PhD is an apprenticeship to do research. If we left it until later in life, we wouldn't have anyone trained to do research.
Nonsense. Look at how many job openings there are for PhDs - people can't hire them fast enough. I was in a call the other day saying how hard it is to even find PhDs to interview, let alone get them to accept your offer.
It's not uncommon to see PhDs after five to seven years in industry in the C-suite. That is less common for MBAs, from my view.
For A, PhDs ought never be short of work. They were likely good students, they also spent 3-5 years looking at the state of the art in their field. Something should've rubbed off along the way. So a PhD is usually a good signal (but many smart people decide not to do PhDs, as well).
For B - the supply of places which will pay you to do actual research rather than company-specific goal stuff is much reduced compared to the supply of PhDs. This is why you'll find plenty of PhDs at Google doing software engineering - sure they're PhDs and they're smart, but they're likely not still publishing much. Whereas MSR/IBM research groups do publish (and in many ways act like a university).
For C - the chances of getting to tenure-track professorship are low, and when people say there aren't enough jobs for PhDs this is generally what they mean: there aren't enough academic jobs. The usual process is to become a postdoc and then to become junior faculty - a professor, but not tenured. Then you must seek promotion into a tenured role. Sometimes people skip steps, but this is the path people I've talked to have taken. Each stage is competitive: you have to convince a professor to take you on as a postdoc, then you must convince hiring committees you should become faculty and that you have the potential to become tenured. All the while, you must deliver enough research output to back that up.
At each of those stages, there's a relatively high attrition rate, and this is amongst people who obtained PhDs.
Now, there are hot topics that may see something around this amount of growth for a while, but many don’t. For example, in biology, genetics and biochemistry may somewhat have that growth rate, but if you’re in any other subfield such as ethology, forget it. If you’re a historian, did research on some ‘obscure’ language such as Middle English, or are working in some corner of mathematics that doesn’t have direct applications, you can forget it, too. If anything, the number of tenureships in those fields is shrinking.
Of course, many tenured professors have a lot more graduating students during their career. In medicine, one a year isn’t uncommon. That would require a field that grows by over 12% each year.
The deceptive thing about this is that, as a phD student, you don’t see those 30 competitors. You’ll only see maybe eight to ten. Some of them graduated a few years before you, others will start after you graduated.
Which means that having more PhD students means you can explore more of your own ideas by assigning them to it. It also means more teaching assistants and so on for your course. So the chances you will take and graduate one at a time are very low. Also, you know some fraction won't make it, so you need to account for that too.
I also sit on a graduate admissions committee, and it's hard to find good students.
We have work enough for easily 10x as many STEM students in the US.
> it's hard to find good students
Maybe because you provide such a bad deal for the majority of your students?
Can you explain this further? I don't follow how, for example, if I wait 10 years to begin a PhD there is somehow no research being conducted.
Even if everyone suddenly waited 10 years, there would only be a lag for the first generation.
The people I know who've gone for a PhD later in life were already doing research and generally took the path because the credential opened professional doors (e.g., to become a supervisor or director), not necessarily facilitated greater amounts of research.
To play devil's advocate, you could in theory increase the amount of research by opening avenues for non-traditional PhDs. There will still be those who start a program at 21, plus those starting at 40 who would normally not engage in a program.
If this is true, then you almost have to concede the model is broken. There will be a glut of PhDs compared to faculty positions. One of the reasons so many PhDs are disgruntled is because of the false expectation that every PhD is on track for a professorship.
I picked a random public school to see if their program states the intent is to produce faculty[1]. It seems like research is the other half of the goal along with teaching. Lots of research is done outside of faculty positions.
"The doctoral degree, i.e. the Ph.D., is primarily intended for students desiring a career in research and/or collegiate teaching."
[1]https://cse.engin.umich.edu/academics/graduate/phd-in-cse/
PhD training prepares people for the academic career and faculty positions. Less than half of PhD graduates actually end up in the academic career. So yes, you could say the system is ...not perfect.
Imagine if a trade school only had a 50% employment rate for their intended purpose and forced the students to work at a fraction of their market value for the privilege of attending, while profiting off that labor...I don't think imperfect would be the word used.
In 17th century Leiden, 40% of surgeons' apprentices dropped out. In Utrecht's Coopers' Guild, only about 10% of apprentices became Masters.
