Hi all! I'm a young researcher (doing my master's) and I'm trying to get a hold of AUCs in ROC curves for my research protocol. I'll be using a software, but i'd like to understand better!
Briefly, I'm validating two depression screening scales against the golden standard of depression diagnosis. So far, its clear to me how to generate the ROC curve however, I'd like to calculate the AUC for both scales. All patients (n=184) will receive (or not) a depression diagnosis by a physician and will also be evaluated with both screening scales separately and blinded by nurses. 1 of the screening scale is nominal, binary and the other is categorical, rank order (however a cut-off of >10 indicates depression). I've been reading multiple articles/blogs to understand/locate the different methods for generating an AUC and I can't pinpoint which method is better to use and how we should prioritize one over the other. I haven't found any guidelines on this with the assumptions needed for the different methods. How do I chose which method applies best when calculating AUCs? **If more info needed, ask in comments! Thank you all!