I have data from an ordinal scale, with the score ranging from 1-4. Each subject has 9 assessments on this scale (study is a 9-way crossover design), and there are no missing data. What is the proper statistical analysis for ordinal data that also can account for the correlation within subjects? SAS is the preferred software for analysis, by the way.
Asked
Active
Viewed 1,549 times
1 Answers
2
I'm guessing SAS might have a multilevel ordinal modelling facility. You'd just specify the subject variable as a random effect. If it's not available in SAS you could try the ordinal package in R and the clmm function.
John
- 21,167
- 9
- 48
- 84
-
Further question- Would proc genmod or glimmix work? – Bosley Apr 30 '13 at 13:45
-
GENMOD with the repeated option sounds good. – John Apr 30 '13 at 13:57