I've averaged multiple effect sizes within the same study to a single mean effect size per study, however I know this method underestimates the study's standard error and overestimates study precision. Can I implement other methods (RVE, or a three levels model) within the Comprehensive Meta-Analysis program, or do I need to use R or another program?
Hello Nicole,
The CMA manual (https://www.meta-analysis.com/downloads/Meta-Analysis%20Manual%20V3.pdf) is pretty much mute on this point, unless the multiple effect sizes are distinguishable by: (a) being reported for different subgroups; and/or (b) representing different outcome variables. You could try contacting the support line: [email protected] to see whether they could offer more helpful advice.
And, yes, there is R, with a number of good program libraries for meta-analysis (e.g., robumeta, clubSandwich).
Good luck with your work.
See this paper
Article Methods to estimate the between-study variance and its uncer...
- Badges
- Science topic
- Similar topics
- Computer Science
- Program