[R-meta] calculate effect size for pre-post control design
Yingkai Yang
yykj|y|@u|p|@n @end|ng |rom gm@||@com
Tue Jan 22 02:30:18 CET 2019
Hi,
I am wondering how to calculate effect size for pre-post control design.
For example, I have a data:
group A: pre mean= 83.91, post mean= 83.24, preSD = 14.9, postSD= 14.88, n=
40;
group B: pre mean= 80.54, post mean= 80.71, preSD = 15.28, postSD= 15.36,
n= 41;
pre-post r= 0.995 (all data)
When I used the code in Metafor website
http://www.metafor-project.org/doku.php/analyses:morris2008?s[]=morris, I
got an effect size of -0.055.
This is confusing，since the the true effect size [The intervention
(between-group) effect for change] I got from the raw data is -0.56.
Then, I used the software of Comprehensive Meta-analysis 3. I found that,
when calculating effect size from pre-post control design, you can select
to use change score SD, or post score SD. When I choose change score SD,
the effect size is indeed -0.56, which is correct.
So, my question is what is the difference between change score SD and post
score SD (or pre score SD). How to select these SDs?
best,
Yang
[[alternative HTML version deleted]]
More information about the R-sig-meta-analysis
mailing list