[R-meta] Subgroup analysis best practices

[Tamar Novetsky]

Hello all,

I have a dataset with topline and subgroup-level treatment effects, and I
would like to be able to answer questions like "[how much] does this
subgroup perform better/worse than average?". I know that I can use ANOVA
to test whether there is variation between the groups, but I am having a
hard time figuring out the best way to compare each subgroup to the others
or to the overall effect. I would rather be able to say how a subgroup's
effect compares to the overall effect (or the effect of all people not in
that subgroup), rather than make pairwise comparisons between the
subgroups. Are there best practices on statistical tests to use for this
purpose?

Thanks,


*Tamar Novetsky* *(she/her)*
Data Scientist I
Eastern Time Zone

	[[alternative HTML version deleted]]