No matter where you are in your research from starting out and seeking the gap in the knowledge, to rounding off that last few touches to a journal article, there will come a time when your desk looks like this.


Perhaps your desk doesn’t look like this, and you are tidy in the way you read papers, perhaps reading them all on digital devices (honourable undertaking that I haven’t yet mastered – I dislike sitting down to read at length if the thing I’m reading is not printed in black on white paper). But no matter how you read, or what your desk looks like, we all reach the point where the mountain accumulates and threatens to bury you under an endless web of “author publications,” “paper citations” and “cited by.” Here follows a few pointers from my experience on literature searching, reading journal articles and citation management.  It’s really advice for beginning environmental science PhD students, but it might be valuable to others interested in research methods.

Literature searching

When embarking on a new topic, the easiest way to find information is through a keyword search. You can do this simply on abstract database sites like Web of Science. If you’re using one of these it’s worth restricting the search to the last 5 years for an initial investigation. The latest papers will tell you the state of the research environment for your chosen topic right now. They can help you identify the current knowledge gaps and by going through a few of them you can usually identify a few common papers.  By looking at the papers they all cite, you will eventually find your way to the seminal works in your field and possibly review articles.  Seminal papers (highly cited) can identify the key authors in this specialism. By looking at this author’s other work and following the “cited by” web from the seminal papers, it is easy to gain a very quick impression of the literature in your chosen field.  Review articles can be very helpful but bear in mind unless you are interested in first principles, generally the more recent a review article, the more useful it will be to new research.

When literature searching in an expansive topic, the question will soon arise “how far back in time is too far back?” and/or “how remotely related to my research is too remotely related?” The days of limited access and hidden texts in vaults is gone. Today scientists have easy access to reams of information dating back to the turn of the 20th century and beyond.  For an initial literature search, I would advise not casting the net too far back.  In the early stages of a project, the most important thing you can do is find out what people are doing right now, and how your proposed project can fit in to the current body of literature.  When you are deeper into your project, perhaps devising methodologies for experimentation and statistical analysis this is when older/less related studies can be helpful.  At these stages, keyword searching for specific techniques can be very useful, and don’t discount the use of textbooks and manuals.  During my PhD viva, I was accused of re-inventing the wheel somewhat when I failed to cite papers from the late 1980s and early 1990s that had used a similar technique.  This was not out of disrespect, it was due to not looking far enough back down the chain of literature to identify these early papers.  This is how sometimes a reliance on the most recent papers can catch you out – if the papers do not cite these older methodologies and results then how will you find out about them?  I have learnt to cast the net wider and further back when considering experimental design by other researchers in a given field.  In the latter stages of a research project, seeking out comparable results and items to discuss can easily turn into yet another expansive search. In my PhD, I ended up citing papers from atmospheric chemistry and microbiology sources, despite my main focus being soil physics.  It’s easy to forget in your specialism that environmental science is a systems science subject and as such you may need to draw in discussion from a wider range of source material than you might have initially thought.

Finally DON’T restrict your literature search to the peer-reviewed journals. There is a wealth of information to be found online, using Google (consultancy reports, governmental reports, white papers…) and offline in books and magazine publications. Wikipedia, even though you’ll never cite it, can give you the lowdown on a new topic very quickly and provide invaluable keywords for database searches. Though you will streamline and academise your sources later on (particularly when writing up!) you shouldn’t restrict sources of information at the early stages.

Reading scientific papers

When confronted with 25 pages of tightly arranged font, equations and figures, it is excusable to feel daunted (especially if there are 20 more to follow).  The first thing to ask is: how much of this do I need to read to get the messages I need at my current research stage?  There will always be the papers that you’ll read from end to end; seminal papers, review articles (relevant sections) and the most recent work in your field.  You’ll end up coming back to many of these papers time and again, to look at different sections. So there’s really no need to go through each one with a fine tooth comb. Instead: (the following assumes you’ll always read the full abstract first – recommended in any case)

  • If you are beginning a project, focus your reading on the introduction, discussion and conclusions.  Find out the justifications for their research, the aims of their research, what they found out and what they suggest doing next.  Delving to deep into the methods and results can waste time at this stage, there’ll be a lot to get through and you can keep reference to the most useful papers to read in depth later.
  • When developing your methodologies focus on the methods sections (funnily enough). Follow method citations even if the paper itself is outside of your field. You might not be interested in the results of the cited paper at all, just the method they used to get those results. Whatever you do, don’t look at the paper and go “oh, ‘Journal of Cattle Management’, that’s not relevant to my research” when the authors have used an experimental or statistical method that could be.
  • When you have results you’ll probably want to compare them to others in the literature. You can identify results in abstracts much of the time (numerical abstracts are quite ‘in’ nowadays if you compare to abstracts from 20/30 years ago). Looking at results sections of relevant papers can also give you a great deal of information on how best to visualise your own data and the types of statistical analyses that are most useful.
  • Writing up and discussion stages.  Joy, this is the fun part. Sit yourself down and conduct another thorough literature search on your own observations. Has anybody found this before? How did they explain it? Be prepared – you might end up outside of your primary subject area again. Prepare to let this stage hoover up your time. Think of yourself as the Sherlock Holmes of the literature. Have you exhausted all leads? All possible explanations from the papers you know of?  Keep an eye on introductions, results and discussion sections.
  • Conclusions, finalising writing up, preparing papers. Well done. Now be prepared to do another literature search. Years have probably passed by now (sorry but it’s true!) since you searched your main keywords right at the beginning. You don’t want your last cited paper to be from 2 years ago. If you’ve kept abreast of the literature throughout, well done! If not, the 2013 papers in your subject area are calling you.

Citation management

I think it’s nearly enough about this. Writing about literature searching is almost as wearing as doing it. The last thing I want to do is highly recommend to all new researchers that you use a citation management system.  I let my university’s inability to provide the software for free to a student’s personal computer to put me off use of this software. As a result I wrote a thesis and managed all 300+ references manually. That is, by hand. That is, checking the reference list and the text match up and no mistakes have been made, slowly, tediously. I am still doing so in the papers relating to my PhD topic as I have no reference manager set up. Copy-and-paste is no way to go about this while sensible options are available.

Now I am staff, I have a staff computer and so I can have citation management software installed. My platform of choice is EndNote, which also has a web-based version to cite on the move!  There are also free web-based programs (i.e. Mendeley) which I haven’t used but seems pretty good to me (but anything would seem pretty good to someone who processed their PhD citations manually. If there is one thing you take away from this blog article it is not to do that).

Good luck in all your literature searching and reading endeavours!

In my next ‘life in environmental science’ article, I’ll talk about the topic I introduced earlier: environmental science as a system science discipline.  Using my research career thus far and its progress from soil to atmospheric science as an example, I’ll show how the many strands of earth and environmental science are highly interrelated. The next article on Ground to Sky will return to the science, and I will be discussing current thinking on how we can cool down our hot cities.