Why not fix doctoral programs in length?

It's simple: cap the program at a fixed number of years (TC: five?) and let the market clear with whatever people have done in the meantime.  It's not fair to people who get sick but if that's the only cost maybe it's still worth doing.  (Is there a credible way to make exceptions?)  And instead of a dissertation require one good published article.

Anyway, that's the proposal in the new Louis Menand book, The Marketplace of Ideas: Reform and Resistance in the American University.

There is a behavioral argument for this policy — it is anti-procrastination – and a zero-sum status game argument for it, namely that if more people went on the market "unfinished" the stigma would lessen and everyone would save some time.  The overall rank ordering probably wouldn't be much different.

But are these people ready?  Menand has an effective zinger:

The argument that they need the [extra] training to teach the undergraduates is belied by the fact that they are already teaching undergraduates.

Overall his book is a stimulating read, whether or not you've spent more than five years in graduate school.


Many universities already do require completion in so may years; this is a common practice in Canada. Five to seven years is the usual limit.

There are also universities that permit submitting a collection peer-reviewed articles instead of a dissertation. One is no where near enough. A science doctorate should generate several papers. Three to six seems about average, and I've known candidates to produce twelve in four years. I'm less convinced of your second argument. With several publications, it's difficult to see a clear thesis or thought process. A thesis still serves a purpose.

Comments for this post are closed

ditto Bruce.

And there is a *big* difference between teaching a single, narrow course as a TA and running an entire class (which some, but not most, graduate students do).

Comments for this post are closed

In engineering, at least where I went, there's the expectation that your dissertation will be your major papers, e.g., Chapt 1 = Intro, 2 = common lit survey, 3 = paper 1, 4 = paper 2, 5 (opt) = paper 3, 6 = conclusions. But it's my understanding the publishing time for econ papers is much longer than for engineering papers.

Comments for this post are closed

In most life science disciplines it is 3-5 papers that are expected. They don't necessarily have to be published yet, but at least submitted to credible journals. I can't imagine, at least in these fields, a justification for writing "A Big Book".

Comments for this post are closed

In the UK PhDs are capped at 4 years. If the department as a whole doesn't have a good 4 year completion rate, that department won't get more PhD funding. So there is strong internal pressure to get the students to complete on time and not drop out. Students who drop out in year 1 don't count, so at the end of year 1, weak/unfocussed students are likely to be encouraged to leave. I don't know if that is the ideal system, but it works for most things. One big problem is no allowance for maternity leave, which is much more common in some fields (women's studies?) than others (physics?).

Comments for this post are closed

One of my favorite quotes is from Antonin Scalia in his dissent from McCain-Feingold to the effect that evenhandedness is not fairness.

I'm in a startup lab and requiring anything on par with anyone else is not fairness. There are already too few compensating rewards for joining a new advisor to make the costs equal to others with much greater rewards. I actually want to keep working after my requirements are met because I discovered that leaving with just a PhD is a huge disappointment.

A similar idea is to require a stipend for all students. It would be fine to pay them more but you have to pay them something. In my case this is the main motivation they have to get rid of me.

However, if you take 5 years of someone's life and then boot them out without regard to their situation I guarantee there will be some blowback.

Comments for this post are closed

Why not make the PhD length cap appropriate to the discipline, with a sub-goal to keep some of these folks off the streets?

Econ: 5 - 6 years
Engineering: 5 years
Math: 5 years
Business: 2 years
History: 5 years
English: 25 years
Sociology: 35 years

Oh, and with annual compensation cap, too: after 5 years if you are still in a PhD program you must pay the same tuition as a full-time undergrad with no grants or scholarships, no subsidy.

And I love this, although I'm sure I misunderstand it:
Some loosening of the job market process would be welcome.

Such as hiring more people even if they are not needed?

All this bloviating about PhD programs is going to get even more humorous in the coming decade as the number of high school seniors declines and the number of suckers willing to pay $35-$40k annually for a degree declines.

Someone should start the "Unofficial Problem College List" like CR's Problem Bank List, as many of those small colleges and "universities" you never heard of start to go under. Shrinking the market for PhDs even more.

Hey, how about an FDIC for colleges? Oh wait, even better, we have federal student loans, non-dischargeable in bankruptcy....

I guess all the perpetual students could go to law school.




