How many hours per week should a graduate student work?

Here’s one of those things from Blogland that is flying around the Twittersphere today..

The original post revealed a leaked email  “sent to the entire graduate student body enrolled in the well-regarded astronomy program at Unnamed Academy” containing such gems as this:

We have received some questions about how many hours a graduate student is expected to work.  There is no easy answer, as what matters is your productivity, particularly in the form of good scientific papers.  However, if you informally canvass the faculty (those people for whose jobs you came here to train), most will tell you that they worked 80-100 hours/week in graduate school.  No one told us to work those hours, but we enjoyed what we were doing enough to want to do so.  We were almost always at the office, including at night and on weekends.

This missive has already provoked a number of responses (e.g. here and here), but I couldn’t resist putting in a few comments myself.

The first and most obvious thing is that I don’t think the faculty members mentioned above were telling the truth. It’s by no means a new phenomenon for oldies to pretend that they worked harder than the younger generation. “When I were a lad…”, etc. This is either  form of delusion that accompanies ageing or a kind of one-upmanship designed to create a impose some sort of authority over the junior members of the department.  A supervisor who demands such things of a PhD student is likely to be someone who regards a grad student simply as a form of cheap labour and doesn’t care at all about their development as a researcher or indeed as a human being.

The following sentence gives the game away

No one told us to work those hours, but we enjoyed what we were doing enough to want to do so.

It is clearly intended to mean No one told us, but we’re sure as hell telling you…“.

My advice to a young PhD student would be: if your supervisor tells you to put in 100 hours per week on the project, find another supervisor –  because he/she clearly hasn’t put sufficient thought into the practical feasibility of your project. The fact is if you have to work 100 hours per week to get your work done you must be exceptionally inefficient or working on a stupid project or simply nuts. Or all three.

The email is correct in saying that it’s “productivity” that counts. I’m sure there are many people who can sit at their desks for 11 hours a day without producing anything very much at all. It’s not the hours that matter, but what you do with them. In no way will indulging your outside interests (sporting, cultural, political, or “other”…),   or simply relaxing, detract from your ability to do research. I think such diversions actually improve your work, as well as (of course) your general well-being.

I had plenty of outside interests (including music, sport and nightlife)  and took time out regularly to indulge them. I didn’t – and still don’t – feel any guilt about doing that. I’m not a robot. And neither are you.

In fact, I can think of many times during my graduate studies when I was completely stuck on a problem – to the extent that it was seriously bothering me. On such occasions I learned to take a break. I often found that going for a walk, doing a crossword, or just trying to think about something else for a while, allowed me to return to the problem fresher and with new ideas. I think the brain gets into a rut if you try to make it work in one mode all the time.

But there is an element of truth in the paragraph quoted above. There were indeed many times during my time as a research student – and have been since – that I worked extremely long hours. I wouldn’t say exactly that was because I “enjoyed” it, but that I wanted to know the answer and couldn’t get the problem out of my head.  I’ve stayed up into the early hours of the morning trying to finish a crossword too. Not because I had to, but because I couldn’t put it down unfinished. I know that makes me a saddo in many minds, but I think that’s the sort of obsessiveness and tenacity a researcher needs: becoming so absorbed by the task in hand that you don’t notice the passage of time.

Anyway, as  a research student I certainly didn’t work 80-100 hours per week routinely, although I might have done a few times when things were getting interesting. I think an average working week of 40 hours is perfectly fine for a PhD student, as long as you use that time efficiently and are prepared to step up a gear when motivated to do so.

It’s been a while since I last had a poll, so let’s see if we can generate some statistics on this…

About these ads

43 Responses to “How many hours per week should a graduate student work?”

  1. Anton Garrett Says:

    Minimally, as many as is necessary to produce a decent thesis. Above that, as many as you want. Different quantitative answers for different people and different topics.

    The faculty members who were polled are, of course, the minority who went on to be faculty members…

    • If all you want is a decent thesis, sure. But suppose you want an academic job and the thesis is at least in part a means to an end. Let’s face it: 2 people are in the running for a job, one has twice as many refereed-journal publications or whatever the metric is, but also worked 100 hours a week instead of 50. Will the committee convert this to papers per hour? Fat chance.

      Yes, I agree that one should do other things, and also that working more does not necessarily lead to more productivity (however it is measured). However, there are cases where it does. Actually, it doesn’t matter if it does for you; it matters if it does for the other folks working 100 hours per week applying for the same job. :-(

      • One might wonder whether mentioning extra-curricular activities on the CV is a good thing, at least in part to offset this. I don’t think it is, since quantifying these is even more difficult than quantifying publications. And, as Stephen Jay Gould pointed out, tennis and skiing count more than softball and bowling.

      • Anton Garrett Says:

        “If all you want is a decent thesis, sure. But suppose you want an academic job”

        Yes, that’s why I said “as many as is necessary to produce a decent thesis. Above that, as many as you want.”

        Of course you can go the high risk route and declare that as your thesis problem you will seek to prove Goldbach’s conjecture or that there are no odd perfect numbers. If true, these will probably be provable in not more than 50 pages after having had one world-class idea.

