At SigCSE this year a session participant said, roughly, "research shows that there is no correlation between 'early skill' is CS and a student's ultimate success in the field." Unfortunately, I wasn't able to get a citation of the study or studies that came to this conclusion.
I teach high school CS and have a significant number of students for whom my class is their first significant exposure to "rigorous" CS (where "rigor" means, more or less, attempting to develop the foundational skills and confidence to go on in CS, rather than "just playing"). I also have some students in every class who are real "rock stars." I've been telling my students that they start where they start and their curiosity and persistence have a greater bearing on their success than their initial "acceleration" so I was thrilled to hear that there might be actual research that backs this up – especially if there is insight into how best to support the students who need time to build confidence and a foundation.
Have you heard of research along these lines? Any pointers on finding it?