Summary: I think graduate school is largely a mistake for everyone involved.¹ Graduate students are poorly paid, low-skill but high potential, junior employees who need a lot more supervision and mentorship.

Problems:

1. Graduate students in many fields are thrust into a role where they need to decide on a research direction. Developing research taste is difficult and expecting graduate students to succeed at this is a great way to be disappointed.²
2. In the lab sciences, it seems less common to need to choose your own research direction. But, you still need to attach yourself to a particular lab.³ And the choice of lab will have huge ramifications for the value of your work and your career prospects. So, the situation is actually worse than in fields that require research taste, because you’re forced to make this single high consequence decision right at the beginning of grad school when you have the least knowledge and experience to make a good decision.
3. Very few other people have strong incentives to help graduate students succeed. It’s often necessary to be pushy and ask for help or feedback.⁴ By comparison, in many other jobs, your coworkers depend on your work and are constantly asking you about it. This lack of aligned incentives in grad school can also lead to habits that are counterproductive in future careers, like under-communicating, having an overly adversarial or transactional mindset, or sometimes solving problems by being dogged and pushy rather than just finding a more solvable problem or a work-around.
4. Students rarely have effective managers or mentors. Even a small amount of explicit management effort goes a long way.⁵ Effective managers and mentors are common in the private sector, and people with a bad manager can just switch jobs.⁶
5. Some first year graduate students are fully prepared for the independence and challenge of professional research and do a phenomenally good job. Graduate school is failing these students too because there’s almost no way to promote them quickly into a position of power and influence!⁷ Sometimes 24 year olds successfully start massively successful companies. It seems that some 24 year olds should also be well funded principal investigators.⁸
6. Students are paid an arbitrary rate completely unconnected to their market rate as researchers. ⁹ Some fields have high industry earning potential and students have two month summer internships that pay more than the student receives for the entire rest of the year. This filters for students that are already financially well off. Other fields have very little industry earning potential. This really filters for financially secure students.
7. This rate is also unconnected to how much their advisor/manager values them. Good students are rewarded by eventually getting tenure 15 years later. This is silly.
8. Graduate students are locked into this pay rate for 4-8 years, a time frame during which their skills have increased massively and they could have (easily) more than doubled their income in the private sector.
9. Graduate students are normally not able to move between labs. It’s not uncommon for someone to hop between two or three jobs over a few years before settling into one that turns out to be a good fit for a longer time frame. This is basically impossible in graduate school where switching advisors is difficult, much less switching universities!
10. When graduate students leave a lab, their projects often die because many graduate students are working on a solo project. This is a significant drag on research progress.
11. Graduate school is notoriously bad for mental health. This paper by Val Bolotnyy presents evidence that “24.8% experience moderate or severe symptoms of depression or anxiety - more than two times the population average.”¹⁰ This is almost surely exacerbated by the above issues. It’s hard to feel good about yourself when it seems like you’re working extremely hard on problems that no one else is really invested in.
12. There’s a lot of potential for abuse. Students often can’t leave a research team or university during grad school without completely leaving academia or switching fields. Students on visas are particularly vulnerable because they can’t leave their program without completely leaving the United States.¹¹

The thing that ameliorates some of these problems a little is that most graduate students are smart and hard-working. But, this also makes some of the issues regarding mental health and poor management worse because students regularly burn out. There’s a strong stigma against “quitting” that plays right into the diligent attitude of many grad students. By contrast, in the tech/software engineering world I’ve worked in since grad school, everyone is constantly encouraging friends in bad jobs to quit.¹²

There seem to be more non-university research teams popping up lately both in industrial contexts but also as non-profit independent research labs.¹³ I’d love to see these organizations hiring junior team members without PhDs as an alternative pathway into professional scientific research. Pay them competitively, train them and pair them with excellent managers. It might turn out to be an incredible investment.