Hacker News Reader

Comment by fdgsdfogijq | Hacker News Reader

PhD students are mostly cheap labor to carry out research for faculty. It's not a feeder to a faculty job. Most of them are immigrants who view it as a way to get into the USA, and so its a worthy trade off.

Its like in investment banking where analysts are hired to crunch numbers and make powerpoints. Banks hire analysts that have the proper training for those duties, knowing they don't have the potential to rise up the ranks and will be kicked out in a few years. Banking analysts/associates and Banking higher ups are then actually two separate careers.

Academia is basically the same

View on HN

This may be the case in the sciences and engineering, but is much less so in the humanities and the social sciences. They are very different situations.

It's different but the same, in the sense that your primary occupation is PhDing, you aren't paid well, and the most obvious pathway to reasonable employment is joining some faculty.

In the humanities, social sciences, and sometimes math, it's mostly teaching instead of research. The students are paid a bit worse (usually). The overall deal is a bit better because a TAship really can be a part-time job if you work efficiently and don't "care too much". Of course that also means the research work is unpaid. Whereas the type of research work done by most STEM PhDs is more like a 40 hr work week on top of occasional TAing, so the overall deal is worse but at least you're being paid for the thing that can get you a decent job afterwards. Different tradeoffs.

It's really the applied/clinical life/health sciences, education, and a few other "professional pathway" fields that work much differently. Most people getting those degrees aren't interested in TT faculty positions and are working a full-time or part-time "grown up" job outside of the university, so the power dynamics and expectations are quite a bit different.

Actually, post-docs are the cheap labor. PhD students are given a stipend and many/most have their tuition waived. The overhead of running a research department in the US is fairly high and leaves little room for high salaries[1]. Anything can be improved, and its easy to criticize from the outside.

[1] In the main-stream case, not withstanding the egregious greed that sometimes occurs at the management level. But no industry is immune from human nature.

Thanks for mentioning this as I was thinking the same. My SO is a PhD student and she often mentions her pay is greater than a post doc (or she would be taking a pay cut to pursue a post doc)

> PhD students are mostly cheap labor to carry out research for faculty.

This is true.

> Most of them are immigrants

This is false.

At least in the molecular biology field, most PhD students are not immigrants, while the majority of postdocs definitely are.

>> Most of them are immigrants

> This is false.

Doctorate means different things in different fields. In particular, getting a doctorate in Education or some of the applied Life Sciences fields looks more like a masters degree with a long paper at the end and you typically work full-time at a "real job". You or your employer may even pay your tuition. Ie, not the type of PhDs the article is talking about.

Digging into SED data, the "Citizen / Resident" numbers are propped up by two categories:

1. the applied biological sciences / health sciences, many of which do not resemble a traditional PhD program in any way.

2. education, which again functions very differently from a typical PhD program in every way: who pays, how they pay, what you're doing during and after, etc.

You can then track by field in the rest of the SED data and see that a huge number of these people aren't on the traditional academic track and are not doing a phd in the sense that most people in STEM mean; ie, they're working other full-time jobs, probably not publishing or barely publishing, never planning on going into academia afterwards, and so on.

More details: The NSF runs a survey of earned doctorates (SED) each year.

These reports include a table entitled "Research doctorate recipients, by detailed field of doctorate, citizenship status, ethnicity, and race: <year>". Let's look at 2021.

In Computer Science, 953/1545, or about 62%, were Temporary visa holders.

That's a lower bound! The other category includes permanent residents and the survey happens at the end of the PhD. Many PhD students receive permanent resident status during their PhD -- particularly those from places other than India and China, or who immigrated to the US prior to starting their PhD (eg for undergraduate or work). There are a variety of pathways, bu suffice it to say that many people end those 6 years no longer on a temporary visa.

Additionally, there are about 100 PhDs in 2021 that reported neither "citizen / permanent resident" or "temporary visa", who are almost certainly majority immigrant who don't (or at least don't think) they fall into either category.

All-in, I've seen estimates that up to 70% of earned doctorates in Computer Science are awarded to immigrants.

For all earned doctorates these numbers are closer to 33%, so flipped in the opposite direction. But if you start breaking down by field you will notice that the vast majority of fields where citizens are over-represented are exactly the fields where "PhD study as research assistant/TA" is less of thing and full-time employment during the PhD is much more common/feasible.

> But if you start breaking down by field you will notice that the vast majority of fields where citizens are over-represented are exactly the fields where "PhD study as research assistant/TA" is less of thing and full-time employment during the PhD is much more common/feasible.

This claim is very questionable. Your comment seems to completely ignore the humanities, for example, except for hand-waving away education.

tldr: I'm not ignoring the humanities; it does function like you describe. But ALL of the humanities combined were only 2.6K PhDs out of 52K total; Education is larger than all of the humanities combined, health sciences is about the same size as all of humanities combined, clinically-heavy life sciences are about the same size, CS is about the same size, etc. etc. The humanities are an outlier, and also small.

