In a recent question, an OP asked
Do I have to take courses before I can start doing research despite having a master and highly relevant skills for the topic in question?
Generally speaking German students embarking on a PhD are better prepared than many of their US counterparts. So, even if most US students concentrate on coursework and exam preparation during the first two to three years of the PhD, postponing diving into research, that doesn't mean that you have to do things that way. If the coursework and exam preparation don't take up all your time (on top of possible teaching assistantship duties, typically 20 hours per week), and you get connected early on with an advisor who won't hold you back, there's no reason why you couldn't dive right into research from the start.
In the ensuing discussion someone wrote
For schools that offer TAships, the TA load varies enormously by school and discipline, but 20 hours per week is at the high end.
and someone else added
Although students often unofficially work more than 20 hours per week, most universities set an official upper limit of 20 hours per week, and that's also the limit imposed by visas for foreign students.
So now I'm curious whether 20 hours per week is actually at the high end as the commenter claimed, and what the actual average is. But someone tried to ask about this not too long ago and the question was closed because it was a polling question.
I need some help formulating an acceptable question. Here's a starting point:
Reference request: I'd like to find a study that sought to find the average number of hours worked per week by grad student teaching assistants in the US.