      • Anton Garrett Says:

        Incidentally I recall banging my head against a wall for about 6 weeks trying to prove a mathematical conjecture in statistical physics for which there was strong numerical evidence. At that point I wondered if it might actually be false; I still believed it but thought it would be worth briefly investigating the opposite to prevent myself wasting more time if I was wrong. In one morning I reduced it to a Yes/No question which looked like it should have been treated by mathematicians, and an hour in the University Library got me the answer – the conjecture was in fact false, and failed in regions of parameter space well beyond those explored numerically. I was one of three who published the result independently and ‘simultaneously’, ie our submission dates were all before the earliest publication date.

      • Anton Garrett Says:

        “One might wonder whether mentioning extra-curricular activities on the CV is a good thing”

        I know that serious sportsmen were advised NOT to mention this on some applications to Cambridge. But in my opinion the really serious sportsmen export a discipline from their sport to their studies. The 2nd XV are a lot more likely to create havoc than the 1st XV.

      • telescoper Says:

        I get asked that about applications for PhD places. I think it helps if it’s a serious interest to which the applicant is clearly dedicated – for the reasons Anton mentions. But if it’s just very vague or seems like a passing fad then it’s counter-productive.

  2. I never found long hours by default were helpful. After a few hours of trying to solve a mathematics problem, it’s clear you’re doing something stupid and need to ‘reset’ your assumptions by doing something else – going for a walk, playing an instrument, whatever.

    Sometimes I would get obsessive and stay awake until the wee hours trying to solve a problem – but often those solutions were not he best (though they usually did pave the way for a better answer). But I never worked anything close to 80 hours a week.

    On the other hand, your research is always on your mind. If you include the amount of time spent puzzling on a problem while walking, doing chores, sat on a bus, staring into a pint in distraction instead of chatting to your friends, or trying to get to sleep at night – yeah, maybe that would add up to some pretty long hours.

    • Anton Garrett Says:

      Agreed. There were entire days when I did nothing and intended not to. but once I reached the stage of obsession with a theoretical physics problem I’d do nothing else for many days, 16 hours/day. At that point, whether you are driven or called is a fine line.

      Most of the friends I’d made as an undergraduate got jobs in London – quite a few were from the SE anyway – and it was a pleasure as a research student in Cambridge to visit them about fortnightly, spending the day seeing the tourist sights which as a Mancunian I’d not known before meeting up with them in the evening. After a few months of that I realised I’d seen more of London than most Londoners.

  3. Annoyingly I can’t find the paper but I’m reasonably sure that in many professions anything above a 40 hour week becomes detrimental.

    It’s not citing any primary research but this is a nice history of it: http://www.ehow.com/about_6692488_history-40_hour-work-week.html

    A long day to finish something (especially when you’re in the zone) can be very helpful but in general 8 hours/day (with a short break every hour or so) seems to best.

    • Bryn Jones Says:

      As PhD student I worked 55 hours per week. I found that working any more led to no greater productivity: I would become so tired that my efficiency declined so much that I accomplished less. Working more than 55 hours per week would also result in mistakes, necessitating repeating work.

      • yep – i worked ~60hrs a week when i was a grad student and a (young) postdoc – and that seemed “sufficient”.

        if you’re not immersed in your field then you’re unlikely to think of the new ideas… for that reason, i’d say that ~35hrs/week as a grad student (at least in astronomy) suggests that you’re not aiming for an academic career.

        working in research is just like being self-employed. if you’re not working (almost) every hour – then (as phillip says) the person who is will be the one who gets the job.

      • In other words, anyone who claims that one should work a “normal” number of hours should publicly state that he will inversely weight all selection criteria with the number of hours worked when considering job applicants. (But, let’s be honest, as long as the list of candidates is not public knowledge, with external referees who rank the candidates and provide justification for the ranking, and this also being public (this actually happens in some countries), what does this matter? Some people will continue to hire their mistresses and torpedo the applications of their competitors, the competitor non-mistress types resigned to the scrap-heap of academia.)

      • Bryn Jones Says:

        Hold on. Working in research is like being self-employed for those people who have control over their research direction, such as people on fellowships and academic staff. Working in research can be very, very different to being self-employed for many postdocs and many PhD students whose research objectives and methods are determined by others.

  4. 100 hours of work a week leaves only 6.8 hours per day for everything else including sleep. I don’t believe anyone who claims to regularly work such hours. If this email is real (it seems a bit strange that no-one seems to know which institute is the origin of it) then the signatories are certainly lying to themselves as well as their students. In my experience, graduate students more often need to be told to spend less time in the office than more time.

    • Bryn Jones Says:

      Yes, spot on. The claim in the missive that some people worked 100 hours per week is wrong: it is a lie. It deserves to be exposed as such.

    • “100 hours of work a week leaves only 6.8 hours per day for everything else including sleep.”

      No, it leaves (168-100)/7 = 68/7 = 9.83 hours.

