If I want to analyze the correlation between certain biomarker (continuous variable) and the clinical parameters of patients such as age (grouped into < 50 yrs & > 50 yrs groups), TNM stage (include 3 groups, Stage II, III & IV) and PR status (grouped into -ve & +ve PR groups), which statistical test could be used? given that the data are nonparametric
There are a number of problems here, like:
1) Data are not "nonparametric".
2) grouping age into categories is never good.
3) It's not clear what statistical hypothesis you want to test.
4) If you want to analyze the performance of a biomarker, you should use the biomarker values to predict the outcome (not sure what it is in your case). The interesting things to look at are the PPV and NPP and their improvement by including your biomarker into the prediction model.
I strongly suggest to sit down with a local statistician (one with some experience in such biomarker discovery studies) and discuss your project.
Jochen Wilhelm I want to determine whether there is a significant difference in biomarker level among the 3 groups of TNM stages or the 2 groups of age or PR status
Regarding the second point about grouping age into categories, I have noticed several papers did the same (grouping age).
I agree with Jochen Wilhelm . I am attaching a recent paper of ours that you may find helpful. Best wishes, David Booth PS Somebody else did it doesn't mean "it" is a good idea.
- Badges
- Science topic