How to Write a Literature Review
A literature review is not a list of papers. It is an argument about what a field knows, what it disagrees about, and what is still missing — built from the papers you have read. This guide explains how to write literature review the way examiners and supervisors expect: how to search, how to organise what you find, and how to write it up so it reads as one coherent story instead of a stack of summaries.
On this page
- The answer in short
- What a literature review actually is
- Search: finding the papers that matter
- Organise: from a pile of papers to a map of the field
- How to write literature review: the structure that works
- Writing with a critical voice
- Common mistakes
- Your checklist
- What to do next
- Frequently asked questions
- Sources checked
The answer in short
A good literature review starts with a clear question, searches systematically for the papers that answer it, and then organises them by theme — not by author, and not in the order you found them. Read each paper well enough to say what it claims, how it supports the claim, and what its limits are. Then write by theme: what the field agrees on, where it disagrees, and what gap your work will fill. A review that only summarises papers one by one is a bibliography with sentences; a review that synthesises them is research.
What a literature review actually is
A literature review does three jobs. It shows you know the field: the key papers, the main debates, and the standard methods. It justifies your work: by showing what is missing, it explains why your project is worth doing. And it helps you think: writing the review is often how you discover what your own question really is. In a thesis, the review is usually a full chapter; in a paper, it is the introduction and background; in a project report, it may be a few pages. The shape changes, but the job is the same.
A review is not a summary of everything ever published on a topic. It is selective: you include the papers that matter to your question and explain why they matter. A review that tries to cover everything ends up covering nothing properly.
Search: finding the papers that matter
Start with a question, not a keyword. “What methods have been used to measure X in Indian rivers?” is a question; “rivers” is not. Turn the question into search terms, including synonyms — different authors use different words for the same thing. Search Google Scholar first: it covers journal papers, conference papers, theses, books and preprints across fields. For biomedical work, search PubMed as well — it holds more than 40 million citations for biomedical literature from MEDLINE, life science journals and online books. For physics, mathematics and computer science, check arXiv for preprints.
Use Scholar’s tools properly: “Cited by” takes you forward to newer papers that built on a paper; “Related articles” finds similar work; “Since year” limits results to recent work; and the author search finds everything by one researcher. Save useful papers in Scholar’s My Library with labels, so you can find them again. Keep a search log — the terms you used, the databases, and the date — so your review is reproducible and your supervisor can see how thorough you were.
Stop searching when new papers stop adding new ideas. You will never read everything; the goal is to cover the field’s main lines, not every paper.
Organise: from a pile of papers to a map of the field
Reading without organising is wasted effort. For each paper, write a short note: the question, the method, the main result, and one limitation. A reference manager such as Zotero stores the papers, the notes and the citations together, and it creates the bibliography for you in the style your institute requires — Zotero is free and open source, and it works with Word, LibreOffice and Google Docs.
Then group the papers by theme. Themes are the structure of your review: methods used, results found, populations studied, or debates in the field. A theme is not “papers by author A” — it is an idea, such as “studies that find X” versus “studies that find the opposite”. Draw the map on paper before you write: three to five themes, each with its papers, and a gap at the end that your work addresses.
How to write literature review: the structure that works
Write in this order. First, an introduction that states your question and why it matters, in one paragraph. Then one section per theme, in a logical order — from the general to the specific, or from the earliest work to the latest. In each section, synthesise: “Three studies found X, but two found Y, and the difference seems to be the method.” Compare papers with each other instead of describing them one at a time. End each section with what it means for your question.
Finally, a conclusion that names the gap: what the field does not yet know, and how your work addresses it. That gap is the justification for your project, and it is the sentence your examiner will look for. Keep the writing plain and the paragraphs short — one idea per paragraph, one theme per section.
Writing with a critical voice
Critical does not mean negative. It means judging: is the sample big enough, is the method appropriate, do the results support the claim? A critical review says “this study’s result is limited because it used only one site” and “these two studies disagree, possibly because of different methods” — not “this paper is bad”. Praise strong work clearly too: a review that only criticises reads as unfair.
Cite as you write, not afterwards. Every claim that is not yours needs a citation, and the citation belongs next to the claim. Keep a consistent citation style throughout — your reference manager handles the formatting, but you must check that every citation in the text appears in the bibliography and vice versa.
Common mistakes
- Summarising paper by paper. Group by theme and synthesise; do not list.
- No clear question. A review without a question has no shape.
- Citing papers you have not read. Read the paper, not just the abstract.
- Ignoring disagreement. Where papers disagree, say so and explain why.
- Writing the review last. Write it early — it shapes your project.
- Citations added at the end. Cite as you write, and check the bibliography matches.
Your checklist
- ☐ Review question written in one sentence
- ☐ Searches done in Google Scholar and the right field database; search log kept
- ☐ Each paper noted: question, method, result, limitation
- ☐ Papers grouped into three to five themes; gap identified
- ☐ Reference manager set up; citations added as you write
- ☐ Draft read once for argument, once for citations
What to do next
Read the guide on how to use Google Scholar and research databases effectively for the search tools in detail, and the guide on how to write a research proposal as a student for turning your gap into a plan. Check the Research Guides hub and the PhD and postdoc opportunities page for current calls. Set up a free ORCID identifier at orcid.org if you do not have one.
Frequently asked questions
How many papers should a literature review cover?
There is no fixed number — it depends on the question and the field. A project review might cover twenty to forty papers; a thesis chapter many more. What matters is coverage of the main lines of the field, not a count. Stop when new papers stop adding new ideas.
Can I cite a paper I have only read the abstract of?
No. Read the full paper before citing it. Abstracts can mislead: a result that looks strong in the abstract may rest on a small sample or a method you would not accept. If you cannot access the full text, find another paper or ask your institute’s library.
Should a literature review include my own opinion?
Yes — a review is your argument about the field, not a neutral list. State what you think the evidence shows, where it disagrees, and what is missing, and support every judgement with the papers you cite.
Sources checked
- Google Scholar — https://scholar.google.com/ — checked 19 November 2026 (covers journal and conference papers, theses, books, preprints; “Cited by”, “Related articles”, “Since year”, author search, My Library with labels, citation export).
- PubMed, U.S. National Library of Medicine — https://pubmed.ncbi.nlm.nih.gov/ — checked 19 November 2026 (more than 40 million citations for biomedical literature from MEDLINE, life science journals and online books).
- arXiv.org — https://arxiv.org/ — open-access preprint server for physics, mathematics, computer science and related fields (checked 19 November 2026).
- Zotero — https://www.zotero.org/ — checked 19 November 2026 (free, open source reference manager; collects, organises and cites research; works with Word, LibreOffice and Google Docs; over 9,000 citation styles).
- ORCID — https://orcid.org/ — free persistent researcher identifier (checked 19 November 2026).
- The search method, theming, structure and critical-voice advice in this guide are editorial guidance from IISER Connect, not rules stated by any institute. No paper count or review length is stated as a rule, because these depend on the question and the field.