2
$\begingroup$

I am conducting a retrospective analysis on the effect of three different treatment option (intervention treatment) on a special type of cancer, so I have three groups: 1, 2, and 3.

In the baseline patient demographic table for the three groups, I have multiple variables like age, gender, presence of other disease, presence of lymph node involvement, etc. All of them are coded categorical variables (0,1), I want to know if there is a significant difference in each variable for the three groups.

Some articles use the Chi-Squared test and others use Kruskal-Wallis H test.

Which one is the correct one? Or if there is another possibility?

$\endgroup$
2
$\begingroup$

Kruskal-Wallis is used to compare data from each group that can be ranked. The variable of interest should be at least of an ordinal level. Common types of variable like age groups/categories, level of satisfaction in a 5-point scale, etc. are suitable candidates.

Chi-square is used to compare data from each group that cannot be ranked. Gender, race/ethnicity across groups are common examples.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.