Grad school for Americans is not a good investment, except maybe MBA, but that should belong to its own category. Often too short, less than 2 years, not really enough time for decent research, yet expensive, the chance of getting sponsorship is slim.
Grad school in US, thus is made for foreign students to get their foot on this country. It has a few attractions: It is short. Cost less than Bachelor Degree. It offers a chance to get familiar with the culture . Count as advanced degree so bigger chance to land a H1-B visa under condition you have a job offer. And most importantly, American Universities know this. They specifically design the programs to target foreigners, and made a shit load of money out of it.
If you look at it from a purely economics perspective, sure. But many grad programs are not that expensive. It's hard to overestimate how valuable it is to get the grad class experience (much smaller classes, direct interaction and discussion with world experts in their fields, and really interesting discussions/viewpoints due to those foreign students). There's no learning experience quite like it.
I'd say my experience at a top school in a STEM PhD was the opposite, classes were largely useless, often superficial and the priority was clearly to do independent research.
Those programs, with the mentioned quality would not be cheap. But that might be true for US citizens, because usually the school charges less from them. I think I missed somewhat the definition of the grad school here. Majority of the grad students come here for Masters not PhDs, and for the former, the stuff you mentioned, like smaller classes, interaction with professors cannot really be taken for granted, and from my anecdotal experience, it is not true.
PhD students don’t take ‘classes’ - they do their own independent research. It’s a research degree. If someone else can teach you a class on it then it’s clearly not cutting edge enough to do a PhD on!
You learn about what's come before by reading papers and reproducing experiments.
I can't imagine how a class for PhD students works - everyone has such focused research topics that how can you find something relevant to teach to more than one person at a time?
When I did my PhD on optimising dynamic programming languages I would guess there were maybe ten other students in the entire world doing something vaguely similar at the same time. How do you do a class for that in one institution?
Most PhD programs involve some classes at the beginning - especially programs which don't require a master's degree first. The classes are graduate-level progressions from similar undergraduate classes. Typically, PhD research goes on more slowly during the first year or two of the program while classes are taken, and the dissertation research topic is not completely fixed. For example, a PhD student planning to research programming languages would take the most advanced language classes available, plus math and optimization classes, for at least a year while doing preliminary research before choosing a dissertation subject.
I could see how CS/similar degrees would be that way. I'd assume most STEM are?
For political science in at least one program it was 30-36 credit hours for a master's degree. PhD was the same thing with an additional 18+ elective hours and another 30+ combined class/dissertation hours, and then the actual dissertation. Classes with few exceptions consisted of "workshops" and "seminars" over various topics (terrorism, multinational corporations, etc), which almost all involved weekly topical readings (5-7 journal articles and book selections) and weekly/frequent writings on those readings with a substantial paper for the final, sometimes with a midterm. Independent hours were also common with 1 on 1 work with one of the professors over a chosen paper topic.
As I said, discussion in class was expected, so you couldn't really wing it, and you had to be prepared or it would be obvious in front of your professor an another 5-20 students. I'm guessing this format is semi-common in the humanities/social sciences, it was for a linguistics program I looked into. Tons of reading and writing with the intent of guiding your towards a weird unexplored niche.
Of course PhD students take classes. Most universities make you do 36 credit hours before you can pass your qualification exams. They just don't have courses on specialized topics but there are plenty of relevant graduate level courses you can take in the department that end up being very useful for you.
Grad school in US, thus is made for foreign students to get their foot on this country. It has a few attractions: It is short. Cost less than Bachelor Degree. It offers a chance to get familiar with the culture . Count as advanced degree so bigger chance to land a H1-B visa under condition you have a job offer. And most importantly, American Universities know this. They specifically design the programs to target foreigners, and made a shit load of money out of it.