As an early career researcher, I am on a constant journey filled with speedbumps and stomach-tightening moments; along with joy from conducting novel research and continuous enlightenment and self-improvement. I’m sure these feelings are no strangers to the many other ECRs we have in our readership. This October marked the beginning of my second year of my PhD studentship and in the midst of monitoring my progression, I have been reflecting on what I have achieved over the last year. The first year of my PhD has been quite the learning curve in many ways; not just in gaining an understanding of a new subject area or developing protocols based around equipment that was foreign to me. I’ve had to make the transition from my undergraduate and masters’ degrees – structured, with set deadlines and requirements – to try and see the bigger picture of a PhD study. Each step you take in this process feeds in to the end goal and should be treated with such consideration.
One of my main projects has been to write a literature review that reflects the research questions of my study and coherently explains the state of the subject area and the gaps that my study aims to fill. Having spent the four previous years writing essays and literature reviews for university, I felt confident in my abilities to bash out a decent review. In order to make sure I was engaging with all of the relevant material, it was decided that I would undertake a structured (systematic) review… And that’s where I became a little lost. I had never done a review before that was so tightly developed – throughout my previous degrees, it was all about what articles you could find that fit the question you were trying to answer (and most importantly, could they be found on Google Scholar?). The terms “search strategy” and “data extraction” seemed a foreign, complicated concept. I felt like I was wading through murky waters. Now that I’m finally actually writing a review, I thought I would share some tips I’ve learned from my experience of the initial stages.
Define your question. What exactly is it that you’re interested in investigating? Is it one small tight question or multiple broader questions to form a story of research? Know what it is that you’re looking for and allow this to inform your search criteria.
Set up your search criteria. What kind of articles will you read for this review – just peer-reviewed journal articles or are you going to include grey literature, conference proceedings, etc.? What population/method/analysis are you interested in? What language should the literature be written in? Write up an inclusion and exclusion criteria before you start looking.
Create a search strategy. Set up a search string of keywords to find all articles pertaining to your research question. Create multiple strings if that is the nature of your review. I had never heard of Boolean operators before starting my PhD but these have been lifesavers for any searches I have conducted – particularly in my review. Figure out which databases are the best to find relevant literature on and set up an RSS feed so you can remain up to date on articles you could include in your review as you write it.
Don’t be scared by the search results. Finding out a couple of thousand articles have popped up in response to your search strategy is unbelievably daunting. But I promise you, a huge number of these are irrelevant and have only the loosest connection any of the keywords you’ve chosen. Doing a title screen is boring and time-consuming; make sure you have a strong cup of coffee and a good Spotify playlist while you hit the “delete” key over and over again.
Don’t be scared by the huge number of articles after a title screen. Have you spent hour whittling down your literature only to find you’re still left with hundreds that hold promising titles? I went from 15,000+ to 300+ and groaned. This is when the abstract screening begins – so get yourself another vat of caffeine and some power ballads and get ready to read. Keep your exclusion/inclusion criteria in front of you and delete once something doesn’t meet your guidelines. If possible, get another reviewer to go through this process with you – it strengthens the argument for articles left in the review.
You can still get rid of articles after the abstract screen. The abstract might appear to highlight the study as seminal to the research area but upon reading the whole article, you realise it doesn’t address the topic quite the way you need it to. That’s fine, get rid of it. The review should be providing you with resources to inform your specific research – don’t worry about the rest.
Data extraction forms are a godsend. I had read about half the articles for my review before being told about data extraction forms – my desk was littered with scribbled notes on aspects of studies that MIGHT be relevant (but probably not); I had one huge Excel spreadsheet that was reasonably incoherent by the time I had included all the information I thought was necessary. Once I adapted a simple data extraction form to my review research questions, I sped through the reading process and felt a huge weight of my chest. Make sure to think about what you need to know when designing one of these. What were the methods? What were the results? How did they analyse them? What type of study was it? Do the key findings match your research questions? What information from the studies do you need to include in tables and figures in your review?
Don’t buy in to everything an article says. Remember to be critical when reading for your review. Many of the papers might showcase great research, but equally many might not. It is easy to get bogged down by a mountain of reading and forget to think about the limitations and advantages of a particular study – but always mark down if you think something is missing or not accurately interpreted. Writing a review is as much about establishing the current quality of the research area as it is about outlining what has been found so far.
And with that, I’m going to get my head down and re-engage with my current stage of the review – battling the complexities of data synthesis.
Featured images: “Piled Higher and Deeper” by Jorge Cham