How to Write a Literature Review: Structure, Synthesis and a Matrix You Can Copy
In short
- A literature review summarises and synthesises what has been published on a topic.
- Organise it chronologically, thematically or by method, and say which.
- Synthesis (showing connections between sources) is what separates a good review from a list.
- A synthesis matrix makes the connections visible before you write.
The first literature review most students write reads like an annotated bibliography: one paragraph per source, in the order they were found. Markers are looking for something else: a map of what is known, where researchers agree and disagree, and what is still missing.
What a literature review is
The UNC Writing Center defines it as a piece that summarises and synthesises published work on a topic, sometimes within a set period. "Literature" means any body of material on the subject, not only literary texts. Unlike a research paper, a stand-alone review does not develop a new argument of its own; it aims to "summarize and synthesize the arguments and ideas of others". Reviews are most common in the sciences and social sciences, and they also open most dissertations and research papers, where they set up the question your own work answers.
Before you write
- Clarify the brief: how many sources, which types, what date range.
- Find two or three published reviews in your discipline and look at how they are organised.
- Narrow the topic. "Social media and mental health" is a field; "Social media use and sleep in university students" is a review.
- Search systematically: our guide to finding scholarly sources covers databases, Google Scholar and "cited by" chains.
Three ways to organise the body
- Chronological: sources by date. UNC warns this only works if the sequence shows a meaningful trend, for example how a field's view changed after a key study.
- Thematic: sources grouped by theme, debate or issue. This is the most common and usually the strongest structure.
- Methodological: sources grouped by approach, for example surveys versus experiments, or qualitative versus quantitative studies, so you can compare what each method can and cannot show.
Whichever you choose, state your organising principle in the introduction, so the reader knows what to expect. The usual shape is introduction, body, conclusion; the conclusion often names the gap your own research will address.
Summary versus synthesis
UNC draws the key distinction: a summary recaps one source; a synthesis reorganises information from several sources to show connections, new interpretations or how a field developed.
Summary: "Smith (2020) found that students who used phones in bed slept less. Jones (2022) found a link between screen time and anxiety."
Synthesis: "Studies agree that late-night phone use is associated with shorter sleep (Smith, 2020; Lee, 2021), but they disagree on direction: survey studies cannot rule out that anxious students both sleep less and use their phones more (Jones, 2022), a limitation the only experimental study so far tried to address (Patel, 2023)."
(The studies in this example are invented to show the form.)
The synthesis matrix
A synthesis matrix is a table with your sources as rows and your themes or questions as columns. Filling it in before writing shows you, at a glance, which sources speak to which theme, where they agree, and where the gaps are.
| Source | Theme 1: effect on sleep | Theme 2: direction of cause | Theme 3: method and sample |
|---|---|---|---|
| Source A (year) | finding + page | … | survey, n = …, country |
| Source B (year) | … | finding + page | experiment, … |
| Source C (year) | … | … | … |
Each column then becomes a section of your thematic review, and each cell becomes a sentence, written in your own words (see paraphrasing vs plagiarism).
Writing tips from writing centres
- Be selective. Include only what relates directly to your focus.
- Use evidence. Back up your interpretation of each source.
- Quote sparingly. Short quotes only, to stress a point or keep a coined term.
- Keep your voice. Start paragraphs with your own point, then bring in the sources that support it.
- Paraphrase carefully and cite every source clearly.
- Revise for concision and remove unnecessary jargon; check every citation against your reference list.
How many sources, and how recent?
There is no universal number: a short undergraduate review might discuss 10 to 20 sources, a dissertation chapter many more. The brief, your supervisor and the published reviews in your field are the guides. On recency, fast-moving fields such as medicine or technology usually expect mostly recent work, plus the few foundational studies everyone still cites; slower-moving humanities fields may expect a longer historical range.
Common mistakes
- One paragraph per source, in the order you read them.
- Describing what studies did without saying what they show together.
- Ignoring studies that contradict the rest, instead of explaining the disagreement.
- No clear end point: a good review closes by naming what is still unknown.
Writing the introduction of a review
Open with the topic and why it matters now, define its scope (which sources, which period, which population), state how the review is organised, and, if the review introduces your own research, end with the question your study will address. A short paragraph on how you searched for sources is expected in some disciplines and appreciated in most.
A paragraph template
Topic sentence (your point about the literature) → sources that agree, grouped in one citation → sources that differ, and why (method, sample, period) → your evaluation → what remains open.
A note on integrity
A literature review is built on other people's work, so accurate attribution is everything: every claim about a study must be traceable to that study. Do not cite sources you have not read; if you rely on another author's summary of a study, cite it as a secondary source in the way your style guide prescribes.
Sources
- The Writing Center, University of North Carolina at Chapel Hill, Literature Reviews.
- Purdue Online Writing Lab, Paraphrasing.