Hello!
In the previous post, I presented some of the approaches that have helped me carry out a literature search related to my research project. Often, this process culminates with a written article or chapter where we summarise the most exciting findings and try to provide a digested overview of the field to the readers. Hence, assuming we have narrowed down possible topics and collected and structured multiple search results, we should have enough information to start planning our review document. Therefore, in this post, I would like to share some practical suggestions that made this writing task much more manageable and entertaining.
2. Identify trends for each category analysed
3. Keep notes of interesting figures
Review all references
As a first step, I reiterate the importance of reviewing all the references that may have a chance of being cited in the main work and create a dedicated section on the reference manager (such as Mendeley or EndNote) for these articles. Next, I would go over every entry to confirm it includes the correct details in the corresponding fields. In fact, this check should be performed every time a new paper is retrieved. This way, we ensure that the bibliography added at the end of the review is correct and complete and allows readers to refer to the relevant papers they find particularly useful for their work.
Identify trends for each category analysed
Next, when delving into the results, it is helpful to analyse our table of structured results and ensure that all fields are understandable on their own. Short sentences that I initially added to the table were often not detailed enough, and I would repeatedly return to the original papers to collect more information. An additional benefit of having this comprehensive description of each article is that it will later help identify and highlight relevant differences between studies, especially among those applying similar methods.
My strategy was to go through each column (representing one topic of interest), identify frequent instances and create categories from them. Often, we are dealing with recent literature (e.g., last five years), and by identifying such patterns in the data, we have a better understanding of established and novel practices in the field. Moreover, these categories should also support defining the sections of our written document and, their frequency can even help recognise which parts require a more extensive discussion.
In practical terms, I skim through each column and create a bullet-point list containing the category names with a frequency counter and, for each one, including the related references followed by a short note of the detail that makes the work belong to that list.
Finally, these lists can also be represented through insightful visualisations that summarise the findings and give readers a quick glance at the work.
Keep notes of interesting figures
As hinted above, figure panels can be beneficial to communicate and complement the details described in the main text. In this regard, I found it helpful to keep notes of appealing figures from the papers I reviewed while filling the summary table in Excel.
Note that re-using figures from published work often implies submitting a request to the copyright holder (there is generally a reference for this on the article’s webpage). It is recommended that this process is carried out as soon as possible since it can take some time to be approved, and be aware that there may be associated costs. The latter point is one of the reasons why it is convenient to have notes for figures highlighting the same aspect: if we are not able to support the costs for a given copyright request, we can check our list for alternatives that have a lower or no cost.
Finally, I would add that after converging on a final set of images, it is helpful to place all the panels side-by-side to get a complete overview and decide if they communicate the information we intended.
Slow and steady writing
“The secret of getting ahead is getting started.” – Agatha Christie
Regarding the writing step, the crucial aspect is to start! I prefer booking a few hours without distractions to sketch each section. Then, I believe maintaining the flow is more critical than re-reading what we just wrote. Overall, it is easier to review a section iteratively than to ensure we write the perfect sentence at each keystroke. Therefore, in a first draft, my approach focuses on transferring the ideas on the mind to the text on the document, with as few hesitant pauses as possible.
Furthermore, I found it more productive to split the work into smaller writing tasks and distribute it over several days. First, writing for very long periods can get overwhelming, with the risk of decreasing the quality as we get more tired. Additionally, breaks help to return to the text with a fresh mind, making it easier to revise previous sections.
At this stage, the preparation work from the two previous steps also comes in handy. On the one hand, writing readable sentences simply meant merging the notes added on the bullet-point lists. Additionally, given that the corresponding reference was already present in my lists, I could immediately cite it in the text, reducing the time spent looking for the sources. On the other hand, I would often consult the list of noteworthy figures and decide where to place them in the text (since there was no hard limit on the number of figure panels, it also gave me more freedom).
Conclusion
These are some of my suggestions to draft a literature review, and, again, I would reiterate that the ideas presented here are based exclusively on my personal experience. Overall, I found the data analysis and writing phases much more entertaining than the initial “review” step. Most likely, as I got a clearer vision of the previous work in the field, it was easier to have a more defined direction in subsequent stages.
I am looking forward to finalising my work and applying what I learned in my research project.
Please feel free to share your thoughts and experiences about literature reviews!
Have a great day!