dot-run_leave1out: Fit Multiple Meta-Analytic Models For Leave-One-Out Analysis

In daniel1noble/orchaRd: Orchard Plots and Prediction Intervals for Meta-analysis and Meta-regression

Fit Multiple Meta-Analytic Models For Leave-One-Out Analysis

Description

Iteratively refits a meta-analytic model, leaving out one level of a specified grouping variable in each iteration. This internal function handles the actual model refitting process for the leave-one-out analysis.

Usage

.run_leave1out(
  model,
  group,
  vcalc_args = NULL,
  robust_args = NULL,
  phylo_args = NULL
)

Arguments

model	A fitted metafor model object containing a data element and call object.
group	A character string specifying the column in model$data that defines the groups to be omitted one at a time.
vcalc_args	Optional list of arguments for the variance-covariance calculation using metafor's vcalc function. See leave_one_out() for details.
robust_args	Optional list of arguments for robust variance estimation using metafor's robust function. See leave_one_out() for details.
phylo_args	Optional list of arguments for phylogenetic matrix calculation. See leave_one_out() for details.

Details

The function creates a subset of the original data for each unique value in the grouping variable, removing that group and refitting the model using the same specification as the original model. If variance-covariance matrices or phylogenetic corrections were used in the original model, these are recalculated for each subset. If a model fails to fit, NULL is returned for that group.

Value

A named list of models, each fitted after omitting one group. Names correspond to the omitted group IDs. Any failed model fits will be NULL.

Author(s)

Facundo Decunta - [e-mail]

daniel1noble/orchaRd documentation built on April 17, 2025, 3:59 a.m.