// Reporting Kruskal-Wallis-test - results from R //
The Kruskal-Wallis-test in R usually consists of three parts:
- First, the Kruskal-Wallis-test itself, which tests the medians of at least three groups for differences. In addition it is common practice to report the effect size Eta² for the test itself.
- Second, in case of a significant Kruskal-Wallis-test, post hoc-tests to determine between which of the groups a difference exists, should be calculated.
- Third, the effect size r for group differences that showed a low enough p-value.
The reporting is limited to the essentials: the test statistic (H-value), the degrees of freedom, the p-value and results for post-hoc-tests, consisiting of p-value and effect sizes respectively.
The effect size can be classified with comparable studies or research field specific thresholds, optionally with Cohen (1992): A Power Primer.
Please be aware that an effect size for the test itself is not really desirable since you are in almost all cases more interested in the specific pairwise differences. It is more frequently used to compare your study with others, for instance when it comes to a priori sample size calculations.
Sample formulation:
=================
"The Kruskal-Wallis-test showed differences between the three observed groups (H(2) = 32.02; p ◃ .001; η2 = .63).
Post-hoc-Analysis showed differences between the control group and meditation group (p ◃ .001; r = .52) as well as the control and mentored group (p ◃ .001; r = .92)
Both effects are large, according to Cohen (1992).."
⏰ Timestamps:
==============
0:00 Introduction and overview
0:09 Kruskal-Wallis-test results
0:22 Post-hoc-test results and respective effect sizes
0:34 Classifying results
If you have any questions or suggestions regarding the write-up for the Kruskal-Wallis-test in R, please use the comment function. Thumbs up or down to decide if you found the video helpful.
#useR #statorials
Support channel? 🙌🏼
===================
Paypal donation: [ Ссылка ]
Amazon affiliate link: [ Ссылка ]
