One corollary that would be interesting to investigate with that dataset would be level of understanding about how things work between those two groups. My anecdata leads me to believe that more layers of abstractions discourage deep understanding (both because it's harder and there's less need) but I'd love to see that proven/disproven.