfaction runs
through freedom and self-efficacy

Self- Life
Income Freedom
efficacy satisfaction
Review
1. Conceptual distinction and selection of appropriate model
based on your research question; importance of theory
2. Mediation
• Barron & Kenny
• Sobel calculator
• Bootstrapping SPSS Macro
• Prepare and manipulate data for analyses
3. Run and interpret analyses, including supporting diagrams
• Advance extensions:
• Multiple mediator models / Chain mediation
Multiple Mediators
Model 4
Moderated Mediation
The effect of the mediator differs as a function of another variable (moderator).

(Edwards & Lambert, 2007)


Extensions: Moderated Mediation

(Edwards & Lambert, 2007)


Moderated mediation
Example
Testing Moderated Mediation
• Become clear about your theory or your
hypotheses as to which kind of moderated
mediation you expect
• Use PROCESS add-on
• Alternatively, specify the assumed path model
in SEM program.
• Alternatively, use multigroup analysis in SEM.
Resources: Moderated Mediation
• Edwards, J. R., & Lambert, L. S. (2007). Methods
for integrating moderation and mediation: A
general analytical framework using moderated
path analysis. Psychological Methods, 12, 1-22.
• Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007).
Addressing moderated mediation hypotheses:
Theory, methods, and prescriptions. Multivariate
Behavioral Research, 42, 185-227.
Mediated moderation
Mod.

X Med. Y

Identity

IG Rel.

X*Mod. Med. Y
Example
adapted from Montoya and Pittinsky (2011)

• When and how is group identification related to outgroup liking?

• Identification (‘Identity’) and Intergroup relations (IG Rel) were manipulated


• Identity – high vs. low
• IG Rel – competition and cooperation

• Trust and outgroup liking (‘OG liking’) measured on a Likert scale.

IG Rel.

OG
Identity Trust
Liking
Example
adapted from Montoya and Pittinsky (2011)

Outgroup trust Outgroup liking


8 Cooperative Competitive 7 Cooperative Competitive
7 6
6
5
5
4
4
3
3
2
2
1 1

0 0
High ID Low ID High ID Low ID
Example
adapted from Montoya and Pittinsky (2011)

IG Rel.

OG
Identity Trust
Liking

Identity

IG Rel.
+ (sig) + (sig) OG
ID*IGRel Trust
Liking

n.s.
Testing Mediated Moderation
• Use PROCESS add-on
• Or any SEM software (e.g., AMOS, MPlus)
– MPlus: Useful code:
http://offbeat.group.shef.ac.uk/FIO/mplusmedmo
d.htm#modindex
Longitudinal Mediation
• Cole & Maxwell (2009); Maxwell & Cole (2007)
Break
Mini-Quiz
Multiple Moderators Multiple Moderators
Mini-quiz
Revised Timetable (Session 3)
Session 1: Mediation
For this week and next, please read Baron & Kenny (1986) and Maxwell & Cole (2007)
Have SPSS downloaded on your laptop
Have Hayes’ Process Macro downloaded and installed on your laptop
 
Session 2: Moderation

Session 3: Advanced Topics in Mediation and Moderation

Session 4: Writing up Results // Defining your Research Question and Testing in SPSS
PPT Slide of research in your area
Your data!
Have your prepped/scored data in SPSS for this class session

Session 5: Refining your Research Question / Analysis


Practice makes proficient – two ‘homework’ write-ups, exchange with a peer [pairs]
Pairs
Troubleshooting, hands-on practice and feedback [breakout rooms]

Session 6: Student Presentations of Results


Short presentations with the main findings that will serve as the basis of your final assignment

Future: Introduction to Structural Equation Modelling // Introduction to Longitudinal Data Analysis


Next…
• Questions about Moderation Exercise?
• Statistical Power for Mediation
• Writing up Exercise 1 and 2
• Bespoke data questions
Questions about Moderation Handout?
What is statistical power?
• Correctly rejecting the null hypothesis (H0) when it is true in
the population.
• Type 1 error: False positive (α) [typically = .05]
• Type 2 error: False negative (β)
– Power = 1 – β [typically = .80]
• In essence, the trade off between reducing the potential
threat of both types of error.
• Statistical power needs 4 pieces of information:
– α, β (these are typically specified by the researcher),
– the effect size (from previous literature), and
– the sample size (what we can ‘control’).
Mediation: Statistical Power
Now, connecting to
your own research or area of interest.
Next Steps
Session 4 Session 5
1) Find a paper in your area… 1) Write-up Exercise 1 & 2
• Mediation, moderation, or advanced • Based on your handout for
topic
• Prepare a single PPT slide: each week, write up a
– RQ separate Results section for
–
–
Hypotheses
Supporting image
Mediation and Moderation.
– Results / Take-home message These may each be 1-2
– Be prepared to share in ~2 mins paragraphs, with
• Make notes: 2) Post RQ to Discussion
supporting Figures andForum
– What was clear, unclear about the writing or
presentation? captions.
– What would you borrow?
– What would you avoid? • You will exchange these
– Highlight or pull-out precise language /
structure to share.
with a peer in class.
Questions?

2) Have your data prepped/coded


Housekeeping
You asked?

• Statistically compare Betas:


https://www.youtube.com/
watch?v=O5OpUAybVOE
• Bootstrapped confidence
(please excuse his
intervals:
breathing!)
https://www.youtube.com/
watch?v=djsZg_iZNT0
Bespoke Data Questions?
Poll
tinyurl.com/PSY40630RQ
Break

You might also like