What statistical analysis is appropriate for four groups for which students are to rank from 1 to 5 level of difficulty?
There are five subjects that have five levels of difficulty set from 1 to 5. Each student picks the level of difficulty for each of the five subjects. Looks like a 5x5 table. Not too sure what statistical analysis is most ideal and what relationship to look out for in determining statistical difference in the subjects difficulty for the students. All students offer all the subjects.
Hi Olutoyin.
One way to go about it is to convert the subject difficulty to numerical scale. One summation formula to use is: (Actual score – possible lowest score) / (lowest possible score range) * 5.
Using the formula above, Level 1 difficulty = 0; Level 2 = 1.25; Level 3 = 2.5; Level 4=3.75; and Level 5=0.
The subjects can remain as categories: 1--Cology; 2--Cognosy; 3--Xmistry; 4--Ceutics; 5--Clinical.
You can then determine mean or median (depending on distribution) of each subject with the selected scores, to see their average difficulty level. To do association test, use one-way ANOVA or Kruskal Wallis (again, depending on distribution).
Good luck.
Performance of the students on the five courses are the data to be analyzed. The null hypothesis is that the courses are equally difficult. This will be tested using ANOVA. If the test is significant then a post-hoc analysis to get the ranking of the courses in terms of difficulty is done by Duncan's multiple range test, for example.
Thanks Alhaji. The difficulty levels are already determied as 1 to 5. What is the reason for the re ranking? See the attached data file? I believe since ranking is already specified it may not meet parametric test requirements, most likely.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis