I have a study where I'm looking at post-surgical limitations after shoulder surgery between two groups of baseball players. Patients were told to select all that apply among: Pain, Stiffness, Weakness, Decreased athletic performance, or other. Patients are divided as pitchers or non-pitchers.
What is the proper way to analyze the distribution of these limitations between pitchers vs. non-pitchers?
My current approach is using Cochran's q test, is this correct?
You need to define the response (outcome) variables clearly according to your need.
First, make the decision on multiple response variables or single response variable. The example of multiple response variables is treating pain, stiffness... separately. And use Fisher's exact test to compare the two groups on each response variable.
You can also combine multiple variable into single one. For example, you can count how many items the player checked. Suppose you have 5 items totally, then one play can get 0, 1, ..., 5 score. Then use Kruskal-Wallis test to compare the difference between groups.
If your Cochran's q test means https://www.medcalc.org/manual/cochranq.php, it is incorrect.
