Generates a comprehensive narrative interpretation across one or more bayesma fits – overall effects, heterogeneity, publication bias, model averaging, sensitivity, model comparison, and convergence diagnostics. The function auto-detects what is provided and assembles only the relevant sections.
Usage
interpret(
...,
null_range = c(-0.1, 0.1),
effect_label = NULL,
credible_level = 0.95,
quiet = FALSE
)Arguments
- ...
One or more fitted bayesma objects (named or unnamed). Accepted classes:
bayesma,bayesma_mv,bayesma_robma,bayesma_egger,bayesma_metareg,bayesma_robma_sensitivity,bayesma_comparison.- null_range
Optional length-2 numeric vector for direction/ROPE probabilities. Defaults to
c(-0.1, 0.1)on the natural scale.- effect_label
Optional character override for the effect label used in narrative text (e.g. "log_or"). Inherited from fits when
NULL.- credible_level
Credible interval width used in summaries. Default
0.95.- quiet
If
TRUE, suppresses progress messages during assembly.
Value
An object of class bayesma_interpretation – a list with one
element per detected section plus a meta slot. The print() method
renders the full narrative report.
Examples
if (FALSE) { # \dontrun{
fit <- bayesma::bayesma(data = dat, yi = "yi", sei = "sei")
rob <- bayesma::robma(data = dat, yi = "yi", sei = "sei")
egg <- bayesma::egger(data = dat, yi = "yi", sei = "sei")
sens <- bayesma::robma_sensitivity(data = dat, priors = my_priors)
interpret(fit, rob, egg, sens)
} # }