I am using metaprop to compute pooled binomial proportion and confidence intervals, entering for each observation the numerator (items of interest) and denominator (n) for each group. It is working well. However, I would like to conduct the same analysis over subgroups of the data. The metaprop option by() allows the analysis to be stratified and stores the individual estimates and confidence intervals per group, but it seems to display only one overall pooled result. I am using the Clopper-Pearson exact method. Does anyone know a formula that will derive the overall estimates (which I wish to do per subgroup) from the individual group estimates? Alternatively, is another option available in metaprop to store these?
-
Login or Register
- Log in with
