How to Write a Literature Review: Steps, Format, Annotated Example and Free PDF Examples (2026)

Quick answer

To write a literature review, define a focused question, search several databases, screen what you find, map how the key papers connect, take structured notes, then group the studies into themes and write about each theme, not each paper. A good review compares and judges the evidence and ends with the gap your own work fills.

What Is a Literature Review?

A literature review is a piece of writing that collects what has already been published on a question, groups it, compares it and judges it. It tells the reader three things: what is known, where studies agree or disagree, and what is still missing.

You will meet literature reviews in two forms:

• As a chapter or section inside a thesis, dissertation, grant proposal or research paper. Here the review sets up your study. It ends with a gap that your work fills.
• As a stand-alone article (a "review paper"). Here the review is the contribution. Journals publish these as narrative reviews, systematic reviews, scoping reviews and meta-analyses.

A literature review is not a list of summaries. "Smith (2019) found X. Lee (2021) found Y. Kumar (2023) found Z." is an annotated bibliography with the formatting removed. A review joins sources together: "Three cohort studies found X, but the only randomised trial did not, possibly because it used a shorter follow-up." That joining is called synthesis, and most of this guide is about how to do it.

This guide covers the review itself. If you need the whole paper around it (methods, results, discussion), see how to write a research paper.

Updated October 2026.

Types of Literature Review (Compared)

Before you search, decide which type of review you are writing. The type sets how you search, how many papers you include and how long it will take. Most student assignments and thesis chapters are narrative reviews. Health and social-policy questions often need a systematic or scoping review.

• Narrative (traditional) review: a broad, expert overview of a topic. The author chooses what to include. Flexible and fast, but open to selection bias.
• Systematic review: answers one narrow question using a written protocol, a reproducible search across several databases, set inclusion criteria, two screeners and a quality appraisal of each study.
• Scoping review: maps how much research exists on a broad topic, what kinds of studies were done and where the gaps are. It uses a systematic search but usually does not judge study quality.
• Integrative review: combines different kinds of evidence (experiments, surveys, qualitative studies, theory) into one picture. Common in nursing and education.
• Meta-analysis: pools the numbers from comparable studies with statistics to get one overall effect size. It is usually done inside a systematic review.

Other names you may see: umbrella review (a review of reviews), rapid review (a shortened systematic review for a deadline) and critical review (a narrative review with a strong argument).

How to Write a Literature Review in 9 Steps

These steps work for a narrative review in a thesis or assignment. For a systematic review, the same steps apply but each one is stricter and written down in advance (see the PRISMA section below).

1. Define the question. Turn a topic ("social media and sleep") into a question ("How is evening social media use linked to sleep duration in 13–18-year-olds?"). In health topics, PICO (Population, Intervention, Comparison, Outcome) helps. Write down what is in scope and what is not: years, languages, study types.

1. Search more than one database. No single database has everything. Combine a broad index with a subject one: for example OpenAlex plus PubMed for health, or OpenAlex plus arXiv for computer science. List your keywords and synonyms, combine them with AND / OR, and save every search string with the date. Ocean of Papers searches OpenAlex, PubMed, arXiv, bioRxiv, medRxiv and Europe PMC at once, and shows a free PDF button when a legal open-access copy exists.

1. Screen in two passes. First read titles and abstracts and drop what is clearly off-topic. Then read the full text of what is left and apply your inclusion rules. Keep a short log of why papers were dropped. You will need it when your supervisor asks.

1. Map the field with a citation map. Searching finds papers that use your words. A citation map finds papers that are connected to your papers, even when they use different words. Start a map from your best paper:
2. - Switch to Prior works to see the references many of your papers share. These are the foundations you are expected to cite.
3. - Switch to Derivative works to see newer papers that cite your set. These keep your review up to date.
4. - Look at the clusters in the graph. Groups of papers linked by shared references often line up with the themes you will write about.
5. See what a citation network is and how to find papers that cite a paper for more detail.

1. Read and take structured notes. Use the same fields for every paper: question, sample, method, main finding (with numbers), limitations, and how it relates to other papers. A spreadsheet or a tagged reference library works better than loose notes.

1. Synthesize by theme, not by paper. Look across your notes and group findings: where do studies agree? Where do they disagree, and why (different samples, methods, years)? What has nobody studied? Each group becomes a section of your review.

1. Choose a structure. Thematic (by sub-topic) is the most common. You can also go chronological (how thinking changed over time) or methodological (what different methods found). See the template below.

1. Write. Start each paragraph with a claim about the literature, support it with several sources, and end with what it means for your question. Write the introduction and conclusion last.

1. Cite consistently. Use one style (APA 7, Harvard, Vancouver, IEEE…) and a reference manager. Export your sources as BibTeX or RIS and let the software build the reference list.

