[R-meta] Help with explaining why I should use rma.mv

Timothy MacKenzie |@w|@wt @end|ng |rom gm@||@com
Sun Feb 6 17:16:07 CET 2022

Dear Experts,

I'm writing a quick section in my paper explaining why I used a
multilevel meta-regression approach (using rma.mv).

But to simplify what that means, I get a bit stuck.

Below is the jist of what I've written, but is this technically accurate?

Tim M
"When studies give multiple effect sizes, then effect sizes from the
same studies are often viewed to be more similar to each other than
those from other studies.

This fact should make our inferences about the size of treatment
effect more conservative. Because, the effect sizes from each study
don't provide independent measures of evidence. Rather, they are
hierarchically dependent.

By analogy, this is similar to the issue of nested data structures
commonly found in educational research. For example, students (like
effect sizes) from the same classrooms (like studies) tend to respond
more similarly to an intervention resulting in hierarchical dependence
emerging in such data.

More information about the R-sig-meta-analysis mailing list