How to write a literature review

A literature review is not a summary of everything you read. It is an argument about the state of a field, and the sources are the evidence.

Updated 21 September 2026 · 9 min read

The commonest mistake is treating a literature review as a reading list with commentary — one paragraph per source, in the order you happened to read them. That produces something nobody can use, including you. A review organised that way cannot show a pattern, because patterns run across sources rather than through them one at a time.

What it is for

A literature review does four jobs at once:

1. Shows you know the field well enough to work in it
2. Maps what is already established, and where researchers disagree
3. Identifies a gap — something unresolved, untested or unexamined
4. Justifies your own study as the thing that addresses that gap

The fourth job is why it exists. Every earlier job is building towards “and therefore this study”.

Organise by theme, not by source

This is the single change that turns a weak review into a good one. Group sources by the argument they contribute to, and discuss several together in each paragraph.

The second version makes a claim, supports it with multiple sources, and explains a disagreement. That last sentence is the sort of thing a marker is looking for.

Finding themes in what you have read

Build a matrix before you write. A table with one row per source and columns for the question asked, the method, the sample, the finding, and the limitation. Once it is filled in, the themes announce themselves — you will see three papers using one method and getting one answer, and two using another and getting a different one.

Themes usually come out as one of:

• Chronological — how thinking changed over time. Use only when the change is the point.
• Methodological — grouped by how the work was done.
• Theoretical — grouped by the framework the researchers used.
• By sub-question — the most common, and usually the clearest.

Naming the gap

“No one has studied X” is the weakest possible gap, and often untrue — usually it means you have not found the paper yet. Stronger forms:

• A contradiction: two credible bodies of work disagree and nobody has resolved it.
• A limitation shared across studies: every study used undergraduates, so nothing is known about older adults.
• An untested assumption: the field assumes something that has never been directly examined.
• A new context: the finding is established in one setting and has not been tried in another where it might not hold.

Being critical without being dismissive

“Critical” does not mean listing flaws. It means judging how much weight a finding can bear. A small sample is not a reason to discard a study; it is a reason to treat its conclusion as provisional and say so.

Useful things to weigh:

• Does the method actually measure what it claims to?
• Does the sample support the generalisation being made from it?
• Is the conclusion larger than the data?
• Who funded it, and does that bear on the framing?
• Has it been replicated, and if not, why not?

Shape

• Introduction — scope, and how you chose what to include. Say what you excluded and why; an unstated boundary reads as an oversight.
• Themed sections — the bulk. One theme each, each ending with what it establishes.
• Synthesis — what the field collectively knows, and where it stops knowing.
• The gap and your study — the paragraph the whole review was built for.

A checklist before you submit

• Most paragraphs cite more than one source
• No paragraph is a summary of a single paper
• Every section ends with what it establishes, not with the last study mentioned
• Disagreements between sources are named and, where possible, explained
• The gap is specific, and plainly leads to your study
• Your inclusion criteria are stated
• Reading your section headings alone shows the shape of the field