NEWS.md

In kappaGold: Agreement of Nominal Scale Raters (with a Gold Standard)

kappaGold 0.4.0

• new function kappa_test_corr add test for difference in kappa in correlated groups (same subjects) through bootstrap
• modify API:
  • rename entries to se (instead of SE)
  • kappam_gold has parameter refIdx= to specify the column of gold standard
  • kappam_vanbelle works with single matrix of ratings (ratings=) and refIdx= allows to specify one of the groups of raters

kappaGold 0.3.2

• enhance description of package (upon request from CRAN)
• little code optimizations

kappaGold 0.3.1

• add datasets
• drop dependency on obsolete bootstrap-package: last bits

kappaGold 0.3.0

• kappam.gold
  • use z-quantile (std normal) for confidence interval (not t-quantile)
  • nRaters refers to number of evaluated raters (excluding gold standard)
• kappam.fleiss implemented
  • based on irr::kappam.fleiss
  • allow to set ratingScale=
  • jackknife method for bias-correction and SE
• kappaTest implements a homogeneity test for kappa's from independent groups of raters
• rename: kappaInference becomes simulKappa (more clear)
• drop dependency on obsolete bootstrap-package: use own implementation of jackknife

kappaGold 0.2.3

• kappa2: acknowledge that we have SE only for classical kappa2 (not for robust variant)

kappaGold 0.2.2

• kappa2: add std. error for unweighted kappa

kappaGold 0.2.1

• Cohen's kappa kappa2
  • simplified calculation of random agreement
  • return more information as list
  • allow to set ratingScale=
  • allow for robust= variant

kappaGold documentation built on April 4, 2025, 1:02 a.m.