Comments for this post are closed

anon suggests that Economics PhD takes 5-6 years. Business PhD 2 years. This is nonsense. Having been through an american-style phd program, I can say that my first 2.5 years were devoted to coursework done together with the econ students (econometrics (plenty), micro, etc.). We did not take macro courses, but had to take psychology and other courses that econ students typically do not take.

Comments for this post are closed

Is this the biggest problem? I'd prefer we work on un-bundling teaching and research first.

Comments for this post are closed

I think the variability between departments and disciplines would make it difficult to establish publication targets. Depending on field, 0-1 first author papers is considered normal. In others, 5+ is considered normal. Even between subfields it can vary widely: big machine physics students often publish zero papers while designing new diagnostics, whereas solid-state physics students can publish multiple times.

For students working with untenured faculty, there may be a drive to publish lots of papers. For students working with tenured faculty, I have heard of cases where the faculty member is so obsessed with high impact factor journals that s/he will combine the entirety of the doctoral research from multiple students for a single paper. My guess is that the academic freedom and styles of faculty will be cramped with a publication target.

As for teaching undergradautes, I will dispute CFG's point, since it is common for junior faculty whose only teaching experience is as a TA for a single course are routinely handed entire courses with no additional training. It is rare that additional time in the PhD program better prepares a person to teach undergraduates.

All that said, I would be in favor of capping the length of PhD programs, since beyond a certain point one transitions from "learning and training" to "cheap labor" (that's what postdocs are for, right?). As others have pointed out, many departments have such rules revolving around funding, although adherence to those rules is often lax. I also will agree that maternity (and paternity) extensions ought to be the norm. As the average age of the PhD student increases, it's important to recognize that some students will choose to have children while in school.

Comments for this post are closed

Four years is too little. Two years of classes plus a year or so on your first project and you've got to start writing. That cuts the independent research phase out of the Ph.D. entirely.

The UK system has time limits, but their Ph.D.s end up doing an extra postdoc anyway. I don't see the point.

Comments for this post are closed

Another useful reform would be to award degrees for contributions to the peer-reviewed literature regardless of training. I'm about to receive an advanced degree from a Dutch university on the basis of three peer-reviewed articles, and apparently they've been doing it for years. If you can get through publishing hell, surely you have qualifications that could be recognized by an advanced degree.

Comments for this post are closed

"you just have to put up with illness, personal dramas, and the exceptional difficulty of some work."

Or not. We could just come out and admit that academia selects against health problems, families, extra-curriculars, and stretch goals more than almost any other job. Put it in the mission statement:

"We seek people with no outside life distractions and an intense passion for a moderately (but not too) ambitious career."

More than anything, I can tell you that having a kid is just a bizarre situation. It is incredibly difficult but what is more I can sense they don't want to get rid of me explicitly for that reason (rather they don't want to be known for doing it). However, while I work harder than anyone here I get the impression they don't like the idea that I'm proving that it can be done (I agree, because it really can't be done by a 'typical' student). Maybe they need a PhD+K designation.

Btw, classes are BS. I got more help for my research by a single seminar of a visiting professor than all my classes combined. I highly recommend students trying to set up seminars. You don't know who to get, but at least you have the right criteria.

Comments for this post are closed

Ugh. We are dealing with pronouncements like this at our University where administrators with experience in one discipline take a one-size-fits-all approach and think that funding or time restrictions that work in one field should be applied across the graduate college. For example, in fields where postdocs are common the PhD is often not the "terminal" degree. In those fields students use graduate school to get a good postdoc where they can do the work that gets them a real job. In other fields where postdocs are uncommon PhDs frequently take longer because the training is often more comprehensive.

Ultimately any restrictions on the time to complete the degree must be designed to benefit the student. A good system should reduce student procrastination and should also have little tolerance for the use of students as sweatshop labor in exchange for a degree with few job prospects (English, at some Universities, I am looking at you). However, it is mistaken to believe that a degree in the 7+ year range is always bad or is always a sign of poor student performance.

In my own field, archaeology, the typical student arc includes classwork (~3 years) sufficient to write a good grant application that can get funded (1+ years) that leads to fieldwork (a few weeks to many months sometimes split over more than one year) that then leads to lab work (1-3 years) and dissertation writing. Of course during that time there is also publishing and teaching. A good student finishes in 7 years and shorter degrees are often signs that a student entered the PhD program with prior work experience or is in a subfield of archaeology that doesn't need fieldwork or one of the other steps. This time to completion is also typical of many discipline with a significant fieldwork component such as cultural anthropology, sociology, geology, etc. A serious proposal to shorten degree times in these fields must propose either how to shorten the research cycle of grants, fieldwork, lab work, and writing or how to offload some of the training into other outlets (postdocs, more intensive required undergraduate training, more internal grant money, etc.)

