Though the pool of students was biased towards those who excelled in high school which post 1980..1990 probably means some prior exposure (i.e., not your "total beginners"), this was a big part of the original idea behind SICP-6.001 at MIT. AFAIK, they gave up on that experiment 15 years ago and moved to Python: http://lambda-the-ultimate.org/node/3312
Being faculty / wordy, there is probably much written about it attempting to draw conclusions. There were definitely some "smart, but total beginner programmers" in the mix, although I don't know how scientifically separated the performance of the "two 'some's" might have been.
Relatively recently I found the SICP lecture series on YouTube from the early 80’s. I’d read most of the book already but it was so much better with Sussman and Abelson teaching it.
The audience for the series I’m thinking of is adult professionals- programmers I believe. And I’d heard they were from DEC or HP or some such. So in addition to the material itself, you have the blessed quirkiness of its two authors, alongside a whole lot of late 70s/early 80s attire and rudimentary graphics. Quite a feast… lol.
Being faculty / wordy, there is probably much written about it attempting to draw conclusions. There were definitely some "smart, but total beginner programmers" in the mix, although I don't know how scientifically separated the performance of the "two 'some's" might have been.