Thursday, December 04, 2008

Too Much Like a Thesis

The common practice of having a doctoral thesis in the sciences be comprised of manuscripts that have been published or submitted (or that are about to be submitted) increases the chance that the advisor and others will participate in a major way in the writing of the papers/thesis. I touched on this yesterday, but want to discuss this specific issue more directly today.

I have found it to be a not-good use of time for a student to write a thesis and then turn the thesis chapters into papers. It is far more efficient (time = a*grant$^2) to go straight to manuscripts and add any extra thesisy stuff in Appendices or ancillary chapters to the thesis.

Furthermore, as a reviewer and editor, I have seen many a manuscript that was 'too thesisy' and needed significant revision. This commonly happens when a thesis chapter is transformed into a manuscript but is not transformed enough and retains too many thesisy elements.

The review comment "This reads too much like a thesis" is a negative one. The comment typically refers to the fact that the details of the study -- or the background material of the research -- are explained in excessive detail and at a more elementary level than what is appropriate for a journal article.

I have also seen this comment applied unfairly to a student/author. Just because an author is a current or recent student doesn't mean that their writing is automatically too thesisy.

A journal article should not be thesisy, but a thesis should be -- that is, a thesis should contain detailed information. The thesis is an archive of the work that was done, and may contain all sorts of information that should be documented somewhere, if not in a published paper. Some of the detailed explanation parts of the text, however, are not so useful even in a thesis; e.g. if a student spends pages explaining some basic background information that could easily be summarized in a few sentences with a few key citations.

I prefer the papers-as-chapters mode of thesis construction because it helps both the students and me, but it's not a perfect system, as it works best if the student can and does write without too much assistance and if the other co-authors (including me) provide timely but not intrusive comments (a topic for a future post).


Anonymous said...

I agree that papers -> thesis is the way to do things. What I am not keen on are theses that are just collections of papers, with very little additional information or depth. This seems to be the case in the Dutch system, where I am currently working, and it has been obvious during the defences that I have attended that the students sometimes lack basic knowledge that (a) should be in the thesis and (b) should be essential to pass.

Anonymous said...

I rather object to the thesis-as-chapters method that is becoming more popular in my field in Europe.

I am involved in a number of cases in which journal articles have been more or less blatant plagiarisms - or were published in conference proceedings for dubious conferences (dubious = location such as Orlando + pay your fee in advance + no reviews actually sent out + not many people physically attend the conference). One conference (they usually have "International" in the title somewhere) I recently saw was a "Virtual" conference sponsored by a "University" wholly owned by a Far Eastern religious sect.

You pay your fee and get your publication. You collect publications and make a thesis out of this.

Is this science?

I realize that there are good mentors out there who have their people avoid conferences like this. And who make sure that their people are learning how to write. I myself end up having to conduct scientific writing workshops and scientific talk-giving workshops for my people (I have also made a course out of this so it counts for my teaching load).

I also think that the mentor needs to butt out of the writing. If it is a thesis = an examination artefact, it needs to be by the student alone. If it is cooperative research, then both names are on it and both do the writing, but this is not an examination arteface.

My two Eurocents :)

Ms.PhD said...

Yeah, but. I actually think more papers should be more thesisy in the sense that many papers these days are overly brief on the critical information.

Once upon a time, it was considered necessary to provide enough information to actually be able to independently reproduce the results from the publication alone.

This isn't really doable from most papers, but it is still (mostly) doable from most theses.

So I'd vote for information sharing, not paring down.

Having said that, I think your approach of "papers first, thesis later" actually makes sense and is the most common approach. Unfortunately in the sciences a thesis is no longer considered an official publication.

But my graduate program didn't tell me that "staple your papers together" was their unwritten official policy, so I wasted a lot of time trying to write a Thesis only to have them tell me to take out everything that wasn't already peer-reviewed.

Uh, thanks guys. Wouldn't want to make you, uh, evaluate what I did or anything like that!

Luckily I had already printed and bound a copy of what I affectionately refer to as The Long Version to keep for myself. Thanks, Kinkos.

Anonymous said...

This has been interesting to read, as I am a grad student who is planning on using the papers-as-chapters mode of writing my thesis. Given the nature of my field and project, my advisor will be a coauthor on most (if not all) publications resulting from my thesis. I'm also planning on writing up these chapter-articles and getting them submitted before I even defend the thesis (as I want publications when I go on the job market). In the end, I'm writing all the drafts and doing a majority of the writing. Would it really be better for me to write a regular thesis and then go convert it to publications...and only then make any changes suggested by coauthors?

By the way, with regards to your discussion on collaborators, I'd love to hear your opinion of dealing with problematic/unpleasant/even toxic collaborators...particularly when you're stuck with them and can;t get out of it (ex: they are your advisor's collaborator).

Unknown said...

As someone who wrote a thesis-y dissertation, I have to second your comments.

Yes, my advisor was significantly involved in the writing - it took basically a year (of doing not much else) to go through all the writing and editing. I'm happy with the result. I was working with a very well known scientist, and still do. But.

I have but one publication from my dissertation (the project was too cohesive to break into more than one manuscript, anyway). But compare that to the "three papers stapled together" manuscript that a lot of my peers did... even with a paper from an outside project, I'm still one publication behind them on the job market.

It's all about publications. Get more of them, you'll look better in the end.

chall said...

When I wrote my thesis (in Sweden) the most common way to do that (in biological sciences/chemistry) is to take the three-five papers where you are first author on at least 2, and then write a "summary" of about 40-50 pages with background and "wheere does my research fit in the grand scheme of things". The summary is placed before the papres and they are all bound togeter in a little book. That way you can have manuscripts in there but still send them off without too much trouuble.

