Why making a Literature Review matters, and how it’s done.
Any quest into a new field requires knowledge on what has been done over the years already. There is no point in re-inventing the wheel, unless you’ve found a way to make a wheel friction-less or afloat. Even for that, one would still require the basic knowledge that this is novel. And this should be used with certain terminology that’d convey the uniqueness of your new invention.
A PhD is a quest to the edge of human knowledge and the push the boundary a bit further than where you found it. It’s a contribution of your intellectual work, to the sum of all the work done till now in your field. The picture here explains this very simply.
So the first step of any PhD is a literature review. And it’s done for a couple of reasons.
- Novelty. For you to reach the boundary point of all knowledge in your chosen field. Without this, it’s really hard to know what is already known. And novelty is critical for truly investing your time in research.
- Boundary. For your understanding of what’s been tried, and what’s beyond this boundary, and what’s attainable from within the circle now. Eg: Sure you can try to experimentally prove that time travel is possible, however, all of humanity has not progressed enough to reach the point where you have the necessary resources and knowledge to be able to do that. So knowing the boundary is key.
- Terminology. Scientists and all researchers alike, love this. And it makes sense, and our lives easier. Otherwise, it ends up becoming like the scene from the movie Transformers, where Sam might know all about outer space travel and existence of space robots, but if no one’s able to understand his lecture, then it’s meaningless. Similarly, to make sure everyone’s on the same page, scholarly terminology is important.
4. Citations. Seeking your unique answer needs help from others, since you can go on proving everything necessary for your proof, by yourself; especially when it’s already proven. And you need to know who did it, and when. That’s essentially how citations help you, and it redirects the audience to that proof/publication.
How to make a literature review
Firstly, my professors at Oxford, said as a true scholar, you shouldn’t be doing a ‘review’ of the literature, rather an ‘analysis’ of the literature. The difference is, in the latter, you use some of your independent thinking as well, to see if the literature is reliable (in terms of methodology, sample size, statistical significance, repute of journal which in turn validates the reliability of the peer review process), and how other scholars are agreeing or disagreeing with this article.
So from here on, let’s call it a literature analysis. The big question here is: what is literature? In this context, it doesn’t refer to Shakespearean sonnets, rather to scientific journal articles where researchers publish their work, which is validated by an ‘independent’ peer review process where experts in the same field review the work as worthy of publishing. But there are also grey literature, where considerable amount of knowledge lies. These include, media articles, reports by governmental and non-governmental agencies and even masters theses.
From here on, it might get a bit detailed, and would vary vastly by fields of study. I research somewhere in the intersection of Environment Sciences, Engineering, Policy and Urban Planning, Development Studies. So it certainly is pertinent to the details for the above fields. Roughly 100,000s of new articles are found on these topics, everyday! So obviously you can’t keep up with reading every single one of them. That’s where filters really help.
Define your search as narrow as possible. If your researching on effects of climate change on volcanoes in Iceland, you have to narrow it down to exactly that. Just ‘volcanoes’ could give you information on the Mt. Fuji tourism or just ‘climate change’ will redirect you to a page with the American President denying it. But you’ll have to understand each of the terms above to a certain depth, after which you can start aligning them. So for example, you can start with ‘Geology of Iceland’. Or, ‘Effects of climate change on volcanoes’.
Once you’ve identified the areas where you want to be more aware, then the first step is to identify your databases. The most popular ones are Google Scholar, Web of Science and Scopus. The latter are paid, but the former, is formidable enough to give you a good enough insight. Most Universities have subscriptions to these by default anyway.
Next is using Boolean algebra to create your search strings. Look here for the details. It’s basically to accurately type out what you’re searching for. And it’s plenty useful in filtering out your results. For example- Climate Change gives about 264,000 results. Climate Change AND volcanoes yield about 91,000. (Climate change AND Volcanoes) NOT Japan gives about 50,000. (Climate change AND Volcanoes AND Iceland) about 26,000 and (Climate change AND Volcanoes AND Iceland) NOT (Ice Caps) gives about 129. It’s simple and effective. Then just a few clicks and you are right through to the journal article where you can read the abstract (short summary) and decide whether you’re gonna read further.