On the other hand, getting rid of the guilds in the 1820s and replacing it with less formalized vocational training during industrialization didn't change the system for better:
>A lengthy contribution in the journal of the Dutch Society to Advance Industry in 1891 criticised the condition of Dutch training in detail: ‘Nowadays boys leave one boss after the other for the greatest triviality . . .. The boss continuously fears that the boy will leave him to use his acquired skills with another boss. Bosses therefore slow down training by putting boys to work at specialised repetitive tasks, so that they bring in the highest profits.’
CS programs may be different, but engineering and physics were definitely geared for academics
I ask because all of the PhDs I've worked with outside of academia are from engineering and physics, save one from sociology and three from psychology. (The latter were geared to human factors of engineering problems).
You can’t convince me people are taking four years out of their professional lives at their prime to get a simple promotion?
Let's face it, PhD is a social signal as much as it is a vocational toolset. And it also signals a specific value system, like putting a high value on educational credentials. So if the selection board for a lab director is made of similar directors who have PhDs, you've already got a self-selecting pool of people who value PhDs, regardless if they provide much value in practice. Cognitive bias is strong and it would be tough for someone who's put in the work for a PhD to admit it's of little value, even if it were the case. Occasionally, people will get promoted who do not have a "proper" PhD, and they are (rightly or wrongly) derided for it behind closed doors, even if it has no bearing on their research competence. As a side note of irony, in many cases that I've worked with, getting a PhD makes someone do less research, not more. Because of the status signaling, they get more administrative leadership duties, not unlike a dean at a university.
And the PhDs that I know, who do it for a "simple" promotion, don't leave their job. (simple in quotes, because in my experience, they are rather large promotions).They complete their PhD as part of their job, in addition to their normal duties. It sounds like you are stuck in an academic mindset, where there is only one way to skin the PhD cat.
Well you’re mistaken because I also did my PhD after working for a few years, and while employed for an industrial lab.
And I recommend things because of my experience not despite it.
That may be where we're talking from two different angles. Much of this discussion seems predicated on the idea that you need a PhD to do research (i.e., it's vocational in nature). My claim is that you do not, but that it's possible people will pursue a PhD for other reasons (because it's more about signaling than many realize).
Yes, some people are naturally talented enough to do research without training, or with training other than a PhD.
But if you aren't exceptional, then promoting this is not realistic and is setting people up for disappointment. Some people can learn to play an instrument without a teacher... most would be better served with some structure and guidance for an experienced tutor.
I think the idea that a PhD program is the only adequate training for research is flawed. Most of the researchers (PhD and otherwise) have a fairly methodical approach to the endeavor. If there's a method/process, it can be taught. The non-PhDs that I know doing research did not have some innate talent for research, but were taught by other more senior researchers.
IMO, most research is borne from the combination of a solid grasp of a problem that needs to be solved, taking the time to build deep understanding of the topic, and then following a methodical approach to testing a hypothesis. None of that takes exceptional talent or half a decade in a PhD program to learn. (I say this as somebody who's been through a PhD program that I enjoyed. It's not from any bitterness). The biggest hurdle for many seems to be funding and resources, rather than some exceptional talent. And to that end, I'd concede that a PhD credential can open up doors for resources, but I hold that as distinct from the skills required from research.
If you're talking about the exceptional, once-in-a-generation breakthroughs, sure, I can see why a PhD program with an exceptional advisor may be necessary. The the vast, vast majority of research is small and incremental.
I had a high-school intern who worked with me and hit the top of Hacker News with a new compiler optimisation she invented.
I still think a PhD is the best for most people in most cases, if you want to be doing genuine research.
IMO, employment research offers many advantages. For one, it forces somebody to focus on the application and how it would genuinely be of value, so that one doesn't get bogged down into "neat" but ultimately impractical efforts that are little more than academic exercises. IMO, it opens the projects up to more interesting problems because they don't get cornered into the "publish or perish" paradigm. People easily get paid 3-4x (or more) what they would as a TA. They have daily access to experts in their field rather than relying on piecemeal meetings with their advisors. It also tends to open one up to larger resources for their investigation; for example, the people in PhD programs at my last job were given ample access to multi-million dollar rocket test stands and the anonymized data from them; if that were a university resource it would be hard to gain that same level of recurring access. Now I'll grant that there are certain PhD programs that are better suited to specific types of fundamental research that would otherwise be difficult to get funding while an employee. Nuclear physics would be one example. But I would venture that the majority of research is on the applied sciences/engineering side.