Expanded release-facing documentation to describe the package’s broader workflow coverage: univariate, multivariate, multilevel, and network meta-analysis; within-study covariance construction; diagnostics; publication-bias models; resampling sensitivity analyses; reporting helpers; review graphics; and exploratory random-forest meta-analysis.
Added README workflow guidance that points users to the primary entry point and relevant vignette for each common analysis goal.
Organized pkgdown reference and article navigation into workflow groups.
Expanded the mars() reference documentation with accepted argument values, data-shape guidance, and a more explicit return-value description.
Added examples for exported network resampling and permutation helpers and converted many long-running examples from \dontrun{} to \donttest{} so they remain visible as executable templates.
Notes
Random-forest meta-analysis helpers remain exploratory and are documented as moderator-screening and prediction tools rather than replacements for the likelihood-based MARS estimators.
