Fun educational psychology experiment idea if anyone wants to run it by an IRB and get a paper published:
Get an entire compsci 1 class to take the RAADS-R. Then measure how easily/quickly they understood call/return stack semantics when the concept was first introduced (based on quiz/homework score, self reported "this makes sense" level, or something else).
And scatter plot the two against each other. Do you get two neat clusters?
Does scores from raads-r actually cluster? Based on nothing but intuition (probably driven by poor quality sources) I would expect no pronounced boundary there.
@azonenberg Ok, so you are looking for correlation (and not necessarily clustering)?
@robryk Potentially, yeah. It might be more of a line than two blobs.