I am comparing scores on a language proficiency exam (ordinal scale) in Modern Standard Arabic (MSA) before and after a major curricular change. The curricular change involves integrating the study of the Levantine dialect of Arabic into a previously MSA-only curriculum. I am using Wilcoxon and Mann-Whitney to compare scores before and after the change. My research question is whether or not the curricular change affected student performance on the language proficiency exam. I will also compare the proficiency gains from Beginning to Intermediate Arabic courses on the same exam before and after the change. Any suggestions on what other analyses I might run to investigate this further?
It seems like you want to see if there is a significant difference between pre and post test scores. Wilcoxon is a rank test but your data seem to be from normal distribution. If this is the case than paired Student's t- test would be suffiecient
Thanks, Peter! The data is actually non-normally distributed and I'm using Wilcoxon because it's the standard for comparing ordinal ratings on the ACTFL scale which the exam produces. Nonetheless, I've also run t-tests in addition to Wilcoxon and MW and the results are similar. Just looking for something new or different that I might not have thought of... :)
Hi Jeff!
I think that perhaps you would be able to check if there is any relationship between the two perfomances (on tests), if there is any correlation between pre-change and post-change results (with Spearman).
Hope it helps!
Good luck!
HI Jeff,
Have you tried to transform your data using lognormal transformation or other to see if is becomes normalized?
- Badges
- Science topic
- Similar topics
- Linguistics
- Philology
- Language