Methods help: How to find significant differences between groups with ordinal data?

[hrh2101]

In examining how students performed at the detail level of an assignment, I have received data that is essentially in the form of a checklist. Basically, for each detail item, a student will have a score of 0 (missed the item), 0.5 (received partial credit for the item), or 1 (received full credit for the item). The point is not to aggregate the data and compare the means of the students' overall scores, but to dig deeper into this detail to discover which groups performed better on which items. There are approximately 400 subjects distributed among three groups which are not equal in size.

Since the mean score between 0 and 1 is basically meaningless, the ANOVA/MANOVA methods we used on the aggregate scores don't seem to be the right approach, but those are the methods with which I'm most familiar when comparing groups. What methods are recommended for examining group differences when it comes to ordinal data?

Thanks in advance for any pointers!

[Biostat]

Hi

if your data is ordinal you can use Kruskal Wallis Test for comparing different group that measured based on the ordinal scale.

[hrh2101]

Thank you very much, Biostat! I will look for some methods literature on this test. Let me know if you have literature on the topic that you recommend.

[Biostat]

Hi

for doing this test by SPSS you can use the following link :
https://statistics.laerd.com/spss-tu...statistics.php

and for learn how this analysis works see below:

http://www.vassarstats.net/textbook/ch14a.html