115 compare cor statistic not z score

[arcaldwell49]

Fix for issue #115 raised by @Lakens

This should resolve the issue, update the documentation, and incorporate a unit test to avoid the issue in the future.

[Copilot]

Pull request overview

Fixes issue #115 in TOSTER where compare_cor(method = "fisher", alternative = "equivalence") reported an unstandardized numerator as the z statistic, despite computing p-values from the standardized value.

Changes:

• Update compare_cor() to report the standardized z-statistic ((diff - bound) / SE) for TOST-style alternatives.
• Add documentation clarifying Fisher-z handling of null bounds and clarify what statistic represents.
• Add a regression test to ensure result$statistic matches the standardized z and reproduces the p-value.

Reviewed changes

Copilot reviewed 5 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
R/compare_cor.R	Fixes reported statistic to be standardized (matches p-value computation) and clarifies Fisher-z bound behavior in docs.
man/compare_cor.Rd	Mirrors documentation updates about standardized statistic and Fisher-z bound conversion.
tests/testthat/test-corr.R	Adds regression test for issue #115 to prevent reintroduction.
R/boot_ses_test.R	Adds roxygen import for stats::runif used unqualified in the implementation.
NAMESPACE	Adds importFrom(stats,runif) (consistent with roxygen-generated namespace).
.gitignore	Adds an ignore pattern for a Claude local settings file.

[Lakens]

Excellent, thanks for the update (am using this function in a new paper, which is how I noticed this).