cs_get_summary() now returns raw numeric values instead
of formatted character strings (e.g., 0.5 instead of
"50.0%"). This facilitates downstream analyses and plotting
(e.g., in sensitivity analyses). Formatting is now handled exclusively
by the respective print() and summary()
methods.print() and summary() methods
to be lighter and more maintainable. The output no longer relies on the
cli package, which is now used solely for messages and
warnings. This is an internal change to improve code readability;
calculation results remain unaffected.cs_anchor(),
cs_percentage(), cs_distribution(),
cs_statistical(), cs_combined()) now calculate
group-wise statistics if the group argument is
provided.garden_club,
trackmania, columbo) to demonstrate various
analysis methods.clinical_significance() is
deprecated. The package now offers several different approaches to
clinical significance analyses, namely:
cs_anchor() (Anchor-based approaches)cs_percentage()(Percentage-change approaches)cs_distribution()(Distribution-based approaches)cs_statistical()(Statistical approaches)cs_combined()(Combined approaches)The function clinical_significance() used the combined
approach (distribution-based and statistical). So the identical analysis
can be carried out with cs_combined().
The function check_cutoff() is deprecated.
When method = "HLM, participants with three or more
data points will be used. Before, participants with at least three data
points (i.e., 4 or more) were used.
You can now specify a significance level at which you want the
analysis to be based on with the significance_level
argument. The default is significance_level = 0.05. Note
that for this is the Phi max level when you set
method = "HA" as outlined in their article.
The package now offers to combined approaches: The classic combination of statistical and distribution-based approaches but also the combination of statistical and anchor-based approaches
We included the function cs_get_model() to retrieve
the hierarchical linear model that is fit during the distribution-based
approach if the HLM method is requested.
The hechler_2014 dataset was included to ensure
reproducibility of the JSS article.
|>magrittrcheck_cutoff() as in other
plot functionssummary() and plot()
methodplot() to give the user
more control over the overall appearanceaes_() was deprecatedNEWS.md file to track changes to the
package.