Why is it "hand-waving away education" but "look at the humanities" when there are nearly 600 more Ed PhDs than there are all of the humanities COMBINED in SED 21?

In Ed, 48% are primarily self-funded and 12.7% are primarily supported by an employer meaning a majority aren't "traditional" phd students on the path to a potential faculty position. A ton of the Ed students on fellowships are employer-supported by other means and even those with research assistantships or teaching assistantships are on same path (working full time in a central office or K12 school). The vast majority of Ed PhDs have no interest in higher ed eployment; they're getting their EdD/PhD because of the role it plays in public school compensation bands and promotion pathways. There are breadcrumbs that paint this picture throughout the various SED tables, but also, you can just... talk with people in the field.

Many of the applied/clinical life and health sciences function the same way albiet with slightly different funding mechanisms and candidate incentives. The point is, no interest in TT faculty jobs. Combined with humanities this is 2x-3x larger than humanities.

WRT non-classical phd setups and immigration, humanities are the outlier rather than the norm. And in any case small compared to STEM -- many individual STEM fields are nearly as large as all of the humanities combined, for example.

> But ALL of the humanities combined were only 2.6K PhDs out of 52K total

Ok, I just looked at the SED (Survey of Earned Doctorates) to see how they break it down, and I notice that "Psychology and social sciences" is huge. Larger than education, larger than math & CS, almost as large as engineering.

So when you combine psychological and social sciences, humanities, "non-science and engineering fields not elsewhere classified", and a few other fields, you've got a huge number of degrees that seem to be ignored in your discussion, and that would explain the conclusion: "For all earned doctorates these numbers are closer to 33%, so flipped in the opposite direction."

I also think you're stereotyping education and life sciences.

> Most of them are immigrants who view it as a way to get into the USA, and so its a worthy trade off.

Simply not true; I know of innumerable cases where European universities keep same-country academics on indefinite post-doc or other roles, refusing to ever grant tenure.

Most end up working in different EU unis and switching every 2-3 years, moving house regularly, all while having a horrible work-life balance.

The comparison to banking/consulting is bad for a few of reasons:

1. Analysts and consultants make very decent livings. Spending your mid 20s as a banking analyst/associate is a very nice life for someone in their 20s. PhD programs barely pay a living wage, and in large cities simply don't pay a living wage.

2. When deciding whether a lottery is worth playing, you are looking at both probabilities and outcomes. You covered the probability aspect well -- that's the similarity between banking and professoring. The massive difference between the two is in the outcome. Banking higher ups have won a lottery that is worth winning. Professors, on the other hand, max out just above FAANG entry-level wages. Even then, it's only after decade+ of top-decile performance.

3. You might counter #2 with things like "work-life balance", "lifestyle job", "job security", and so on. None of these are true anymore.

3a. Universities are run by middle managers who are mostly less competent and more petty than their private sector counter-parts.

3b. Moreover, tenure doesn't mean anything. At this point, it's nothing more than a gentleman's agreement with an MBA. I have literally zero doubt that over the next 20 years every single university in the US will go through a round of layoffs that target tenured faculty, and will do so without declaring financial exigency. This already happened at a lot of places during COVID, despite massive bailouts and even at places that increased enrollment without increasing spending ex-headcount. If you got into academic for the stability 30 years ago, your bet was pretty good. These days you'd have to be an idiot or have your head buried under a hundred feet of sand to make the same bet.

"Tenure" now is nothing more than "continued expectation of future employment"; i.e., it's the only "normal" job in an industrial sector (edu) where the majority of employment contracts are not only at-will but even come with with your dismissal date attached (post-docs and even phd students). Can you imagine working at a software company but they tell you up-front they will have to fire you in 3 years? And also only pay $55K? LOL. When push comes to shove, in most modern faculty constracts, there isn't that much of a substantive difference between "tenure" and "at-will".

4. The glide-path out of a STEM PhD is usually at least as lucrative than out of an unsuccessful analyst position, but the scars of poverty make it feel like an insanely lucky success. I've seen 32 year olds with Mathematics PhDs cry because they get a job doing nothing related to their PhD but that pays the same amount that their students are making at their first job out of undergraduate.

So, I think a PhD in a technical subject can still make sense, but there's no rational reason to be an R1 faculty member outside of a teaching track. Well, except one: "massive ego that can't play well enough with others to hack it in the private sector". And that's certainly who you'll find yourself surrounded by if you choose that route.

I agree banking is a good job and has good outcomes. So does a Phd at a good school in an in demand field.

But my point about PhD students and Tenured Faculty remains the same. They appear to be part of the same job ladder. If you were to view a venn diagram of the skills/background needed for those two roles, there is less overlap than initially appears.

Reply on news.ycombinator.com