I'm doing χ2 test for categorical data with more than two categories without problems. I would also like to know which specific categories differ significantly, though, and doing this gives me some NA p-values. As I already figured out (found here), this is because a certain category might not be present in a given pairwise comparison, leading to expected values of zero. Here is an example:

$$ \begin{array}{|r|r|r|r|} \hline & cat0 & cat1 & cat2\\ \hline clust2 & 0 & 8 & 2\\ clust3 & 9 & 1 & 0\\ clust4 & 4 & 8 & 0\\ \hline \end{array} $$

This works when using the whole table, but not when only comparing $clust3$ with $clust4$, since $cat2$ is not present then. Would it, in a case like this, be valid to exclude the missing category in order to be able to use chisq.test()?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.