Simulations always work well for situations like this. Simulate models
that have the minimum degree of mediation you would consider important
and see what sample size it takes to show significant mediation, say,
80% of the time. One question is how to test for mediation. The Sobel
test has been shown to have low power. Bootstrapping Hs been shown to
work better. Kris Preacher has a paper with both SPSS and SAS code for
doing bootstapped tests of mediation.
Paul R. Swank, Ph.D.
Professor and Director of Research
Children's Learning Institute
University of Texas Health Science Center - Houston
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of
Cesar A. Gonzalez
Sent: Friday, February 22, 2008 9:13 AM
To:
[hidden email]
Subject: mediation analysis- sample size?
I am planning on doing a mediation analysis by doing multiple
regressions (Baron
and Kenny).
In terms getting a sample size required (power analysis) for a mediation
analysis...
I do not know. Has anyone ever done this? ANY info is appreciated.
Thank you.
Cesar
________________________________________________________________
Sent via the WebMail system at bienestar.org
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