I like this approach, go figure, since it is clear that your theories are backed by the papers, which you can explain in more broader terms and applicaitons rather than "just the publication in itself" that can be - as you wrote - a bit more detailed.

I have seen other types of thesises and I can see the happy of writing a thesis with chapters of papers. I do, however, have a hard time seeing that a theisi can be sent in and published as a paper. (I would think one thesis is too little for one paper but maybe not if it is a in a N/S/C journal of course...)

Anonymous said...

I had this problem. We went with the approach that they should be papers and went back afterwards and had to rewrite large portions of the chapter for the paper. I'm in my post-doc and two papers are still sitting on my PIs desk.

Kate said...

It is because my advisor insisted I write a thesis rather than papers -> thesis that I am still trying to publish elements of my dissertation in my first year of a t-t position. It is a massive pain in the ass, and I should be publishing other stuff by now, but I'm having to drag around my committee co-authors and try to get them to work with me on these pubs. Grr.

Kate said...

Also, I agree totally with MsPhD

Anonymous said...

I wrote in yesterday and said I'll be pulling out chapters of my thesis to make into papers, but my first chapter is already published as a paper so; I guess I'm doing it a little bit both ways. I would of liked to have all my "chapters" published as individual papers before the deadline but I don't think I'll be done with the last chapter in time for that.

aceon said...

My thesis, (which I defend tomorrow!) has a little bit of everything. One chapter is a published paper. One will never see the light of day anywhere else. One is pretty much a manuscript ready for submission, and the last one will be broken up into about three papers for publication. I am glad to have the flexibility to include chapters with varying character. In a perfect world, I would opt for more on the published paper side, but it is good to have the opportunity to delve into some details.

Anonymous said...

In the order they were written:

1) Chapter 3-The best of the projects (best because it was 100% mine and proved that I can come up with my own projects, although it certainly didn't rock the world). I started it before the other projects were finished, wrote it up as a long paper, stapled into the thesis, continued more work and polishing on this paper after the thesis defense, and didn't even submit it until after I'd already gotten another paper published as a postdoc. What can I say? Of all my projects, it's the one that I loved the most, because it was the first project that was completely, 100%, indisputably mine. So of course I lavished lots of love and attention on it. It's my baby, and it needed to be perfect.

2) Chapter 4: Written as another long paper, this one on a collaborative project. Stapled into the thesis, with some additional material in an appendix. Submitted shortly after the thesis defense.

3) Chapter 2: My first project. It never really went anywhere. There was a spinoff that I collaborated on with somebody else (wound up being second author on that paper) but most of the results on that project were largely redundant with the rest of the literature. Still, there were a few interesting details, so I wrote a thesis chapter on them, but never submitted a paper.

1) Chapter 1: Introduction, in which I explain the different phenomena that can emerge from light scattering and multiple scattering of light, how these different regimes of behavior are linked to one another, and what parameters differentiate them. Basically, my take on Light, The Random Walk, and Everything.

So, two very thesis-y chapters, and two chapters that were basically papers stapled together. Balanced.

My advisor never had to write any of it, but he did have to kick my butt to get me past the writer's block and onto the task. I thought for sure that the writer's block would go away once I had a postdoc offer, but even then the procrastination was in full gear. But once I got on task, I churned them out in caffeine-fueled bursts.

As to the order in which they were written, isn't there a Ph.D. comic that shows a timeline of grad school, and the time intervals corresponding to each chapter of the thesis?

Sablivious said...

I think papers and theses are different modes of writing and conveying ideas. Multiple authors contribute to and write a paper. A thesis and a paper is distinct and since most people only get to write one thesis, stressful as it is at the time, it is an accomplishment that should be distinct from a multi-author paper. Just an idea, I'm not sure time-wise if it's feasible.

muddled postdoc said...

Most of the people in our department seem to follow the intro - review - conclusions sandwiching the rest of the chapters usually published as papers or in the process of writing to be published. Since our uni imposes a maximum number of years of candidature (4 for scholarship) it saves people the time of having to write all the chapters from scratch at the end in a rush.
I would like to see papers have more background too (i.e. more thesisy) but some of the good publications in my area that most of can hope to achieve are 3 page letters which leaves very little room for background.
One question about theses though, if you have say a title and all your work fits cohesively into it, but you have some other work which (like in my case) is done not at all related to your thesis (but you did it because you had a lull period) and managed to get a good publication or two out of it should it go into the thesis somewhere? It doesn't seem to make sense to include it as a chapter at all but how would you make the examination committee aware of other good work that was done during the time you were doing your PhD?

Tom said...

If there is one thing that I can thank my major advisor for is his insistence that my thesis actually be a compilation of my previously published and to-be-published manuscripts with only a final short chapter tying all the elements together. Even the introduction was to be written as a literature review to be submitted to a peer-reviewed journal. His motto was, if you're going to write it up, it may as well get published. I've adopted that motto myself.

It's nice to see you work along those same lines.

Anonymous said...

When I was a grad student, my advisor said that there were two views on what a PhD was. One view was that it was a hoop to jump through, a step on the path to the next step (often industry, but I suppose could be a post-doc too). For this, stapled papers are (I guess) fine, as long as they are your papers and not someone else's. (With the large lab groups in some fields, that can easily happen, even without plagiarism.) The other view is that a PhD is a chance to write a magnum opus, something to live on for the next decade. As long as you and your advisor share the same view, you're ok. Dissimilar views can lead to problems.

(Personally, I strongly believe in the magnum opus. I turned my thesis into a book and I'm still getting citations, talk-invitations, and requests-to-review based on that book.)