does kruskal-wallis test on matlab take into account when a sample has a large number of ties or is their a specific function to use in such a situation?

I want to perform kruskal-wallis test on a set of data. however my data contains alot of repitation. usually when performing kruskal-wallis test a different formula is used when there is a fairly large number of ties, is there different functions for the two situations? I used

[p, table, stats] = kruskalwallis (test)

[Torsten]

I'd suggest you ask this in a statistics forum. But even then, you will have to explain more about your test data.

[Rawnaq Ababneh]

first of all thank you for your reply. can I ask how toposton the statistics forum since this is my first post to ask a question?

[Torsten]

I'd try this one:

https://stats.stackexchange.com/

But don't forget my advice to elaborate more on the kind of problem and data you have - my impression is the people there are more "elitist" than in the MATLAB forum, if you know what I mean.

[Rawnaq Ababneh]

yes for sure your advice will be taken. thank you so much :)

Answer 1

[the cyclist]

Looking inside the code for anova1 (which is called by kruskalwallis), I see that it uses a formula that is adjusted for ties. It does not seem to have different formulas for "large" number of ties vs. "small" number of ties, as you are asking -- but I would not expect it to.