The guide
How to write a literature review
A literature review asks you to show what is already known about a question, who found it and where the gaps are. The writing is only half of it. Before you write a line you need the right papers, and you need to keep track of them.
That is where most reviews slow down: tabs full of search results, references copied by hand and a draft with too many gaps marked citation needed. A general chatbot will happily fill those gaps with papers that do not exist.
What a literature review is
A literature review is not a list of summaries. It is an argument about what a field already knows, built from the studies that matter most to your question, and it ends where the evidence runs out: the gap your own work steps into.
That is why reviewers and markers look for synthesis. They want to see papers compared, grouped and weighed against each other, with your reasoning about which findings hold up and why.
Three ways to structure it
Pick the structure that best serves your question. Most reviews use the first.
- ThematicGrouped by idea or debate. The clearest choice for most reviews.
- ChronologicalShows how thinking changed over time. Good when the history is the point.
- MethodologicalGrouped by how studies were done. Useful when methods explain the disagreements.
Step by step
- Question
- Search
- Screen
- Synthesise
- Write
Start with a question narrow enough to answer. Search with a few key terms and their synonyms, then follow the references of the best papers backwards and see who cited them since. Screen by title and abstract first, then read the papers that survive.
Before you write, group what you found. For each group, note where studies agree, where they disagree and why. The writing then becomes a series of paragraphs that each make one point and support it with several papers.
Common mistakes
Avoid: One paragraph per paper
Do: One paragraph per idea, with several papers in it
Avoid: Quoting instead of comparing
Do: Say where studies agree and where they split
Avoid: Ending without a gap
Do: Finish with what is still unknown and why it matters
A worked example
Across three studies, shorter sleep predicts weaker next-day recall (Marchetti et al., 2021; Okafor & Lindqvist, 2022), but the size of the effect depends on what is being remembered (Tanaka & Brooks, 2023).