[R-meta] Inquiry on metafor trimfill
li@t@ @ending from dewey@myzen@co@uk
Sat Sep 22 14:43:22 CEST 2018
I think the answer to your question is probably "No". Although it is 13
years old I think Wolfgang's article cited in the help page of rma.uni
as Viechtbauer (2005) gives that advice. Worth reading (or re-reading)
On 22/09/2018 01:12, Chishio Furukawa wrote:
> Dear all,
> I used the "trifill" command in the metafor package, which provided various
> options for estimating the heterogeneity across studies. My objective was
> to estimate the coverage probabilities of the trim-and-fill method (Duval
> and Tweedie 2000) for comparison with other methods.
> Given the original article in 2000, I think it is adequate to use "DL"
> command (standing for formula in DerSimonian and Laird 1996), but I was not
> confident what methodology is most commonly used. Could I ask if there are
> recommendations on which formula to use among HS, HE, SJ, ML, REML, and EB?
> It appears that the estimation result of coverage probability can change
> importantly depending on which method to use.
> Thank you!
> Best regards,
More information about the R-sig-meta-analysis