search

wwiecek / baggr

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

Allow baggr_compare()'ison of covariates (fixed effects) #50

Closed wwiecek closed 3 years ago

wwiecek commented 4 years ago

Once the next PR is done

be-green commented 4 years ago

Fixed effects meaning pooled treatment effects? Fixed effects (as a term) is confusing for economists because it refers to a non-parametric approach to estimating unobserved average differences between individuals! So kind of like random effects as most other people use it.

wwiecek commented 4 years ago

Fixed effects as in fixed and random effects. But, yes, I know about this confusion and won't be using FE/RE without explanation in the package.

wwiecek commented 4 years ago

@be-green Is that in the current version?

be-green commented 4 years ago

Not sure what this means exactly, so I'm gonna say maybe but probably not? Is this in the case where we have covariates other than the randomly varying measurements?

wwiecek commented 4 years ago

Measurements themselves are not covariates. Don't worry about it, I'll add!

wwiecek commented 3 years ago

This is now in the v0.6.9, but the next step is to add `baggr_compare(strata = "covariate") type syntax, which also groups outputs by levels of a given covariate