|
Getting your Trinity Audio player ready...
|
Do you have 20 hours a day to spend reading research papers? Yawn! The number of journal articles published keeps increasing every year. That is why a literature review is so important: it gives busy researchers a way of keeping updated with important developments in their fields, and helps them identify the most important and seminal studies that they should be aware of.
What is a literature review?
It’s not just a paper that cites a lot of other papers! According to Jahan et al. (2016), a literature review provides “a complete summary of the current literature relevant to a research question.” Carrera-Rivera et al. (2022) describe literature reviews as a “critical first step in research to understanding the state-of-the-art and identifying gaps and challenges in the field.”
Essentially, a literature review collects and summarizes existing research on a specific topic, with the purpose of generating “a comprehensive description, integration, or conclusion” about that topic (MacLeod et al., 2021).
Types of literature reviews
Your literature review can be
- Narrative (tells the story of the research)
- Systematic, following a detailed and comprehensive search strategy
- Scoping: a preliminary check of how much research has been done on X and what kind of studies are they
- Integrative (incorporates diverse methodologies, like qualitative and quantitative)
You can also directly work with data, pooling and combining data from multiple different studies and reanalyzing the data to draw new insights. That’s called a meta-analysis.
One really neat thing about meta-analyses is that they can be prospective, where you plan them in advance. You identify the studies or cohorts you want to include even before the final results of those studies are in. This prevents some of the bias that arises when you follow the traditional approach of doing the meta-analysis after the studies are published.
How to write a literature review: Basic steps
If it’s the first time you’re writing a literature review, here’s how you can get started.
- Confirm what your literature review is meant to do: Is your literature review part of a larger journal article or thesis/dissertation? Or is it meant as a standalone journal article?
- Decide on a methodology: You can choose between narrative reviews, systematic reviews, scoping reviews, meta-analyses, etc.
- Identify a research question: You need to come up with a specific and clear research question that you hope to answer through your literature review. A poor research question would be something vague and unspecific, like “urinary tract infection in women”. With such a research question, you’d end up with far too much information to fit into a single paper. A good research question would be like
“Is a history of urinary tract infection a risk factor for developing renal cell carcinoma in postmenopausal women?”
- Make a list of keywords: You’ll use these keywords to search for relevant literature, so include
All the variables you are interested in, like the patient population, intervention, outcomes
Any synonyms or closely related terms (e.g., “obesity management” and “obesity treatment”; “vitamin A intake” for “vitamin A consumption”; “COVID 19” for “SARS-CoV-2”)
Take a look at how Lazar et al. (2024) developed and described their list of keywords in their narrative review on the mechanisms of initiation and progression of lung fibrosis following SARS-CoV-2 infection.
- Find out the best sources to search: Make sure you’ve access to the databases that are most relevant for your discipline, like CINAHL if you’re a nursing researcher. PubMed, Scopus, and Web of Science are all popular and reliable. Pro tip: Apps like R Discovery come with a Reading Feed, which identifies articles based on your interests and delivers personalized recommendations in a simple social media style. Use them to make sure you’re not missing out on key trends or the most recent breakthroughs in your field.
- Set inclusion and exclusion criteria: You can’t possibly include every single piece of research you come up with. Decide in advance what kind of studies you’ll include and which you’ll reject. See for example, how McConvey et al. (2022) set inclusion and exclusion criteria for their systematic review on triggers for referral to specialized palliative care in advanced neurologic and neurosurgical conditions.
- Summarize and synthesize studies you shortlist: You mustn’t just paraphrase the abstracts of the studies you’re reviewing! Add your own interpretations and comments on their methodology, findings, and conclusions in your literature review. Evaluate each study’s weaknesses, strengths, and contributions. If you’re focusing on a particular study type in your review (e.g., clinical trials), it’s a good idea to use a formally developed quality appraisement tool like Cochrane RoB 2.
- Look for themes, controversies, or gaps among studies: As you read through various studies and take notes on them, look for similarities and differences. See how Majeed-Ariss et al. (2015) identified themes in their systematic review on black and ethnic minority patients’ views on self‐management of type 2 diabetes.
You also need to speculate about what could be responsible for differences between studies, especially if they’re reporting starkly different outcomes. Take for example this paragraph from Kiss et al.’s (2021) systematic review on granulocyte and monocyte apheresis (GMA) as a therapy in ulcerative colitis:
However, data on the efficacy of GMA are still debated. The first studies published in Japan showed remission or response rates of up to 60%–80%.4–6 Sands et al reported a study with a large number of patients comparing GMA to a placebo, and they found no significant difference in terms of clinical response.7 This substantial difference between studies could be explained by the heterogeneity of patients’ characteristics, most probably by the varying severity and extent of the disease.
- Write (and rewrite) your review!: Once you’re done with your literature search, start writing the actual literature review. Refer to all the notes you took during the search process.
If your literature review is a standalone paper: Start with an introduction that explains what topic you’ve chosen and why a literature review is necessary on that topic. Depending on the type of review you’re conducting, you may need to include a detailed Methods section that describes your search strategy as well. Your Results section will cover the “meat” of your review and present your insights to readers (be sure to also include essential numerical details like total number of studies included and excluded!).
How to write a literature review using AI
As large language models (LLMs) like ChatGPT become more popular in academia, it’s natural to wonder whether you can simply outsource your literature review to one.
In short: No!
Tools like ChatGPT are capable of generating “academic-sounding” text but they cannot replace the original insights and creative perspectives that YOU come up with, drawing on your years of experience in the field.
Further, even the most recent version of ChatGPT, ChatGPT4o, relies on training data that is around 2 years old. That means it is not aware of recent developments, including recently published papers in your field.
Finally, it’s been well documented that ChatGPT can generate fake citations and references that look genuine. In the study by Kacena et al. (2024), up to 70% of the reference list generated by AI in a review article was fake.
Nevertheless, you can still harness the power of AI in your literature review. Use a tool that has been developed specifically for academia, like Paperpal, and which can actually search the literature and find genuine and relevant published research for you to cite. Be sure to scrutinize all AI-generated text to ensure its accuracy and report using AI as mandated by your journal (in the Acknowledgements section, cover letter, etc.).
