Skip to content

[R-meta] Conducting IPD meta-analysis [R]

3 messages · Sam Meyler, Michael Dewey, Röver, Christian

#
Dear Sam

There are more than 100 packages for meta-analysis listed in the 
relevant CRAN Task View https://cran.r-project.org/view=MetaAnalysis

Without more details about what the output from your first stage is 
likely to look like it is hard to recommend the best option but the two 
general purpose packages which most people use are meta and metafor. The 
authors of both of these monitor this list if you have detailed queries.

Michael
On 02/12/2020 19:13, Sam Meyler wrote:

  
    
  
2 days later
#
Dear Sam,

for IPD meta-analysis, you may in fact need to check out generic
software for mixed models (rather than dedicated meta-analysis
software).

I am currently working on a (Bayesian) IPD meta-analysis, and we are
using the "brms" package (https://cran.r-project.org/package=brms).

For a frequentist analysis, I guess the "lme4", "nlme" or similar
packages may be useful.

Cheers,

Christian
On Thu, 2020-12-03 at 13:26 +0000, Michael Dewey wrote: