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 the contexts in which prepare_ma() is used #70

Closed wwiecek closed 4 years ago

wwiecek commented 4 years ago

Must make sure that 0 event corrections are not accidentaly applied in various steps of data prep

wwiecek commented 4 years ago

I think in particular in pooling, e.g. pooling.R:102, we grab the SEs from prepare_ma(). But should we be using these precomputed SE's at all? This might be a mistake (not specific to logit model)

wwiecek commented 4 years ago

Otherwise we're OK -- prepare_prior() uses prepare_ma(), but using corrections seems sensible to help us set the prior.

wwiecek commented 4 years ago

Also check the context in which the automated warning is emitted.