v3.0.0

Changelog

Major code Changes

• For bootstrapping confidence intervals, bcaboot was replaced by the boot package. Bootstrapping is still slow, but boot performs substantially faster single-core and supports parallel processing.
• The function split_ci does not have a default value for fn_coef anymore and now returns an object of class boot

Documentation

• Updated bootstrap-related reference documentation and vignettes
• Updated the RAPI vignette to use the Flanagan-Rulon coefficient
• Added new split-half reliability research and updated splitting methods recommendations accordingly
• Made various small documentation improvements.

v2.2.2

This release has no features nor bug-fixes; it only contains changes to have splithalfr be complatible with newer versions of R and modern software citation practices:

• Fixed. The citation files now refer to a Zenodo DOI instead of the GitHub repo.
• Fixed. The inst/CITATION file now uses the bibentry API.
• Fixed. Added a "_PACKAGE sentinel" to the package documentation. See r-lib/roxygen2#1491
• Removed. Removed a bunch of dead links in the documentation.

v2.2.1

This is only for obtaining a "Concept DOI" via Zenodo. See the actual v2.2.0 release.

v2.2.0

Changes in version 2.2.0:

• New feature. When making permutated splits (method = "random", replace = FALSE), split_p values above 0.5 yield an unbalanced split. In this unbalanced split, score_1 is calculated from split_p% of the trials and score_2 from (1 - split_p)%. See this test.
• Documentation update. Fixed a bug in the calls to split_ci in the vignettes (thnx go out to Maren Sera) and aligned terminology with the compendium paper.

v2.1.1

Version 2.1.1 consists of some small documentation updates, such as links to the published version of the compendium paper (Pronk et al., 2021) instead of the preprint.

v2.1.0

• Added the option to subsample participants before splitting via subsample_n argument to by_split.
• Updated terminology in documentation to match compendium paper
• It's available via CRAN or as R source package