Most papers are behind the arcane system of paywalls. There’s much debate about that here’s a history and here’s a scientist’s take on that. There are also quite many open access journals coming up. Another way you can get access is to write to one of the authors requesting it through email or through Research Gate (Social Network for Researchers).
Now that you’ve got all the articles you want, you must need a database management software to manage these, know which one’s are read, which ones are good, how to classify/tag them, and importantly, how to easily cite them when writing in word through a plugin. The most popular ones for this purpose are : EndNote, Mendeley and Zotero. The first is paid and both the others are for free. Having used both EndNote and Mendeley, I really prefer Mendeley over Endnote, because of the ease of uploading literature (It automatically fills in the bibliographical information), ease of managing literature, in application search, pulling directly from the internet the documents using the plugin, cloud storage, customer support, collaboration with your group, and most importantly since it’s free, I will not have to worry about subscriptions even after I leave University. Here’s Mendeley. They used to be independent, but now they’ve merged with one of the biggest publishers behind the paywall -Elsivier (who also happens to own Scopus). See how Elsivier now has control over not just publishing, but also with searching and data managing.
That’s the three main processes involved in literature search: Searching in an online Database, Procuring the article, and Managing it among all literature. Now comes the second series of processes- the understanding and analysis. Reading every paper on your desktop will take up a millennium. So, choose wisely, which ones you want to brush through, and which ones have a direct relevance for you that you will go through them in great detail. For this, the abstract helps greatly. If still unsure, the figures, and conclusions help. So I think there are 3 types of reading literature.
- Detailed:
I’d probably print it, highlight relevant points, jot down interesting inferences on the side, and document the summary in my journal and then type it onto my literature review file. Such an engaged process ensures that through the 4 step contact with one paper, that you thoroughly get to know it. I’ve spent perhaps 2 days for 1 such article. And in total, I’d do this for 5–10% of my literature.
At the end of the paper- try to ask yourself some of these questions:
What was the main research question the author set out to answer? Have they answered it? What were their methods? Any new concepts introduced? What’s the quality of the paper? How would this be relevant to my research?
2. Semi-Detailed:
The abstract, the introduction, the figures, the discussion and conclusion. Glance at the methodology to understand how they’ve set out to do their research. Take notes onto your Literature analysis document.
3. Glancing:
There are papers you’ve downloaded, since you know someday it will be useful, or maybe it’s tangential to your literature, or maybe there is a small part that’s useful to you. In all these cases, there might be a time when you will read the paper in slightly greater detail- Perhaps by #2 in a semi-detailed way. Otherwise, just key point notes as to why the paper exists in your data base would suffice. But this too has to go onto your Literature analysis document.
Through all these processes, two essential things that must come out are: relevant Tags for the article. At first, I saved them on folders. But my current professor assured me that these ‘folder names’ will change, and how some papers will be cross-cutting and belong to more than one folder. I realised this in the following weeks, and started tagging them. That has been hugely useful when retrieving literature to recall them, or to just see which ones talk about a specific tag.
Second key aspect is snow balling to other literature from this. Every paper cites many others. And this is very handy for you to not repeat the whole search string process, and instead start going from one citation to the other. Soon, you’ll start depending more on snowballing and searching online. This way, it is a lot more targeted.
So when would you end this mundane process?
At some point, you’ll start saturating- you will see the same papers again and again. And then you know, you’ve read enough. This usually takes between 6 months and a year.
Literature Analysis is an essential part of any research project. And in a PhD it’s the most boring, but also the most enriching part. So, keeping in some enthusiasm packed for this will help set sails smoothly over the sea of literature we have produced. And who knows, someday, when your papers are published, new sailors set sails to fish for your research article!
