Dear All,
I try to meta-analyse studies with pre-post designs, but I don't have
the correlations for the pre-post data.
In many studies you can read that the use of the formulas for
independent designs will underestimate the ES, but I tried this with
correlation of r=.4,r=.9 and with the formulas for independent group
designs and found that the ES for independent groups are always
bigger than the other two when I use the inverse variance as a
weight for a total ES.
Has anyone experience with this problem and an idea how to deal
with this problem ?
Thanks in advance,
Markus
|