I'm in the process of selecting articles for my meta-analysis study. I found an article in which the number of non-event in the control trial was zero. Here is the stats I have:
events non-event N
treatment group 42 3 45
control group 46 0 46
Can I include this article in my study? Which effect size should I use for this article?
Thank you!
Dear Naike Wang,
if you are using R the continuity correction is applied automatically. If you want to edit data manually you have to add 0.5 as suggested by Gyanendra Kumar, beware that you have to add 0.5 in all the cells and not only in cells with 0 value.
Also note that the effect size you choose can change results of the analysis when dealing with studies with 0 cell frequencies, see the paper attached http://bmcmedresmethodol.biomedcentral.com/articles/10.1186/1471-2288-7-5
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis