tukey, 'tukey-kramer' or both in the function multcompare

5 views (last 30 days)
I need to correctly report the method I use for analysis. Can anyone tell what method Mathworks refers to when they use 'tukey-kramer' in the function 'multcompare'? I am confused because the help text on the option for 'ctype' for function 'multcompare' satisfies both tukey and tukey-kramer methods of correction for multiple comparison. My question is does 'multcompare' choose tukey for balanced anova and tukey-kramer for unbalanced anova? Or does it always choose tukey-kramer and therefore has a misleading help text as regards to the tukey method?

Answers (1)

Aditya
Aditya on 25 Mar 2025
Hi Bethel,
The `multcompare` function in MATLAB, when using the `'tukey-kramer'` option for the `'ctype'` parameter, implements the Tukey-Kramer method for multiple comparisons. This method is an extension of Tukey's Honestly Significant Difference (HSD) test and is specifically designed to handle unequal sample sizes (unbalanced designs).
The confusion often arises because both Tukey's HSD and the Tukey-Kramer method are similar, but they differ in their applicability:
- Tukey's HSD is typically used for balanced designs, where each group has the same number of observations.
- Tukey-Kramer is an adaptation that can be used for unbalanced designs, accommodating different group sizes.
In the context of MATLAB's `multcompare` function, it consistently applies the Tukey-Kramer method regardless of whether the design is balanced or unbalanced. This ensures that the method is robust and applicable to a wider range of scenarios, including those with unequal group sizes.
Therefore, if you are using the `'tukey-kramer'` option in `multcompare`, you are always utilizing the Tukey-Kramer method, and the help text is not misleading but rather inclusive of the scenarios where the method is applicable.
Please refer to the following documentation for further details:

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!