My study is aim to show the correlation between the Prostate MRI Scoring and Prostate Pathology Gleason Scores. For prostate MRI Scoring system (PIRADS) we score lesions according to their MRI finding 1 (lowest risk for cancer) to 5 (highest risk). So I think PIRADS is ordinal rather than categorical. The other variable is Gleason scores usually between 6-10 (most aggressive cancer). I think Gleason scores are also ordinal rather than categorical.

So for assessment of the correlation between PIRADS and Gleason scores, what do you think is the best statistics analysis method?

I thought I can do Spearman's correlation, is it okay?

  • 2
    $\begingroup$ I would say yes. However, I also suggest to expand a bit your analysis (linear or non-linear correlation), so check this out: stats.stackexchange.com/questions/8071/… In my view the first answer is super interesting, as it discusses how to chose between Spearman and Pearson correlation. $\endgroup$ – Fuca26 Feb 29 '16 at 8:09

You could try a polychoric correlation, which is for two ordinal variables.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.