I am new to meta-analysis. I did a quick search, and I think I can wrap my mind around generating cohen's d and its variance when there is only a single distribution for each of the control group (C) vs the experimental group (E) within each study. But my mind gets lost when each group (C or E) also has both baseline and follow-up distributions.
Below are my thoughts how this can proceed:
- Ignore the baseline measurements and just use the follow-up distributions to calculate the cohen's d distribution between the C and E groups, but this will ignore the part of the effects related to differences in baseline measurements between the C and E groups(, although quite many studies have similar baseline)
- Construct a distribution for the difference between baseline and follow-up distributions in each group (D_diff). Then compute the cohen's d distribution by comparing D_diff between C and E groups. Like, computing the difference between differences.
I have a feeling that #2 is the way to go since #1 implies a loss of information. However, I am not sure how D_diff can be constructed - could someone please help me start on this? Or, if I am on the wrong track, where should my next step be?
Thanks in advance!