Skip to contents

Changelog

Source: NEWS.md

rwa 1.0.1

CRAN release: 2026-09-28

Improvements

  • Fixed the CRAN macOS ARM test error in the near-collinear fixture of “estimable highly correlated predictors retain their valid fit” (#29). The predictors correlated at 1 - 5e-9, giving a condition number near 4e8 and a platform-dependent rounding error above the test’s tolerance. They now correlate at about 0.99995, which stays highly collinear and estimable with a much smaller rounding error. calculate_rwa() is unchanged, since the deviation was bounded floating-point error rather than a calculation defect.
  • Applied the same treatment to the neighbouring test “additional predictors do not impose a new conditioning cutoff”, whose 1.1e-7 offset gave a condition number near 3e14 and left only about 7x headroom under the 1e-8 tolerance. It now uses a 1e-4 offset. The assertions in both tests were strengthened, not relaxed: the predictor correlation is pinned within (0.9999, 1), the raw relative weights are asserted to sum to the returned R-squared, and rwa() is asserted to agree with rwa_multiregress(). No returned value is clamped or repaired to satisfy an assertion.
  • The R CMD check workflow now runs on macOS arm64 with R release and old-release alongside the existing Ubuntu R release job, covering the CRAN platform that reported the error.

There are no user-facing changes in this release.

rwa 1.0.0

CRAN release: 2026-09-16

Breaking Changes

  • Raw.Significant is now derived from a comparison against a randomly generated variable’s weight, rather than from the confidence interval around the raw weight itself (#26). Raw relative weights are non-negative, so an interval around a weight almost always excludes zero and previously flagged even unrelated predictors as significant. Comparing each weight to that of a random variable is the approach suggested by Tonidandel, LeBreton and Johnson (2009, https://doi.org/10.1037/a0017735) for judging whether a weight exceeds what chance alone would produce (see also the discussion in vignette("evaluating-rwa-method-reference")). This package applies a directional cutoff on that comparison: a predictor is significant only when the lower bound of the difference interval is above zero, so a predictor that performed worse than the random variable is not reported as significant either. Predictors previously reported as significant may now correctly be reported as not significant. rwa() also returns Random.Diff.CI.Lower and Random.Diff.CI.Upper for the comparison the flag is based on, and the descriptive Raw.RelWeight.CI.* columns are unchanged.
  • Because significance requires this comparison, bootstrap = TRUE now runs an additional bootstrap when comprehensive = FALSE, which roughly doubles bootstrap time. comprehensive = TRUE already computed the comparison and is unaffected.

New Features

  • Added rwa_logit() and rwa_multiregress() to support logistic regression and multiple regression.
  • Added new vignette to cover the new regression methods.
  • Added use parameter to rwa() function to control how missing data is handled when computing correlations. Options include “pairwise.complete.obs” (default, pairwise deletion), “complete.obs” (listwise deletion), and other standard options from cor(). (#12)
  • Added weight parameter to rwa() function to perform observation-weighted Relative Weights Analysis (#12), using a weighted complete-case correlation matrix. Bootstrap inference uses iid individual-row resampling, carrying each row’s weight; clusters, strata, and replicate-weight survey designs are not supported.
  • Weighted results from both rwa() and rwa_multiregress() now include n_weighted (sum of retained original weights) and n_effective (Kish’s unequal-weighting effective sample size). Existing unweighted return fields are unchanged.
  • Added a weighted/missing-data vignette explaining filtering, weight scaling, diagnostic counts, and bootstrap limitations.
  • Updated the introductory and regression-methods vignettes to cover the use and weight arguments, the weighted sample-size diagnostics, and the multiple-regression-only scope. Corrected the introductory vignette’s incorrect statement that missing data is handled by listwise deletion; the default has been pairwise deletion.

Improvements

  • Updated all bootstrap functions to support the new use and weight parameters
  • Shared numeric, finite, strictly positive weight validation across point estimates and bootstrap calculations; missing-weight filtering remains mode-dependent.
  • Validate the full joint correlation matrix and predictor invertibility with documented numerical tolerances and actionable errors for missing-data-induced indefinite matrices, constant variables, singularity, and insufficient data (#24). Exact fits remain valid when predictors are not collinear.
  • Preserve legacy missing-data preprocessing, including outcome removal before every correlation mode and weighted predictor-completeness filtering even for all.obs; corrected na.or.complete documentation.
  • plot_rwa() now reports the sum of weights and the effective sample size in the caption for weighted analyses, so weighted charts are distinguishable from unweighted ones.
  • Weighted results return n_weighted and n_effective immediately after n, making them easier to find. Field names and unweighted output are unchanged.
  • Datasets with fewer usable observations than predictors now name the sample-size problem in the singular-matrix error, instead of reporting only an eigenvalue (relevant to the rank-deficient data discussed in #10). Models that were previously estimable, including pairwise-deletion models with few complete cases, are unaffected.
  • Added a plain-English summary of the weighting and missing-data behavior to ?rwa, including guidance on when to use survey weights.
  • Improved test coverage and minor bug fixes

Bug Fixes

  • Fixed weighted single-predictor calculations and random-comparison name collisions, including weight columns named rand.
  • Bootstrap samples now preserve predictor identity, order, and statistic length. Invalid or degenerate samples error instead of dropping variables or recycling estimates; no samples are skipped or retried.
  • Comprehensive bootstrap now computes random comparisons without a focal predictor and correctly labels and maps random/focal intervals to the requested predictors.

rwa 0.1.1

CRAN release: 2026-01-20

Improvements

  • Input validation: Added comprehensive validation for rwa() parameters including conf_level, n_bootstrap, non-numeric variables, zero-variance variables, and singular correlation matrices with informative error messages
  • Code refactoring: Consolidated duplicate RWA calculation code in bootstrap functions into a single internal helper
  • Documentation: Added links to pkgdown site in README; fixed internal function documentation

Bug Fixes

  • Fixed flaky bootstrap test by using a fixed random seed for reproducibility (#20)

Tests

  • Added extensive tests for input validation and edge cases (collinearity, small samples, invalid parameters)

rwa 0.1.0

CRAN release: 2025-07-16

New Features

  • Bootstrap confidence intervals: Added bootstrap = TRUE parameter to rwa() for statistical significance testing of relative weights
  • Result sorting: Added sort = TRUE parameter to automatically sort results by importance (descending order). Set sort = FALSE to preserve original predictor order
  • Comprehensive vignette: New detailed documentation covering methodology, examples, and best practices
  • Enhanced documentation: Updated README and function documentation

Technical Improvements

  • Package compliance: Updated DESCRIPTION with proper Authors@R field for CRAN submission
  • CI/CD: Enhanced GitHub Actions workflow with vignette building support
  • Dependencies: Added boot, purrr, and utils packages for bootstrap functionality
  • Code quality improvements: Fixed long lines in R code to meet CRAN standards
  • Documentation cleanup: Improved code formatting and removed unused variables
  • Enhanced vignette formatting: Cleaned up formatting in comprehensive vignette documentation

Bug Fixes

  • Fixed vignette compilation issues
  • Resolved R CMD check warnings and notes
  • Removed unused variables to eliminate R CMD check notes
  • Improved consistency in code formatting

Version 0.0.3

Re-submission to CRAN

  • Unwrap \donttest{} in examples where unnecessary

Version 0.0.2

Re-submission to CRAN

  • DOI references added to DESCRIPTION
  • Added CodeFactor badge
  • Typos in DESCRIPTION rectified

Version 0.0.1

First submission to CRAN (required to re-submit)