I have a data set with two groups of observations. Group X scored an observation on one ordinal scale (A, B, C, D ,E or F) and group Y on another (A, B+, B, C, D ,E or F) ... note the B+
It is my hypothesis, that there is no difference in dispersion/variation between the two groups ... but how do I test this?
The data is ordinal (ordered, but no equidistant).
Arguably, scores of B+ in group Y should correspond to either A or B scores in group X.
To test the difference in median between the two groups, you can use the Mann-Whitney U test.
The difference in dispersion can be tested with the Ansari Bradley test.
Ansari Bradley test ... never heard of it :-)
I will go google right away - thanks a lot for your help.
That will do it!
An excellent, short, concise and very useful reply -- thank you very much Koenraad Van Meerbeek
- Badges
- Science topic