Hello all,

I am working on a meta-analysis of 30 studies to report the pooled means stratified into 4 subgroups depending on the technique used to compute the mean. I use the following syntax in Stata “metan mean se, randomi by (technique)” and I get the following results:

Subgroups no. of studies mean 95% CI

subgroup 1 20 studies 2.71 2.65 - 2.84

subgroup 2 2 studies 2.60 2.55 - 2.81

subgroup 3 1 study 2.48 2.35 - 2.54

subgroup 4 1 study 2.55 2.49 - 2.63

Is it possible to compare the results of the 4 subgroups altogether as well as the results of each 2 subgroups apart (subgroup 1 vs subgroup 2, subgroup 1 vs subgroup 3, etc....). If yes, which statistical tests should I use and how to run these tests in Stata?

Thank you for your input and best wishes

More Samir Haffar's questions See All
Similar questions and discussions