Is there an equivalent "Reduced chi-squared" test for homogeneity? I see that this is the case for goodness of fit, which helps when the expected values are below 1, but I haven't found if it also works as a test for homogeneity.
Asked
Active
Viewed 23 times
1
-
2What is your experimental design or data collection scheme? How many cells with low expected values? That may help to provide a relevant answer. – chl Nov 22 '20 at 19:34
-
Around 30% of all cells have low expected values – Kazh Nov 22 '20 at 21:17
-
And what is your data collection scheme? Because, in several cases, the [$\chi^2$ test does a good job](https://stats.stackexchange.com/a/4024/930) even with low EV. – chl Nov 23 '20 at 16:47
-
I'll further explain how the data is in the question, thanks for the link. – Kazh Nov 23 '20 at 16:48