How to write a scientific literature review?

May 16, 2009 4:18 AM

I'm at that point in my PhD where my experiment is (almost) up and running, so I have more free time to do other stuff. I've decided that writing a general introduction to my thesis will be less beneficial than trying to get a review article published. How do I go about writing a review article, from the perspective of an unpublished graduate student?

I'll be getting help from my advisers, both of whom are well respected in their individual fields. I'm looking at a number of different areas and trying to tie them together, which hasn't been done in any previous review (that should buy me enough originality, right?). I've done the obligatory Google search, but all the advice out on the intertubes is pretty generic, and usually aimed at small reviews instead of reviews aimed at publication.

I have more articles on the topics than I need, I just need to start writing something now. What do you guys do to streamline writing reviews (not just scientific, any kind)? Any tips on reducing the pain and increasing the fun? How do I go about planning this thing? Tips for writing it without drowning in information?

posted by doctor.dan to Science & Nature (12 answers total)

In my field, I am under the impression that review articles are invited. Are you sure you can get one accepted if you do write one?
posted by Comrade_robot at 6:27 AM on May 16, 2009

I don't mean to be glib, but isn't this something that you should be discussing with your advisor? You'll need their support and permission before submitting a paper.

If you advisor is absent, or unhelpful, talk to your committee members or other trusted faculty about this.
posted by chrisamiller at 6:33 AM on May 16, 2009

It depends on your field, in my (smallish) field I would look up the specialist journal (which I am already aware of) search to see whether my area had been covered much recently, the get in touch and see if they wanted something, perhaps supported by a supervisor who I would ask for advice from if I was still a PhD student. Your supervisor might be able to tell you the best journal if you don't already know and offer advice on the specific topic and how you might frame the paper. I would advise that any paper you wrote like this should best come out of the stuff you're already looking at, e.g part or all of your lit review, as this maximises its value, minimises additional work and the oublication can be used to support your final thesis.
posted by biffa at 6:40 AM on May 16, 2009

I should elaborate, maybe -- in my field, review articles are seen as kind of a feather in a cap, and generally they are given to people who are already kind of famous in the field. I am not aware of any cases (I'm not saying there are any, but I haven't really heard of any) in which an unpublished graduate student has written one. I know of one or two review articles written by PhD students, but they were already doing very very good (published) work (PNAS, Cell, Science, Nature).
posted by Comrade_robot at 6:51 AM on May 16, 2009

So you have an experiment up and running, and in the absence of any results so far, you feel ready to publish a review of the field?

Not. Going. To. Happen.

I work in a similar field to you, and I only know of one PhD student I have worked with who published a review of the kind you are talking about AFTER she had finished her PhD. In her PhD she had published 7 or 8 papers in journals such as Science, Nature Immunology, NEJM and PNAS, and our supervisor is a world leader in the field (and we're talking: has won every prize short of the Nobel).

Her review was published in IUBMB Life. No, I hadn't heard of it either.

By my reckoning, and judging from your previous questions, you are little more than 12 months into your PhD. You should spend your 'free time' working on additional projects to get you those 10 or so papers that will get your foot in the door for writing a review.

If you are super keen on writing something, then write an article for something like this in a magazine more suited to your field. It will get you and your work out there. Do your time, put your head down, and good things will happen.
posted by TheOtherGuy at 7:42 AM on May 16, 2009

Your thesis introduction and a review article are going to overlap a lot, so it's not really an either/or proposition. You need to ask your advisors which project they consider to be more beneficial. Ten hours of asking semi-strangers on AskMe is useless compared to a 5 minute conversation with your advisors. They will know how to propose a review article and the scope of the work.

As far as managing tons of papers, learn to love EndNote. Also Papers for the Mac. The rest is basic scientific writing- identify the important discoveries and researchers in the field. Who is using what techniques now? Compare and contrast results. You'll have to do a ton of reading before you can start writing.
posted by mbd1mbd1 at 7:49 AM on May 16, 2009

In part, what Mbd1mbd1 said.

