For each study, compares the observed naive treatment effect to the
posterior predictive distribution under the fitted model. Uses full
posterior draws for all quantities including standard errors.
Usage
pp_check_effects(fit, study_id = NULL, prob = 0.9)
Arguments
- fit
A meta_did_fit object with method = "sample".
- study_id
Character vector of study IDs to include. NULL (default)
shows all studies.
- prob
Width of the credible shading (not currently used but reserved).
Default 0.9.
Value
A ggplot2 object, faceted by study.