search

wwiecek / baggr

R package for Bayesian meta-analysis models, using Stan
GNU General Public License v3.0
46 stars 12 forks source link

check cross-validation for IPD models #98

Open wwiecek opened 3 years ago

wwiecek commented 3 years ago

Make sure that if leaving out group X, what happens is that estimation is done for control part of group X and then ELPD calculated for treatment part of group X

wwiecek commented 3 years ago

This applies to rubin_full and logit both

wwiecek commented 3 years ago

I checked that it works and wrote a test of syntax, but need to check correctness