1. Define the question — Narrow the topic; set scope, years and study types
2. Search several databases — Keywords + synonyms; save every search string
3. Screen — Titles and abstracts first, then full texts
4. Map the field — Prior works = foundations, Derivative works = recent, clusters = themes
5. Read and take notes — Same fields for every paper
6. Synthesize by theme — Agreement, disagreement, gaps
7. Structure — Thematic, chronological or methodological
8. Write — Claim → evidence from several sources → meaning
9. Cite — One style, exported from a reference manager

Step 4 in Practice: Mapping the Literature with a Citation Map

Mapping is the step most students skip, and it is the one that most often saves them from a missing key paper. Here is how it works in Ocean of Papers:

• Start from a search result. Search your topic, find a strong, recent paper and click Build map. You get about 30 related papers (you can choose 15, 30 or 50).
• Read the map. Bigger circles are more cited papers. Darker circles are newer. Lines show a direct citation, shared references, co-citation or a shared author; click a line to see why two papers are linked.
• Use the three modes. Similar shows the closest neighbours. Prior works shows the older papers your set builds on, which usually includes the classics every reviewer expects. Derivative works shows the newer papers that cite your set.
• Try Timeline view. It plots papers by year and citations, so you can see when a topic took off and which recent papers are already gaining attention.
• Grow the map. Select up to 10 papers and map them together, or use the Explore panel (Similar work, Earlier work, Later work, These authors) and Add to map.
• Export. Download the papers as BibTeX, RIS or CSV for your reference manager, or a PNG of the map for your supervisor meeting.

The free plan includes unlimited search, PDFs and exports plus 5 new citation maps a month; switching modes or expanding a map does not use up a map. See the literature mapping tool page for a full walkthrough, or compare it with other tools in best literature mapping tools.

Literature Review Format and Outline Template

There is no single required literature review format, but almost every good review has the same three parts: an introduction, a body organised into themes, and a conclusion. Use this outline and adjust the headings to your topic.

1. Introduction (about 10% of the length) - The topic and why it matters now - The exact question or scope of the review (years, populations, study types) - How you searched (databases, keywords, date of search). Even a narrative review is stronger with two or three sentences on this. - How the review is organised ("This review first examines…, then…")

2. Body, organised by theme (about 75%) - Theme 1: a claim about what the studies show, the evidence for it, and where studies disagree - Theme 2: same pattern - Theme 3: same pattern - Optional: a short section on methods used in the field, if methods explain the disagreements

3. Discussion and conclusion (about 15%) - What the literature, taken together, tells us - Limits of the evidence (small samples, one country, short follow-up) - The gap: what is not yet known - How your study, or future research, addresses that gap

4. References - Every source cited, in one style

Other structures you can use for the body: - Chronological: early work → turning point → current debate. Good when ideas changed over time. - Methodological: what experiments found → what surveys found → what qualitative studies found. Good when methods explain why results differ. - Theoretical: competing theories, each with its evidence. Common in social science and philosophy.

How long should a literature review be? In a master's thesis the review chapter is often 20–30% of the thesis; in a PhD it may be a full chapter of 8,000–15,000 words; inside a journal article it is usually folded into the introduction (a few paragraphs). Always check your department or journal guidelines first.

Literature Review Example (Annotated)

Below is a short illustrative example we wrote to show the difference between summary and synthesis. The topic, studies and numbers are invented for teaching; do not cite them.

Version A: summary (weak)

Patel (2018) studied 400 students and found that evening phone use reduced sleep. Garcia (2020) surveyed 1,200 teenagers and found a link between social media and later bedtimes. Chen (2022) ran an experiment and found that removing phones for one week did not change total sleep.

→ Annotation: each sentence is about one paper. The reader has to work out how the studies relate. Nothing is compared, weighed or questioned.

Version B: synthesis (strong)

Survey studies consistently link evening social media use to later bedtimes and shorter sleep in adolescents (Patel, 2018; Garcia, 2020). However, these studies are cross-sectional and rely on self-reported screen time, so they cannot show which comes first. The only experimental study found that removing phones for a week did not increase total sleep (Chen, 2022), which suggests the association may partly reflect other factors such as school start times or anxiety. No study has yet followed the same adolescents over a full school year, which is the gap the present study addresses.

→ Annotation, sentence by sentence:

• Sentence 1 (claim + grouped evidence): opens with a statement about the literature, not about one author. Two sources support one point.
• Sentence 2 (critique): judges the quality of that evidence (design and measurement), which is what makes it a review rather than a list.
• Sentence 3 (contrast + explanation): brings in a study that disagrees and offers a reason for the disagreement.
• Sentence 4 (gap): says what is missing and links it to the writer's own work. This is how a review "earns" the study that follows.

Useful synthesis phrases: "Several studies agree that…", "In contrast,…", "This pattern holds for X but not for Y…", "A limitation shared by these studies is…", "Taken together, the evidence suggests…", "No study to date has…".

For an annotated example of a full research paper (IMRaD), see research paper examples.

Literature Review Example PDFs (Free, Open Access)

Reading real reviews is the fastest way to learn the format. Each review below is a genuine, peer-reviewed or widely used review article, free to read through PubMed Central or arXiv. We picked one of each main type so you can compare how they are built.

How to study them: read the abstract, then the methods (how did they search and choose studies?), then skim the section headings. Notice that the headings are themes, not paper names. Finally, read the conclusion and find the sentence that states the gap.

• Combinations of physical activity, sedentary time, and sleep duration and their associations with depressive symptoms and other mental health problems in children and adolescents: a systematic review — Sampasa-Kanyinga, H., Colman, I., Goldfield, G. S., et al. (2020), International Journal of Behavioral Nutrition and Physical Activity. Systematic Review. A clear, readable systematic review. It searched four databases, assessed risk of bias, rated evidence with GRADE and used a narrative synthesis because the studies were too different to pool. A good model if your review cannot do a meta-analysis.
• Active learning increases student performance in science, engineering, and mathematics — Freeman, S., Eddy, S. L., McDonough, M., et al. (2014), Proceedings of the National Academy of Sciences (PNAS). Meta-analysis. A meta-analysis of 225 studies comparing lecturing with active learning in STEM courses. Study how it reports pooled effect sizes, checks for publication bias and tests whether results hold across class sizes and disciplines.
• A scoping review on the conduct and reporting of scoping reviews — Tricco, A. C., Lillie, E., Zarin, W., et al. (2016), BMC Medical Research Methodology. Scoping Review. A scoping review of 494 scoping reviews. It shows exactly what a scoping review does: a broad systematic search, charting of study features and counting, without pooling results. Useful twice over if you are planning a scoping review yourself.
• Effects of screentime on the health and well-being of children and adolescents: a systematic review of reviews — Stiglic, N., Viner, R. M. (2019), BMJ Open. Umbrella Review. A review of 13 systematic reviews, organised by outcome (obesity, diet, mental health, sleep and more). A good example of grading the strength of evidence theme by theme: "moderately strong", "moderate", "weak".
• A Survey of Large Language Models — Zhao, W. X., Zhou, K., Li, J., et al. (2023), arXiv (updated regularly). Narrative Review. A long narrative survey from computer science, organised by theme: pre-training, adaptation, use and evaluation. Shows how a fast-moving field is reviewed and kept current with new versions.
• Machine learning and the physical sciences — Carleo, G., Cirac, I., Cranmer, K., et al. (2019), Reviews of Modern Physics / arXiv. Narrative Review. A selective narrative review that says up front it is not exhaustive, then organises the field by area (statistical physics, particle physics and cosmology, quantum many-body physics, chemistry and materials). A model for being honest about scope.

Writing a Systematic Review? Use PRISMA

If your review is systematic, journals will usually ask you to follow PRISMA 2020 (Preferred Reporting Items for Systematic reviews and Meta-Analyses). It is a 27-item checklist plus a flow diagram that shows how many records you found, screened, excluded (with reasons) and included. Scoping reviews have their own extension, PRISMA-ScR.

What this means in practice:

• Write a protocol first and, for health topics, consider registering it (for example in PROSPERO) before you start screening.
• Report every database and the full search string, with the date you ran it.
• Screen with two people and record how disagreements were settled.
• Assess risk of bias in each included study with a named tool.
• Draw the PRISMA flow diagram from your screening log.

The checklist, flow diagram templates and extensions are on the official site: prisma-statement.org. The statement itself is free to read: Page et al., The PRISMA 2020 statement, BMJ 2021.

A narrative review does not need PRISMA, but borrowing two habits from it (saying where and how you searched, and how you chose papers) will make any review more convincing.

Common Literature Review Mistakes (and Fixes)

• Writing paper by paper. Fix: sort your notes into themes before writing a single sentence. Each paragraph should cite more than one source.
• Searching only one database or only Google. Fix: use at least two sources and record your search strings. Multi-source search tools help, but still note what you searched.
• Missing the classic papers. Fix: check Prior works on a citation map of your core papers. Papers that most of your set cite are hard to leave out.
• Being out of date. Fix: check Derivative works for papers from the last two years, and re-run your search before you submit.
• Describing without judging. Fix: for each theme, say how strong the evidence is and why (sample size, design, consistency).
• Hiding disagreements. Fix: disagreement is useful. Explain possible reasons: different populations, methods, measures or time periods.
• No clear gap. Fix: end with one or two sentences that state what is unknown and why it matters.
• Too many direct quotes. Fix: paraphrase and compare. Quote only when the exact wording matters (a definition, for example).
• Citing papers you have not read. Fix: read at least the abstract, methods and results of every source you cite. Free legal PDFs exist for many papers; see how to download research papers for free.
• Messy references. Fix: use a reference manager from day one and export from your search tool in BibTeX or RIS.

Free Tools for Each Stage of a Literature Review

You can do a complete literature review with free tools. Here is one workflow using Ocean of Papers alongside a reference manager such as Zotero.

• Search: search your topic once across OpenAlex, PubMed, arXiv, bioRxiv, medRxiv and Europe PMC (250M+ papers). Paste a DOI to jump to an exact paper. Click the PDF button to open a free legal copy when one exists (via Unpaywall, PubMed Central, arXiv, bioRxiv or medRxiv). For a single paper you already know, see how to download a research paper by DOI.
• Map: build a citation map from your best paper, then use Prior works, Derivative works and the Explore panel to find what keyword search missed. Our explainer on citation networks shows how to read one.
• Organise: save papers into a library project with notes and tags (a free account is needed to save; it syncs across devices). You can import existing references from Zotero (one-time import), a BibTeX or RIS file, or a pasted list of DOIs.
• Check what you are missing: open Discover in your project. It ranks papers your project does not have yet and tells you why each one is suggested, for example "cited by 6 of your papers" or "cites 4 of your papers". Add the useful ones with one click.
• Cite: export your project as BibTeX, RIS, CSV or text, then import it into Zotero, Mendeley or EndNote. You can also copy a single citation in APA, MLA, Chicago, Harvard, Vancouver or BibTeX.

What Ocean of Papers does not do: it does not send email alerts for new papers, does not sync both ways with Zotero, and does not have shared team projects. If you need alerts, tools such as Litmaps offer monitoring; see our comparisons with Litmaps, ResearchRabbit and Connected Papers.

Literature Review Checklist Before You Submit

• The question and scope are stated in the introduction
• You searched at least two databases and can say which, when and with what terms
• The classic papers (high in Prior works) and the newest work (Derivative works, last 2 years) are both covered
• Every body paragraph is about a theme and cites several sources
• Each theme says how strong the evidence is
• Disagreements are explained, not ignored
• The conclusion states a clear gap
• Every in-text citation is in the reference list, in one style
• For systematic reviews: PRISMA checklist and flow diagram completed

When the review is done, build the rest of your paper around it with our guide on how to write a research paper.

FAQ

What is a literature review?

A literature review is a structured summary and evaluation of published research on a topic or question. It groups studies into themes, compares their findings, judges the strength of the evidence and identifies what is still unknown. It can be a chapter in a thesis or a stand-alone review article.

How do I write a literature review step by step?

Define a focused question, search at least two databases, screen titles and abstracts and then full texts, map how the key papers connect with a citation map, take structured notes, group findings into themes, choose a structure, write theme by theme, and cite everything in one style.

What are the types of literature review?

The main types of literature review are narrative, systematic, scoping, integrative and meta-analysis. Others include umbrella reviews (reviews of reviews), rapid reviews and critical reviews. The type decides how strict your search is, whether you assess study quality and whether you pool results statistically.

What is the format of a literature review?

The usual literature review format is an introduction (topic, scope and search method), a body organised by themes (or chronologically or by method), a conclusion that states the gap, and a reference list. Systematic reviews add a methods section and a PRISMA flow diagram.

Where can I find a literature review example PDF?

The examples on this page are free to read: a systematic review, a meta-analysis, a scoping review, an umbrella review and two narrative reviews, all on PubMed Central or arXiv. To find more literature review examples in your own field, search your topic plus "review" or "systematic review" on Ocean of Papers and open the free PDF.

How many sources should a literature review have?

It depends on level and type. A short coursework review may use 15–25 sources, a master's thesis chapter 40–80 and a PhD chapter 100 or more. Systematic reviews report every included study, from a handful to several hundred. Coverage of the key work matters more than a number.

What is the difference between a literature review and a systematic review?

A narrative literature review is selective and flexible: the author decides what to include. A systematic review follows a written protocol, searches several databases in a documented way, uses set inclusion criteria and two screeners, appraises study quality and reports with PRISMA.

Can I write a literature review for free?

Yes. You can search, read open-access PDFs, build citation maps (5 new maps a month on the free plan) and export references from Ocean of Papers for free, then manage citations in a free reference manager such as Zotero.

Related guides

• How to find papers that cite a paper
• Citation network: what it is and how to make one free
• How to write a research paper
• Research paper examples (annotated, free PDFs)
• Free literature mapping tool (citation maps)

Search free research papers · Build a citation map from any search result · Browse all research guides