Post-hoc-Tests after a Kruskal-Wallis-test will give you clarity between which groups differences in regard to the dependent variable exist.
➡️ Watch next: • Effect size for post-h...
You will do post-hoc-tests which are essentially pairwise comparisons of all the groups you have in your sample. However, the need to counter alpha-error inflation (and draw false positive conclusions) when testing multiple times on the same groups arises.
SPSS is using pairwise Dunn's tests. They are preferred over Mann-Whitney-Wilcoxon tests, because ties are considered, which is likely when you have an ordinal dependent variable and becomes even more likely the larger your sample becomes. Dunn's tests furthermore retain the overall ranking from the Kruskal-Wallis-test.
Afterwards I will interpret the results and utter some words of caution for a significant Kruskal-Wallis-test and non-significant post-hoc-tests as well as why not to use post-hoc power analysis.
📚 Literature:
===========
- Hoenig, J. M., & Heisey, D. M. (2001). The abuse of power: the pervasive fallacy of power calculations for data analysis. The American Statistician, 55(1), 19-24.
- Wasserstein, R. L., & Lazar, N. A. (2016). The ASA statement on p-values: context, process, and purpose. The American Statistician, 70(2), 129-133.
⏰ Timestamps:
==============
0:00 Introduction
0:11 Requirements
0:20 Post-hoc-test results
0:36 Interpreting post-hoc results
1:21 Words of caution about p value and power
If you have any questions or suggestions regarding Post-hoc tests for Kruskal-Wallis-test in SPSS, please use the comment function. Thumbs up or down to decide if you found the video helpful.
#useR #statorials
Негізгі бет Post hoc tests for the Kruskal-Wallis-test in SPSS
No video
Пікірлер