I'm using RevMan to calculate effect size of a pooled analysis of studies with non significant heterogeneity. I have chosen random effects for analysis model and Standardized Mean Difference as an effect measure. Is there any way to estimate the effect size (i.e. g) from the generated data of this pooled analysis using RevMan or is there any other available option for this purpose?
Hi Hovanness
Firstly, why have you used a random effects model when there is no significant heterogeneity? Please read the Cochrane handbook http://handbook-5-1.cochrane.org/
about heterogeneity and choice of model. The standardized mean difference for each study, is in fact Hedge's g. So the pooled estimate surely is a pooled estimate of Hedge's g.
I agree with Adrian above that perhaps a fixed effects model should be applied. The effect size should be in the forest plot.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis