NEWS.md
BREAKING CHANGES
Removes the unnecessary aliases for certain tests: bf_one_sample_ttest, bf_two_sample_ttest, and bf_onesample_proptest.
The output dataframe now only contains results for the alternative hypothesis (BF10) and its log value to avoid cluttered dataframe since all the other columns were essentially mathematical transformations of it and not really relevant.
The output argument’s default value has been changed from too generic "results" to "dataframe".
MAJOR CHANGES
broomExtra dependency with parameters.BREAKING CHANGES
bf_expr now expects not a dataframe, but rather the BayesFactor object.MAJOR CHANGES
bf_extractor now internally uses parameters::model_parameters to extract as many details from the BayesFactor as possible. So the returned dataframe will contain additional details now.
The redundant text related to whether the results in favor of the null or the alternative hypothesis has now been removed, since this information is already present in BF subscripts (01: null, 10: alternative).
MAJOR CHANGES
ggstatsplot-verse, the minimum R version is bumped to 3.6.0.MINOR CHANGES
bf_contingency_tab function no longer includes sampling plan information in the expression.
Internal refactoring to improve data wrangling using ipmisc.
MAJOR CHANGES
bf_meta deprecates messages argument. If users want to see the full output from metaBMA, they are advised to run it independently.
All relevant functions gain a hypothesis.text argument to remove the redundant pretext before the expression containing the results.
MINOR CHANGES
Thanks to Sarah, the package has a hexsticker. :)
Re-exports metaBMA::prior() for statsExpressions.
Removed unnecessary dependence on broomExtra.