Clarification on Tukey-Kramer test in multcompare following the Friedman's Test
23 Ansichten (letzte 30 Tage)
Ältere Kommentare anzeigen
When using multcompare function, Tukey-Kramer is a default post-hoc test performed for the Friedman's test. Given that the most common post hoc test for the Friedman's test is the Nemenyi's test, I wanted to clarify whether:
- the Tukey-Kramer option is identical to the Nemenyi' test. The previous answer on MathWorks (see [1]) briefly says so, but I wanted to confirm if this is true, and, if so, whether anyone can touch the mathematical equivalence between the two.
It seems the Tukey-Kramer test and Tukey's HSD test is used interchangeably in the multcompare. Further clarification on this point would be helpful, too.
References:
0 Kommentare
Akzeptierte Antwort
Shivansh
am 18 Dez. 2023
Hi SungJun!
I understand that you want to know about the use of different post-hoc tests in the Friedman’s test with respect to the “multcompare” function in MATLAB.
For your doubts regarding the use of Tukey-Kramer and Tukey HSD test, they are indeed the same, despite being described under different names in various literatures. MATLAB's “multcompare” function provides the option under both names to accommodate this variability in terminology.
Regarding the application of the Tukey-Kramer option after Friedman's test, it is believed that this is equivalent to performing the Nemenyi’s test. There are many literatures where Nemenyi’s test matches the procedure followed in MATLAB.
I also looked at the book by Hochberg and Tamhane on multiple comparison procedures. It seems to be saying that the original Nemenyi idea, perhaps from 1963, was to use a critical value based on a chi-square distribution, but the use of a Tukey-Kramer value is more powerful. There may be a possibility that Nemenyi's name is still attached to the test nowadays even when the Tukey-Kramer value is used.
Hope it helps!
1 Kommentar
Weitere Antworten (0)
Siehe auch
Kategorien
Mehr zu Analysis of Variance and Covariance finden Sie in Help Center und File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!