This is going to sound kind of weird, but consider starting a blog and using it as a sort of open journal where you keep summaries of papers, the occasional note to self and where you can synthesize the review / introduction you want to write. You're going to need that information to be at your fingertips sooner than you want to believe. A blog lets you publish something without actually dealing with publishing something and gives you a head start on networking with other people in your field.
posted by Kid Charlemagne at 9:25 AM on May 16, 2009

I've got to echo the "not going to happen" sentiment here. Having just read through the instructions to authors for the big journal in my field, the Journal of Virology (which I am pretty sure extends to all American Society of Microbiology journals), I can tell you that they do not take unsolicited review articles.

If you have free time (especially as a first year?), you aren't doing it right. You should be up to your neck in the literature or working on side projects.
posted by The Bishop of Turkey at 10:09 AM on May 16, 2009

If you want to write a review, tell your adviser. It is possible that he or she gets a lot of requests to write reviews and would be happy to dump some of the grunt work on you.
The best way to streamline writing a review is to be a genuine expert in the field - know everything about not just the key results, but who the key researchers are and what they're doing.
posted by nowonmai at 11:01 AM on May 16, 2009

Response by poster: Comrade_robot, high impact journals (Nature Reviews, etc) only invite reviews, which are usually written by those already well-established in the field.

chrisamiller et al, I should have been clearer. I have mentioned this to my advisers and they agree that it would be worth a shot as it is very novel. They have recently published something similar, so I have a good starting point in that sense. As a review, it provides an angle which hasn't been explored yet. As an introduction to my thesis, it provides a nice structure to why I'm doing the bloody thing.

biffa, agree. The only reason I am considering this is because it would be exactly what I would place into my introduction/literature review. Not much point wasting time writing a review for some area where I won't end up publishing.

TheOtherGuy thanks for the honesty (and the link). I know that one of my adviser's previous students received an invitation from Nature Neuroscience to write a review after her manuscript was accepted in Nature. I am in now way aiming for this. That being said, a new postdoc in the lab (fresh out of his PhD, and I know I shouldn't compare) had a review accepted in an area where he has not published or done previous work. The other authors on the manuscript have, which was enough to boos the expert factor.

mbd1mbd1, my university accepts published work and manuscripts in lieu of chapters for theses, with a review being considered an acceptable replacement for a general introduction. Already love Papers, made life a lot easier (with EndNote, my feelings are a lot more mixed, necessary evil at this point).

Kid Charlemagne, my few forays into blogging never really worked out. Might give it another shot, if I can get the WordPress app working on my iPhone...

The Bishop of Turkey, I never expected the review to end up somewhere high impact. It would probably end up scraping the bottom of the barrel, but in my position any paper is better than no paper. I'll pump my advisers for more info on the actual likelihood that this thing would be published/where.

nowonmai, the initial idea for writing the review came from my principal adviser. He personally hates writing the things because he has to be meticulous. I don't mind doing it, and my advisers' names will be on the manuscript so it won't just be little ol' unpublished me.

Thanks for all your comments/suggestions/hefty doses of reality.
posted by doctor.dan at 4:38 PM on May 16, 2009

I'm late to the party, but here goes: I wrote an uninvited review article pre-Ph.D. But I did it with my big fancy advisor as second author. In my view that's the only way to go since you aren't talking about your own published work (I assume). Rather you're going after other people's work and tying it all together, and journal editors usually only get that kind of stuff from crackpots.

So, you need a more established name on board to give you credibility. Also you want them to give you critiques and feedback about it long before the reviewers get a crack at it.
posted by drmarcj at 7:19 PM on May 18, 2009

Response by poster: drmarcj, I will have both supervisor's names on the paper, so there won't be an issue about that. I'm guessing the manuscript will go through all of Dante's stages of hell as before it gets anywhere near a reviewer. Thanks for the tips :)
posted by doctor.dan at 4:09 PM on June 15, 2009