Comments for this post are closed

I'm in the Cognition & Perception Psychology Doctoral program at NYU and this is literally our program (5 years, dissertation == papers), except we're supposed to have 3 papers published.

It makes perfect sense to me.

Comments for this post are closed

Tying the completion of a doctoral degree to peer-reviewed articles published seems sensible but suffers from some major problems.

One of the major ones in my mind is, why would a university "outsource" the evaluation of all of its students to anonymous reviewers? If some graduate student manages to have published a manuscript based on research that his professors feel is full of errors and missteps, is it too bad for the professor, does the student graduate anyway? What about the converse? Shouldn't an advisor or committee who feels a student's brilliant work is not being accepted for publication because of unwarranted skepticism from anonymous reviewers be able to confer a doctorate to the student anyway? What exactly is the student paying $40k a year to the school for if not for the explicit endorsement of its faculty, anyway? (Or at least, why is a government grant billed $40k / yr in students' names if not for the guidance and evaluation of the faculty?)

It also doesn't make sense from the peer reveiwers' perspective. Sure, a professional sense of duty goes a long way to motivating people to maintain the quality of the scientific literature. But after a while, won't reviewers wonder why they have to read and evaluate all these students' papers for free, despite the fact that schools are earning $40k annually in tuition? Policies that explicitly link educational progress to articles published may just serve to erode anonymous, uncredited scientists' desire to help educate students for free.

Comments for this post are closed

"Basically, the implicit understanding was, you have projects that don't work, you get scooped, etc., then you didn't have enough different things going in parallel to start with."

This is actually what the extra time is for. The qualifier is supposed to get rid of the uncapable people. Others who really are just unlucky need extra time to get the PhD. The additional time invested and not having the papers is punishment enough.

Comments for this post are closed

Any limit would have to allow for exceptions. It is not
just maternity leave that can bollix up the works. A
major prof can die or leave, as can other members of one's
committee. A student may have to switch topics entirely
after several years, requiring a lot of catchup. There are
other sorts of contingencies that should be allowed for.

I am old fashioned on this, but I prefer to see at least
some sort of common theme in a Ph.D., even if what is coming
out is ultimately three papers to be published as stand alones
(along with an intro, lit review, and conclusions). I am
nauseated by the number of dissertations I see with titles
like "Three Essays on Microeconomics." Gag.

Comments for this post are closed

In economics, the problem is not so big - nearly everyone finishes in 5 years, with some at 6 and some at 4, with rare exceptions (say, Dave Donaldson, who took 8 though to some extent this was solely by choice; with his record, I'm sure he could have graduated earlier). Nobody in econ writes a proper thesis; three publishable articles are the norm. Obviously economists will have fewer papers than hard scientists since we generally do not credit research assistants as authors. There is no tradition of "first" and "second" authors in economics, so being listed as an author generally means you were the major contributor of the research.

Since econ programs are 2+ years of course work, and since the job market begins in November, graduating in four years means that the three dissertation articles need to be written in 14 months, while RAing or teaching. It's not surprising that this is uncommon. Five years gives 26 months; aside from fields of econ that require a lot of data collection, I don't see why students couldn't finish in this amount of time. US universities know this - many do not fund after year 5.

Finally, the true PhD is a bit longer, since today, at top programs, the majority of students hold masters degrees before beginning.

Comments for this post are closed

Why not admit that the whole PhD system is sick and corrupt, and get rid of it. Its real purpose seems to be to impress galley slaves for senior faculty. Lawyers need only 3 years of post-BA work to be licensed, and MDs a mere 4. Teaching English is not nearly as important as healing the sick or keeping the innocent out of jail, and a two or three year post-BA program leading to a Masters degree should be sufficient for that purpose.

Comments for this post are closed

Overall his book is a stimulating read, whether or not you've spent more than five years in graduate school.

Agreed -- and not just for the reasons observed in your post; I wrote about The Marketplace of Ideas at greater length here. For those of you debating on the time investment, note how short the book is: around 150 – 180 pages. But it packs a lot of material in it.

Comments for this post are closed

This would require changing how papers are published as well. Currently, with the endless rounds of revise and resubmit often lasting years, finishing in 5 years would be nearly impossible.

Comments for this post are closed

Comments for this post are closed