How to "Frame Data" and what "Test statistics" to use?

Dear talkStats community,

This is my first time posting.

I would like some guidance in both the framing of my data and what test statistics to use.

The raw data consists of radiographic images, whether or not a specific image has any one or more of a set of 5 "signs", and a particular health outcome.

The study is retrospective. There are some confounding factors with the population, (e.g., interventions, medications, and environment), however I would like to just discuss those, well, in the discussion.

I would like to know how to frame data & test whether, any particular "sign" or combination of "signs" is able to "predict" or "correlate with" a particular health outcome.

Any and all good comments are welcome & appreciated. If you don't have anything nice & useful to say, then please refrain. Alternatively, at least try to let me know if you just have "cuttingly sarcastic" humor because I'm pretty sensitive. :cool:


Last edited: