Posted By:
Rich H.Glenn,
Thanks for looking at this statistically. I have taken a little bit of interest in this particular question, and really appreciate your input.
I don't profess to be a statistician, but was wondering if I could pose several questions to you about the analysis. Chi square analysis is usually for non-parametric data (let me know if I am incorrect). Could the data be analyzed via a regression test (assuming that the psa or sgc "grades" can be considered parametric)? Then, 2 groups could be analyzed for statistical significance (i.e. 350 vs 150). Would the student T-test be appropriate here? I won't be insulted for you to tell me that this would be totally BS in this scenario...
Or if the data cannot be considered parametric, could the Mann-Whitney U be utilized?
Rich