I got a dataset with a value and class for each observation (see sccreenshot others). I want to compare group averages and get post-hoc statistics per class. So, compare the averages on all observations with class 1 to class 2, class 1 to class 3, and so on.

Is there a quick way to calculate F-statistic and ETA, comparing each class to all the others ? I tried pairwise_tukeyhsd from Statsmodels but it does not provide F-statistic or ETA. I tried to Pinguoin anova packages but they only provide one statistic for the whole dataset.

Apologies if my post is unclear, I'm quite new to this.

Thanks in advance

I got a dataset with a value and class for each observation (see sccreenshot others). I want to compare group averages and get post-hoc statistics per class. So, compare the averages on all observations with class 1 to class 2, class 1 to class 3, and so on.

Is there a quick way to calculate F-statistic and ETA, comparing each class to all the others ? I tried pairwise_tukeyhsd from Statsmodels but it does not provide F-statistic or ETA. I tried to Pinguoin anova packages but they only provide one statistic for the whole dataset.

Apologies if my post is unclear, I'm quite new to this.

Thanks in advance

• python
• anova
• tukey
• posthoc
• eta

1 Answer 1

I think I found it:

https://pingouin-stats./build/html/generated/pingouin.pairwise_tukey.html

This allows you to specify the effect size; one option is eta-square.