[R-meta] Quick question about multiple independent samples within one study
@c @ending from imbi@uni-freiburg@de
Thu Dec 13 14:17:27 CET 2018
I would suggest that you have a look at Rücker et al. (2017) as well as
sections 16.5.4 and 16.5.5 of the Cochrane Handbook
These papers deal with multiple treatment groups, however, the same
arguments apply if you have different age groups etc. within a study.
In summary, it is fine to include multiple independent groups from a
study under a fixed effect model. However, you should combine the groups
within a study under the random effects model.
Best wishes, Guido
Rücker G, Cates CJ, Schwarzer G. Methods for including information from
multi-arm trials in pairwise meta-analysis. Res Synth Methods.
More information about the R-sig-meta-analysis