[R-meta] HKSJ adjusted error by hand

[Yefeng Yang]

Dear community,

I am wondering whether there is a post-hoc way to calculate sampling variances of the estimated regression coefficients from meta-analysis models based on the Hartung-Knapp-Sidik-Jonkman method.

To be more precise,

  1.
I first fit a normal random-effect meta-analysis via `rma()` in metafor package:

library(metafor)
dat <- escalc(measure="RR", ai=tpos, bi=tneg, ci=cpos, di=cneg, data=dat.bcg)
mod <- rma(yi, vi, data=dat),


  1.
then, how can I get the adjusted standard error of the estimated coefficients (in this case, it is an intercept) based on the model object `mod`?

Of course, we can get it using an on-the-fly way:
rma(yi, vi, data=dat, test = "hksj")

But I have a couple of big datasets and want to report both the original standard and adjusted errors. I do not have a high-performance PC and I would like to avoid 're-fitting' the meta-analysis.

Best,
Yefeng





	[[alternative HTML version deleted]]