Medical test probabilities draft (Patreon)
Content
Hey everyone,
Here is a draft for a video about medical test probabilities I plan to publish soon, let me know if you catch any errors or otherwise have feedback that can be incorporated.
I've talked a bit about plans I had for this topic in previous posts, and this video is one I sort of split out from an earlier draft. The target here is a viewer who has either not come across Bayes rule yet, or who has a little but not enough to have become familiar with how it can be framed using odds. It seemed worth trying to pair this with the question of why medical test statistics can be prone to the kinds of fallacies they are.
I still have some open questions about this particular presentation. For example, when you start applying statistics to models that have some kind of continuous parameter to them, rather than just being a binary "you have it or you don't" kind of question, there is a distinction people draw between what Bayes factors are and what likelihood ratios are. Is that worth going into at all?
If you have any thoughts, either on this draft or what you would like to see in any follow-ups, please do feel free to share.
-Grant
Edit: I've swapped the video linked above for the final one now. Thank you for the helpful comments, I rearranged/shortened the intro, swapped around the order of some other parts, and added a note for more info at the end.
Files
Previews only