Hello,

Since your assumption is merely to add subgroups' means (SD), I think the Cochrane calculator might work well for you. Nevertheless, I highly recommend reading the following page.

Please press Show All;

https://www.statstodo.com/CombineMeansSDs.php

Note that in case you were not talking about subgroups i.e. combining data from separate populations or at separate environments, I think random effect meta-analysis would be much better.

You can use the Cochrane formulae. The different populations are the different hypothetical infinite populations of the different random variables (the different facet measures) .

• Thank you Hossam Tharwat Ali and James Leigh . I actually tried the Cochrane formulae for a made up sample that I created with similar condition to the study. However, the result that I calculate normally was slightly different than the one I calculated using Cochrane formula. Is this because the formula always provide a slight error?
• Perhaps the following paper would illustrate my question more clearly. This paper mentioned exactly what I wanted to do at page 380. Specifically, it said that "Where a study provided multiple data for a personality dimension, we used the mean effect size for the meta-analysis." However, I could not find the formulae that the paper use to calculate the mean effect size. Similar papers also reported using a mean or average effect size without provide any formula.

  Article ADHD and personality: A meta-analytic review

  https://doi.org/10.1016/j.cpr.2014.05.002