    • Yes I was told as a grad student to stop working so much by the head of the research group and another academic who pointed out my efficiency must decline enormously after a certain amount of hours….he said he never did long hours even when young.

    • Anders Ehrberg Says:

      Ten day week? Otherwise is it 9.7 hours left.
      Anders Eg

  5. Anton Garrett Says:

    Time is an exceedingly valuable thing. There are not many places where, if you ask someone “What did you do today?” the reply “I spent the day thinking about my problem” is deemed satisfactory. Yet often that is exactly what needs to be done and academics must not be ashamed of explaining that fact, courteously but insistently, to their paymasters.

  6. Loretta Dunne Says:

    I would say enjoy being a grad student, it’s the time I enjoyed best before all the other bits of the job get in the way of being able to focus on research in that way. Incidentally I don’t remember feeling a huge pressure to write a zillion papers a year either, just doing stuff well seemed to be important. If all that matters is writing one paper more than the other person going for the same job, then we risk ending up devaluing what is IN those papers and how good people really are. Too high a workload and too ‘quantitive’ a target just drives people to find short-cuts to achieving some narrow metric, it doesn’t necessarily produce good scientists.

  7. telescoper Says:

    The word on the street is that the original email was sent around the Steward Observatory, University of Arizona.

    There remains a possibility, however, that it is a hoax of some sort…

  8. Andrew Liddle Says:

    I have to say that the original letter looks like a spoof to me. I don’t know any academics who would write in anything like that sort of style, and we all agree that the content is plainly ridiculous. And then there’s the whole unnamed institution fluff.

  9. Being efficiently productive and understanding what’s important (prioritising, being strategic) is the way to get ahead. Hours worked is so crude a measure that it’s not very useful. One of the most important things is someone’s passion for what they are doing. Instill passion in students/postdocs/everybody and there is a lot less worry about hours.

    I send this link to all my students and postdocs. I think it’s great advice.

    http://cosmicvariance.com/2008/04/18/influence/

  10. John Peacock Says:

    Leaving aside the advocacy of clearly excessive hours, the letter does contain some valuable points. The most important one is the emphasis on the need to read and think, and the concern that students don’t spend enough time doing so. This may be where some of the disputed hours could lie. I don’t want my students spending all their time just trying to get their code running – but when they go home in the evening, I would like to think that they might often choose to open a textbook rather than the latest TV soap DVD. This doesn’t mean reading something narrowly focused on their project – just as long as it’s intellectually stimulating. When we select PhD students, we try to pick those who have a passion for science, so that sort of thing ought to happen automatically; if it doesn’t, it may be that the old-timers are failing to produce the right work-place atmosphere, in which we place an emphasis on enjoying learning new things.

  11. Phil Uttley Says:

    I suppose it’s a given that the best way to get ahead in any walk of life is to be good at what you do? Of course it is not easy or even possible to pick ‘winners’ – every PhD student should be given encouragement because it is hard to say whether they have an aptitude or not at an early stage, some are late developers. But in the end I’d say that ability is a much greater factor than how many hours you work and it would be honest to at least state that. In fact, if someone really needed to work 80-100 hours per week to produce decent stuff, I would seriously worry about their ability!

    • “I suppose it’s a given that the best way to get ahead in any walk of life is to be good at what you do?”

      Indeed. Does one need anything else? I believe it was Cliff Stoll who gave the example of a list of rules for winning at chess. One of them was “always make the best move”. Isn’t that enough?

      However, the whole discussion is not worth very much as long as not all jobs are given out based on ability.

  12. Requiring anyone to work such hours is nothing to do with education or career development., It is straightforward abuse.

  13. [...] tone and content have caused quite a stir, with John Johnson, Julianne Dalcanton, Jason Wright, and Peter Coles all weighing in on the matter. Even though the gossip mill has more or less pinpointed exactly [...]

  14. [...] yesterday’s post about the trials and tribulations facing prospective PhD students, and an older post of mine about  the importance of not forgetting to live a life while you do a [...]

  15. [...] article caught my eye on twitter this week raising the question of  ’How many hours per week should [...]

  16. Malcolm Gladwell argues in Outliers that the 10,000 hour rule applies to mastering anything, from how the Beatles learnt their trade to how Chess grandmasters reach the top. It became clear I spent roughly 10,000 hours working during my PhD, as would many others I know who took differing lengths of time to submit.

    Therefore, the number of hours I advise people to work per week isn’t based on how much work they want to conduct, but rather how long they want to spend accumulating the 10,000 hours. I’m happy to believe that extra hours will alter the data, not the quality, of the submission, but have nothing other than a small number of anecdotal cases to back that up.

  17. What about academic staff themselves? When our university first started its “time allocation” quarterly task, along with the boxes for percentages on various tasks, staff were given a box in which they could declare how many hours per week they had worked. The average answer was 55. The powers that be decided this was an embarrassing result (the official answer is 37.5) so of course the box was removed.

    My instinct for either grad students or professors is that less than 40 means you think its just a job; more than 80 means you’re nuts. About 55 in fact sounds like passion balanced with a healthy life.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 3,688 other followers

%d bloggers like this: