How to write a PhD thesis your committee will not approve
nextscientist.com
nextscientist.com
Instead, spend your time writing articles and getting them into journals. You should be able to produce roughly one published article a year [0]. When you have 3-4 published or near published articles, take them and put them together into a dissertation.
Most universities will allow this and it's the approach to developing a dissertation that minimizes your risk (you don't bet everything on one experiment, you minimize the risk of last minute critiques and requests for changes, etc...).
[0] This might seem intimidating when you start out, but if you want a promising future in academia you will eventually need to be producing 5, 10 or more articles a year (with admittedly a lot of collaboration and assistance from other). Setting a strict goal of one article a year as a grad student should be obtainable in most fields and get you on the right track. It will also force you to be disciplined and to produce. When of the biggest risk for grad students is trying to be too perfect or too "big" in their work.
Students don't realise it because they are so focused on getting through, but this opportunity doesn't come again. You are suggesting they should jump straight on to the slavish publication treadmill that academics already spend their lives on. In a pragmatic sense it's good advice, but it's also failing to make the most of the opportunity presented by a PhD.
Maybe I'm just idealistic.
It also discouraged true, equal collaboration because the student must own that work enough to claim it as their own. Personally, I feel I have much more freedom not having the dissertation hanging over me head.
No, that's what you do after you've got tenure.
Sure, but is this really in conflict with publishing articles? The sort of scientific advances that require a full book or dissertation for a proper treatment are few and far between.
Also: doing true, unconstrained, open ended research as a PhD student is taking a big risk, one that is (a) quite likely to prolong the time necessary for completing the PhD and (b) almost certainly a bad idea if not done with the full support of their PhD advisor.
Industry seems to be much easier to get crazy ideas done in.
The only problem comes with the papers that are not solely your own work. I got around this by putting in the preface what I had done in each chapter. It did meant that my thesis was a bit longer than it might have been as it had work that was not mine, but the examiners didn't seem to mind and it had the fortunate side effect of making my thesis a better doorstop.
https://llk.media.mit.edu/papers/ch-phd.pdf
You can tell the author was really in to what he was doing and expressing himself; it was not just a rush to graduate (my own dissertation is boring by comparison). Sure, going the "staple the papers/articles together" approach is safer and common at some popular schools (e.g. UCB), but is it really better for humanity and the advancement of science?
As a finished-my-phd-long-ago researcher, I personally pace myself at publishing one good article a year (I might work on others in collaboration, but nowhere near being the primary author). Producing just for the sake of producing is a big problem in academia, and seriously decreases the signal-to-noise ratio.
So don't be shy about getting very concrete and specific with your committee, including the kind of analysis you want to do, the theoretical framework you want to apply, etc. This is all easier in the sciences--humanities PhD programs are basically abuse factories (http://www.tjradcliffe.com/?p=1588)--but even in the sciences it is all too common to see students get misled by their supervisor and/or committee, who give guidance in one direction one month and a completely different direction the next.
One further piece of advice with regard to a clear thesis question: I've often found it useful to ask students "What's the title of your dissertation?" This just a more concrete way of asking about their thesis question, but it's much harder to fool yourself about. "Well, I"m investigating blah blah blah..." may look like an answer to the question "What's your thesis question?" but it frequently can't be reduced to "Experimental Investigation of Blah Using Blah Models and Blah Theories".
Finally, it isn't just your own perfectionism you have to watch out for. My iron rule for when a dissertation is finished is: the very first time your advisor tells you to change A to B in draft N, and then tells you to change B back to A in draft M > N, you are done. Your resistance to making further changes should go up factorially at that point, because your advisor has stopped paying attention or has become obsessed with minutia, and if given a free rein will continue to fiddle around indefinitely. I've seen students waste a year on this kind of nonsense, when they had a perfectly acceptable dissertation to begin with.
This all but eliminated feature creep, and I could then go back and say "you agreed that investigating this was sufficient." Then it was just a matter of doing what I said I would do, and be done.
It helped focus my attention, because anything that didn't address the thing in my proposal was more obviously tangential. It focused my entire course of research and kept me from spending too much time on side projects. But most importantly, it established in writing exactly what question I was going to answer and roughly how.
I didn't say "I'm going to prove that <foo> does <bar>." I said "I'm going to test to see if <foo> does <bar>." If I do the tests, and show the relationship between <foo> and <bar> I've done science. The goal is to create an experimental plan that tests a theory, not to build the best whatever possible. The job of a Ph.D. student is to demonstrate that they can carry out research, not to build the best widget. If you can, great, but that's not what a Ph.D. means.
I saw endless examples of people who waited until they had better data to put together a proposal. The problem with this approach is that it raises the bar by removing risk from the project. If you spend five years demonstrating the <foo> results in <bar> then your adviser will ask for you to spend another two years doing an actual project extending that result. If you clearly elucidate early on that it's not clear that <foo> results in <bar> then simply seeing if that's the case or not is often sufficient.
Formal proposals are like a contract between yourself and your committee establishing the scope of project that is sufficient to graduate. Doing this as early as possible can only help you, and in my experience any work prior to establishing the scope merely results in a larger scope later.
For better worse, my proposed project failed, and I was extremely lucky that a Plan B emerged at the right moment.
It was an ambitious experiment, i.e., it would have earned me genuine recognition within my specialty. I admit a high likelihood that the project was over my head.
Oddly enough a fellow student dreamed up a project for me that used much of my existing equipment and my equipment hacking skills. It produced one minor publication, but opened up a new experimental technique.
One of the things that makes being a good supervisor so hard is there is no one right way to do it as what each student needs is different. Some students need a lot of guidance and micromanagement, while others flourish when just pointed in the right general direction.
The one thing I would like to see is a limit on the number of students a supervisor can have at anyone time. Students need more than an hour a month with their supervisor. I would try and always spend an hour a day for my students (a lot of that time was group discussion), and at least an hour a week one-on-one time. I have to say I did this mostly for selfish reasons (I really enjoyed talking with my students about what they were doing), but I like to think they benefited from it too.
Discusses how to define a clear scientific question and maintain focus on it in the face of (sometimes subtle) forces that make both of these things difficult.
(Probably no-one will read your papers either, but you can at least list them on your CV.)
Building a career in academia is vacuous, it's not the only thing that can be done with a phd.
Instead, the first time your committee knows that you exist is when you plop your dissertation on their desk a few weeks before your defense. And they say, "what the hell is this?"
Perhaps if it were "forced" to occur regularly, it would be harder for students to spiral into the black hole. I think that the committee should actually be serving in a supervisory role.
In addition, the process evolved to make it extremely difficult to flunk your thesis defense. First of all, your advisor had to sign a form saying that he believed the thesis to be defensible. Then the committee had to sign the same form before a defense could be scheduled. The defense was expected to be practically a formality. For these reasons, a student flunking their defense would actually reflect extremely poorly on the advisor. Now the committee members might make it hot for the student, for instance demanding some additional work or analysis before allowing the defense to proceed.
In some disciplines, it may have served as a final check to make sure the student had actually done the work themselves. There were stories floating around of students who faked their thesis and were not caught until the defense stage.
I've lost around 6 months to these kind of shenanigans and ultimately my supervisor agreed that it could be part of my thesis so that it wasn't wasted time. Took a long time to get that acknowledgement though.
Ultimately a CASE studentship should give you an idea of what it's like to work in industry. This almost certainly means you'll do some non-PhD work, but if you feel that it's becoming a big distraction then you should speak out. It really depends on your supervisor and the company you work for. By contract you are obliged to spend 9 months there over the course your PhD (at least with mine). I tend to go over when I need to. I've helped out on things that are both directly PhD related, kinda-PhD related and stuff that isn't at all PhD related.
Talk to your supervisor first and let him/her know what's going on - at 2 months in they have better negotiating power and could be more subtle (e.g. "I notice X has been spending a lot of time away from the lab, is this work related to his PhD?"). Be honest with the company, tell them that you feel that it's a big commitment and you're struggling to keep up with your PhD work. They're not going to chastise you - if they do then you know you don't want to work for them in the future.
Pushing yourself to publish every year creates noise in the scientific literature and often leads to data cherry picking and massaging which is not just "questionable" as the article claims but completely invalidates everything a researcher does and makes him or her a fraud. The fact that it is tolerated makes it really hard to trust the research coming out of universities nowadays.
If you're doing this for science, then don't complain. If you want a career, there are better options than slaving for 8 years, 80 hours a week.
This is not sour grapes. I excel supremely at the "game". I just don't want to be a part of it anymore.
I wouldn't mind removing the PhD thesis all together and relying on articles instead. It's more agile to borrow from software engineering. Article, feedback, improve on your works beats working on a monograph as a general setup in my opinion.
It's not as rosy as this though, the journal/conference industry is pretty broken but in an ideal world with excellently peer reviewed open access everywhere breaking your research into smaller chunks than a PhD is better imho.
It should also be noted that even if you go the traditional PhD thesis route you'll usually publish articles as well and most good PhD programs will make sure you communicate your PhD work regularly. But I think there's too many cases of scientists basically locking themselves away to write the thesis and only turning in the final product, mostly talking to their advisor.
So I suggest getting smart people in one place, with only informal ties (including informal mentorship, rather than a God-like status of advisor). And then degree based on publications (whether in one institute, or another). PhD thesis is an outdated thing from times, where people didn't publish much and it was (often) their first publication. We should move one and update workflow not believe blind in old one only because it was there for more than a century.