1
$\begingroup$

I was wondering how the weight of each study is determined in a Forest Plot?

Here is a theoretical example of weights appearing in a Forest Plot using the metafor package.

q  <- escalc(measure="OR", ai=ai, bi=bi, ci=ci, di=di, data=q)
q1 <- rma(yi, vi, data=q, slab=paste(study, year, citation, sep=", "), method="FE")

forest(q1, showweights=TRUE,...)

enter image description here

$\endgroup$
2
$\begingroup$

The inverse of the variance is the standard weighting, but the metafor package has a ton of options. Introduction to Meta-Analysis by Borenstein, Hedges, Higgins, and Rothstein is a standard textbook on this that I think explains inverse weighting in an accessible way. The metafor package has fantastic documentation, and you can find more information on the options for analyzing your meta-analysis with ?metafor::rma() or by going to the website: http://www.metafor-project.org/doku.php/features#models_and_analysis_approaches It says there that: "The package provides a variety of models and analysis approaches, including: fixed-, random-, and mixed-effects models using the inverse-variance method (rma() function)" (emphasis mine).

If the effect size in Study 1 has a variance of 10, the weight will be 1/10. If the effect size in Study 2 has a variance of 2, the weight will be 1/2, etc.

$\endgroup$
1
  • $\begingroup$ Perhaps worth adding that your last sentence is only true in general for fixed effects models (which to be fair is what the OP fitted). $\endgroup$ – mdewey Aug 14 '18 at 12:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.