Ordinal variables, before and after evaluation, two different treatments... my nightmare

Hello everyone

  • I have two groups of patients. The first one is operated using the A technique and the second one is operated the B technique.
  • The way of evaluating their results is using an ordinal scale (Grade 0, Grade I, Grade II, and Grade III).
  • Patients are evaluated before the surgery and one year after it.
  • I want to know which is the best treatment (A or B).


  • I cannot just compare the postoperative results, since one group can have worse preoperative results than the other (if I just wanted to compare the postoperative results I guess that I should use the prop_trend_test())
  • I guess I should not use an ordinal variable into a continuous one and perform a t.test() on the difference.

I reckon this should be a very common issue when evaluating surveys. You have a certain situation where subjects evaluate something using a likert scale, then you apply two different treatments to them and then you re-evaluate the situation by the same likert scale.

Any help would be welcome.
Thank you.

One approach (data frame d is not populated)

# Ensure the treatment, pre_surgery, and post_surgery variables are factors
d <- data.frame(treatment = factor(treatment),
                pre_surgery  =  factor(pre_surgery, ordered = TRUE),
                post_surgery  =  factor(post_surgery, ordered = TRUE))

# Fit the cumulative link model
model <- clm(post_surgery ~ treatment + pre_surgery, data = d)

# Display the model summary

This topic was automatically closed 42 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.