Released on CRAN: June 11, 2026
Added leave1studyout(), a new S3 method for evSyn objects that performs a leave-one-study-out sensitivity analysis: the evidence synthesis is recomputed each time with one study omitted. Supports all IC types (GORIC, GORICC, GORICA, GORICAC) and all type_ev options ("added", "equal", "average").
Added support for goric() for objects of class effectlite.
Extended evSyn() to accept a plain data.frame as input in addition to escalc objects from the metafor package.
Extended the order_studies argument in evSyn() and all evSyn_* subfunctions to accept a numeric permutation vector, allowing a fully custom study ordering in addition to the existing "input_order", "ascending", and "descending" options.
Fixed a bug in evSyn_est() where study_sample_nobs passed via an escalc object was a list rather than a vector, causing GORICAC computations to fail; resolved by adding an unlist() call.
Improved and expanded error and warning messages throughout to enhance clarity and diagnostics.
Improved the plot() method for evSyn objects with better handling of axis labels and study ordering.
Internal refactoring of chi-square-bar weight computation and IC weight functions for improved numerical robustness.
Released on CRAN: February 5, 2026
Added support for gorica() for lm objects using the unbiased covariance matrix of parameter estimates.
Extended gorica() for lavaan objects to allow the use of the joint covariance matrix of parameter estimates, enabling simultaneous evaluation of both defined and labeled parameters.
Improved and expanded error and warning messages to enhance clarity and diagnostics.
Released on CRAN: December 20, 2024
This is mainly a maintenance release.
New argument Heq for the goric function: If Heq = TRUE and comparison is the complement, the null-hypothesis is added to the set of hypotheses.
Added support for the evSyn() function for objects of class escalc from the metafor package.
The output of the goric() function now shows clear conclusions and more informative messages.