I have done a bit of experimentation with this (but nowhere near as much as some other people here), and this is what I have found:
<60% (2 studies): substantial decrease of efg%, large leap in turnovers, small increase in fouls
60%-90% (4 studies): everything holds about normal, maybe small increase in turnovers. My results varied a lot, but not any significant increase/decrease
90%-120%(5 studies): small increase in efg% and smaller increase in turnovers
120%+ (1 study): large increase in efg% and turnovers. However, in smaller caps (sub-$50 mil) and draft leagues, getting to 120% means sacrificing on other things, mainly rebounds, so these teams generally get slaughtered on the boards.
These numbers are only for my starters, as it is difficult to find a bench that mirrors the qualities of my starting 5, even when platooning. There are a lot of other variables that go into this as well, such as usage, IRL efg%, fatigue levels, even predisposition to certain qualities, such as high/low turnovers or fouls