[R-meta] Response Ratios in nested studies

[Reza Norouzian]

Dear Fred,

I'm not aware of any work addressing this issue. Obviously due to
auto-correlation, your sampling variances can be erroneously (much)
smaller than what they should be.

I would say that your best bet at this time is to make use of some
form of GEE estimation method to alleviate the auto-coorelation issue.
If your initial model only consists of hierarchical levels (with no
crossed levels), then clubSandwhich package
(https://cran.r-project.org/web/packages/clubSandwich/index.html) can
provide some protection against the ignored auto-correlations.

Kind regards,
Reza

On Fri, Oct 15, 2021 at 8:59 PM Farzad Keyhan <f.keyhaniha using gmail.com> wrote:
>
> Hello All,
>
> I recently came across a post
> (https://stat.ethz.ch/pipermail/r-sig-meta-analysis/2021-October/003330.html)
> that discussed an issue that is relevant to my meta-analysis.
>
> In short, if some studies have nested structures, and the effect size
> of interest is log response ratio (LRR), is there a way to adjust the
> sampling variances (below) before modeling the effect sizes?
>
> vi = sd1i^2/(n1i*m1i^2) + sd2i^2/(n2i*m2i^2)
>
> Thank you,
> Fred
>
> _______________________________________________
> R-sig-meta-analysis mailing list
> R-sig-meta-analysis using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis