GPower for Mediation Analysis and its Alternative

The calculation of power in mediation analysis is a common problem if you want to test indirect effects. There are different possible methods do do a power calculation for mediation models.

How to use gpower for sample size calculations in mediation analysis
How to use an alternative, simulation analysis
How to calculate power for a mediation analysis if you use:
Baron&Kenny (causal steps approach)
Joint significance
Sobel-test
Bootstrapping (e.g. with Hayes' PROCESS macro).

0:00 Start
0:36 Four approaches to testing mediation
1:08 Using G*Power
1:41 Simulation analysis
2:30 Fritz & MacKinnon
3:24 Baron & Kenny / Causal Steps
4:47 Joint significance approach
6:31 Sobel test
6:50 Bootstrap / PROCESS
7:38 Power of .95 instead of .80

Here you can find the paper by Fritz and MacKinnon:
https://www.ncbi.nlm.nih.gov/pmc/arti...

Here you can find the software G*Power:
https://www.psychologie.hhu.de/arbeit...

Here are my other statistics tutorials:
http://www.regorz-statistik.de/en/tut...

How to book a statistics CONSULTATION for PROCESS (mediation, moderation, moderated mediation):
http://www.regorz-statistik.de/en/con...

• #education
• #power_mediation
• #g_power_mediation
• #gpower_mediation