I’m brewing a couple of literature reviews at the moment and I’d forgotten two particularly tedious bits of the process; being sufficiently comprehensive, and accessing the papers ‘off campus’. Bloody hell!
In 2000 I wrote my first literature review as part of my MChem project and the topic was ‘dendrimer catalysis’. There were only about 2 papers on heterogeneous dendrimer catalysis and a quick Web of Science shows that up to 2000 it was possible to be fairly comprehensive with 44 publications from this search. Obviously other key words were used when striving for comprehension. Looking at the graph below, I’m pretty glad I didn’t do it a couple of years later. I remember feeling fairly satisfied that I had all of the papers that I could access. Some PDFs, but lots as photocopies from the library paper copies, also interlibrary loans and articles photocopied on trips to other university libraries. In many ways, knowing the correct reduction from journal size to A4 on the photocopiers was distinctly less frustrating than accessing PDFs.
Web of Science hits for ‘dendrimer and catalysis’ between 1996 and 2016
The point here is that I was striving to be as comprehensive as possible because a literature review is not an essay. In an essay, it may be possible to be less comprehensive in the ‘review’ part of the literature. I wanted to know all the things. I had a thing or two to learn about effectively summarising information and wrote something more akin to a narrative account of the literature than what I’d do now (I’m a big fan of summary tables in review articles and a more ‘tour guide’ type approach).
A plausible definition of ‘sufficiently comprehensive’ is that which you can track down and understand. Actually that’s a pragmatic definition given complaint number two: accessing the damn papers. So far this afternoon I have logged in to the same access system about 5 times through various websites. And clicked through incessantly until I got the damn PDFs. And googled the article titles to find open access versions when I was denied access. It’s making me want to completely revise the definition of ‘sufficiently comprehensive’ to be ‘anything that is open access and to hell with the logons’. But we have all of these expensive subscriptions so should at least use them.